<?xml version="1.0" encoding="UTF-8"?><rss version="2.0" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>ITv6 Blog</title><description>Notes on networking, IPv6, IPsec and post-quantum cryptography.</description><link>https://www.itv6.be/</link><language>en</language><item><title>Post-Quantum Cryptography in IPsec: IKEv2, Standards and Migration</title><link>https://www.itv6.be/blog/pqc-in-ipsec/</link><guid isPermaLink="true">https://www.itv6.be/blog/pqc-in-ipsec/</guid><description>How IKEv2 and IPsec become quantum-resistant: the IKEv2 extensions, IANA code points, ML-KEM and ML-DSA drafts, packet sizes, CNSA 2.0 for IPsec, and a practical migration path. Status as of November 2026.</description><pubDate>Sun, 22 Nov 2026 00:00:00 GMT</pubDate><content:encoded>&lt;aside class=&quot;tldr&quot;&gt;&lt;p&gt;&lt;strong&gt;TL;DR&lt;/strong&gt;&lt;/p&gt;&lt;ol&gt;
&lt;li&gt;&lt;strong&gt;Retire IKEv1, and modernise the classical suite&lt;/strong&gt; if you have not done so already: AES-256-GCM, SHA-384/512 for PRF and integrity, and ECDH groups &lt;strong&gt;20, 21, 29, 30, 32&lt;/strong&gt; (P-384, P-521, brainpoolP384r1, brainpoolP512r1, Curve448). Larger MODP groups (DH-4096) do not help against a quantum computer (&lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/#13-how-far-away-is-a-crqc&quot;&gt;why not&lt;/a&gt;).&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Upgrade your VPN software or hardware&lt;/strong&gt; to a version that supports &lt;strong&gt;RFC 7383&lt;/strong&gt; (IKE fragmentation), &lt;strong&gt;RFC 9242&lt;/strong&gt; (IKE_INTERMEDIATE) and &lt;strong&gt;RFC 9370&lt;/strong&gt; (multiple key exchanges). If your vendor has no roadmap for these, consider other hardware.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Make the key exchange quantum-resistant:&lt;/strong&gt;
&lt;ul&gt;
&lt;li&gt;As soon as your platform supports &lt;code&gt;draft-ietf-ipsecme-ikev2-mlkem&lt;/code&gt;: enable &lt;strong&gt;multiple key exchanges&lt;/strong&gt; combining (EC)DH with &lt;strong&gt;ML-KEM&lt;/strong&gt; (a PQ/T hybrid). Use ML-KEM-768, or ML-KEM-1024 for CNSA 2.0 or the Australian ISM.&lt;/li&gt;
&lt;li&gt;Alternatively, or in addition, where &lt;strong&gt;RFC 8784&lt;/strong&gt; or &lt;strong&gt;RFC 9867&lt;/strong&gt; is supported: mix a &lt;strong&gt;post-quantum preshared key (PPK)&lt;/strong&gt; into the (EC)DH exchange.&lt;/li&gt;
&lt;li&gt;Note: RFC 8784 and RFC 9867 are mutually exclusive within one IKE SA (the responder picks one). Either can be combined with an ML-KEM hybrid.&lt;/li&gt;
&lt;/ul&gt;
&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Keep the setup crypto-agile.&lt;/strong&gt; Prefer &lt;strong&gt;PQ/T hybrid over PQ-only&lt;/strong&gt;: if a flaw is ever found in ML-KEM or in its implementation, the traditional half still protects you until you can switch algorithms (&lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/#why-pqt-hybrid-rather-than-pq-only&quot;&gt;scenarios&lt;/a&gt;).&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Close the downgrade door:&lt;/strong&gt; once all peers support it, refuse proposals without ML-KEM (or without a PPK).&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Know your deadlines&lt;/strong&gt; (&lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/#15-policy-deadlines&quot;&gt;timeline&lt;/a&gt;) and &lt;strong&gt;plan the migration to PQ authentication for 2030–2031&lt;/strong&gt;.&lt;/li&gt;
&lt;/ol&gt;&lt;/aside&gt;
&lt;p&gt;This is part 2 of a two-part series. Part 1, &lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/&quot;&gt;&lt;strong&gt;Post-Quantum Cryptography: Threat, Standards and Timelines&lt;/strong&gt;&lt;/a&gt;, explains why PQC is urgent, who sets the rules, the hybrid concepts and the algorithms. This part applies that to IKEv2 and IPsec.&lt;/p&gt;
&lt;h2 id=&quot;1-ikev2-in-a-nutshell&quot;&gt;1. IKEv2 in a nutshell&lt;/h2&gt;
&lt;h3 id=&quot;11-the-exchanges&quot;&gt;1.1 The exchanges&lt;/h3&gt;
&lt;p&gt;IPsec protects IP packets with &lt;strong&gt;ESP&lt;/strong&gt; (Encapsulating Security Payload). The keys come from &lt;strong&gt;IKEv2&lt;/strong&gt;, specified in &lt;strong&gt;RFC 7296&lt;/strong&gt; (October 2014). RFC 7296 is one of the few IETF Internet Standards (STD 79).&lt;/p&gt;
 &lt;figure class=&quot;viz ike-diagram&quot; data-astro-cid-57zxa6jv&gt; &lt;figcaption data-astro-cid-57zxa6jv&gt;IKEv2 exchanges&lt;/figcaption&gt; &lt;p class=&quot;viz-sub&quot; data-astro-cid-57zxa6jv&gt;Each exchange is a request and a response. Hover over or tab to a message to see its payloads. Blue arrows (→) are requests, orange arrows (←) are responses. Dashed = optional IKE_INTERMEDIATE (RFC 9242), where large PQ key exchanges go.&lt;/p&gt; &lt;div class=&quot;ike-head&quot; data-astro-cid-57zxa6jv&gt; &lt;span data-astro-cid-57zxa6jv&gt;Initiator&lt;/span&gt;&lt;span data-astro-cid-57zxa6jv&gt;&lt;/span&gt;&lt;span data-astro-cid-57zxa6jv&gt;Responder&lt;/span&gt; &lt;/div&gt; &lt;ol class=&quot;ike-flow&quot; data-astro-cid-57zxa6jv&gt; &lt;li class=&quot;ike-row  first&quot; data-astro-cid-57zxa6jv&gt; &lt;span class=&quot;ike-ex&quot; data-astro-cid-57zxa6jv&gt;IKE_SA_INIT&lt;/span&gt; &lt;div class=&quot;ike-msg dir-r&quot; tabindex=&quot;0&quot; data-tip=&quot;IKE_SA_INIT request (unencrypted)
HDR, SAi1, KEi, Ni [, N(...)]
• SAi1: proposed algorithms (ENCR, PRF, INTEG, KE, ADDKE…)
• KEi: (EC)DH public value, or an ML-KEM encapsulation key
• Ni: nonce
Cannot be fragmented by IKE (RFC 7383).&quot; aria-label=&quot;IKE_SA_INIT request: HDR, SAi1, KEi, Ni [, N(...)]&quot; data-astro-cid-57zxa6jv&gt; &lt;span class=&quot;ike-label&quot; data-astro-cid-57zxa6jv&gt;IKE_SA_INIT request&lt;/span&gt; &lt;span class=&quot;ike-arrow&quot; aria-hidden=&quot;true&quot; data-astro-cid-57zxa6jv&gt;&lt;/span&gt; &lt;code class=&quot;ike-payload&quot; data-astro-cid-57zxa6jv&gt;HDR, SAi1, KEi, Ni [, N(...)]&lt;/code&gt; &lt;/div&gt; &lt;/li&gt;&lt;li class=&quot;ike-row  &quot; data-astro-cid-57zxa6jv&gt;  &lt;div class=&quot;ike-msg dir-l&quot; tabindex=&quot;0&quot; data-tip=&quot;IKE_SA_INIT response (unencrypted)
HDR, SAr1, KEr, Nr [, CERTREQ]
• SAr1: the chosen algorithms
• KEr: (EC)DH public value, or an ML-KEM ciphertext
Both sides now derive SKEYSEED and the SK_* keys.&quot; aria-label=&quot;IKE_SA_INIT response: HDR, SAr1, KEr, Nr [, CERTREQ]&quot; data-astro-cid-57zxa6jv&gt; &lt;span class=&quot;ike-label&quot; data-astro-cid-57zxa6jv&gt;IKE_SA_INIT response&lt;/span&gt; &lt;span class=&quot;ike-arrow&quot; aria-hidden=&quot;true&quot; data-astro-cid-57zxa6jv&gt;&lt;/span&gt; &lt;code class=&quot;ike-payload&quot; data-astro-cid-57zxa6jv&gt;HDR, SAr1, KEr, Nr [, CERTREQ]&lt;/code&gt; &lt;/div&gt; &lt;/li&gt;&lt;li class=&quot;ike-row opt first&quot; data-astro-cid-57zxa6jv&gt; &lt;span class=&quot;ike-ex&quot; data-astro-cid-57zxa6jv&gt;IKE_INTERMEDIATE&lt;/span&gt; &lt;div class=&quot;ike-msg dir-r&quot; tabindex=&quot;0&quot; data-tip=&quot;IKE_INTERMEDIATE request (RFC 9242, encrypted)
HDR, SK{KEi(n)}
Carries an additional key exchange (RFC 9370), e.g. an ML-KEM encapsulation key.
Encrypted, so IKE fragmentation works. Repeated once per additional key exchange.&quot; aria-label=&quot;IKE_INTERMEDIATE request (optional): HDR, SK{KEi(n)}&quot; data-astro-cid-57zxa6jv&gt; &lt;span class=&quot;ike-label&quot; data-astro-cid-57zxa6jv&gt;IKE_INTERMEDIATE request (optional)&lt;/span&gt; &lt;span class=&quot;ike-arrow&quot; aria-hidden=&quot;true&quot; data-astro-cid-57zxa6jv&gt;&lt;/span&gt; &lt;code class=&quot;ike-payload&quot; data-astro-cid-57zxa6jv&gt;HDR, SK{KEi(n)}&lt;/code&gt; &lt;/div&gt; &lt;/li&gt;&lt;li class=&quot;ike-row opt &quot; data-astro-cid-57zxa6jv&gt;  &lt;div class=&quot;ike-msg dir-l&quot; tabindex=&quot;0&quot; data-tip=&quot;IKE_INTERMEDIATE response (encrypted)
HDR, SK{KEr(n)}
E.g. the ML-KEM ciphertext. The new shared secret is mixed into SKEYSEED(n).&quot; aria-label=&quot;IKE_INTERMEDIATE response (optional): HDR, SK{KEr(n)}&quot; data-astro-cid-57zxa6jv&gt; &lt;span class=&quot;ike-label&quot; data-astro-cid-57zxa6jv&gt;IKE_INTERMEDIATE response (optional)&lt;/span&gt; &lt;span class=&quot;ike-arrow&quot; aria-hidden=&quot;true&quot; data-astro-cid-57zxa6jv&gt;&lt;/span&gt; &lt;code class=&quot;ike-payload&quot; data-astro-cid-57zxa6jv&gt;HDR, SK{KEr(n)}&lt;/code&gt; &lt;/div&gt; &lt;/li&gt;&lt;li class=&quot;ike-row  first&quot; data-astro-cid-57zxa6jv&gt; &lt;span class=&quot;ike-ex&quot; data-astro-cid-57zxa6jv&gt;IKE_AUTH&lt;/span&gt; &lt;div class=&quot;ike-msg dir-r&quot; tabindex=&quot;0&quot; data-tip=&quot;IKE_AUTH request (encrypted)
HDR, SK{IDi, [CERT,] [CERTREQ,] [IDr,] AUTH, SAi2, TSi, TSr}
• IDi / CERT / AUTH: initiator identity and proof (signature or PSK)
• SAi2, TSi, TSr: proposal and traffic selectors for the first Child SA (ESP)&quot; aria-label=&quot;IKE_AUTH request: HDR, SK{IDi, [CERT,] AUTH, SAi2, TSi, TSr}&quot; data-astro-cid-57zxa6jv&gt; &lt;span class=&quot;ike-label&quot; data-astro-cid-57zxa6jv&gt;IKE_AUTH request&lt;/span&gt; &lt;span class=&quot;ike-arrow&quot; aria-hidden=&quot;true&quot; data-astro-cid-57zxa6jv&gt;&lt;/span&gt; &lt;code class=&quot;ike-payload&quot; data-astro-cid-57zxa6jv&gt;HDR, SK{IDi, [CERT,] AUTH, SAi2, TSi, TSr}&lt;/code&gt; &lt;/div&gt; &lt;/li&gt;&lt;li class=&quot;ike-row  &quot; data-astro-cid-57zxa6jv&gt;  &lt;div class=&quot;ike-msg dir-l&quot; tabindex=&quot;0&quot; data-tip=&quot;IKE_AUTH response (encrypted)
HDR, SK{IDr, [CERT,] AUTH, SAr2, TSi, TSr}
Responder authenticates; the first Child SA (ESP keys) is now established.&quot; aria-label=&quot;IKE_AUTH response: HDR, SK{IDr, [CERT,] AUTH, SAr2, TSi, TSr}&quot; data-astro-cid-57zxa6jv&gt; &lt;span class=&quot;ike-label&quot; data-astro-cid-57zxa6jv&gt;IKE_AUTH response&lt;/span&gt; &lt;span class=&quot;ike-arrow&quot; aria-hidden=&quot;true&quot; data-astro-cid-57zxa6jv&gt;&lt;/span&gt; &lt;code class=&quot;ike-payload&quot; data-astro-cid-57zxa6jv&gt;HDR, SK{IDr, [CERT,] AUTH, SAr2, TSi, TSr}&lt;/code&gt; &lt;/div&gt; &lt;/li&gt;&lt;li class=&quot;ike-row  first&quot; data-astro-cid-57zxa6jv&gt; &lt;span class=&quot;ike-ex&quot; data-astro-cid-57zxa6jv&gt;CREATE_CHILD_SA&lt;/span&gt; &lt;div class=&quot;ike-msg dir-r&quot; tabindex=&quot;0&quot; data-tip=&quot;CREATE_CHILD_SA request (encrypted)
HDR, SK{SA, Ni, [KEi,] [TSi, TSr]}
Creates an extra Child SA or rekeys an IKE/Child SA. Either peer may start it.
An optional KE gives PFS; extra PQ exchanges follow in IKE_FOLLOWUP_KE (RFC 9370).&quot; aria-label=&quot;CREATE_CHILD_SA request: HDR, SK{SA, Ni, [KEi,] [TSi, TSr]}&quot; data-astro-cid-57zxa6jv&gt; &lt;span class=&quot;ike-label&quot; data-astro-cid-57zxa6jv&gt;CREATE_CHILD_SA request&lt;/span&gt; &lt;span class=&quot;ike-arrow&quot; aria-hidden=&quot;true&quot; data-astro-cid-57zxa6jv&gt;&lt;/span&gt; &lt;code class=&quot;ike-payload&quot; data-astro-cid-57zxa6jv&gt;HDR, SK{SA, Ni, [KEi,] [TSi, TSr]}&lt;/code&gt; &lt;/div&gt; &lt;/li&gt;&lt;li class=&quot;ike-row  &quot; data-astro-cid-57zxa6jv&gt;  &lt;div class=&quot;ike-msg dir-l&quot; tabindex=&quot;0&quot; data-tip=&quot;CREATE_CHILD_SA response (encrypted)
HDR, SK{SA, Nr, [KEr,] [TSi, TSr]}&quot; aria-label=&quot;CREATE_CHILD_SA response: HDR, SK{SA, Nr, [KEr,] [TSi, TSr]}&quot; data-astro-cid-57zxa6jv&gt; &lt;span class=&quot;ike-label&quot; data-astro-cid-57zxa6jv&gt;CREATE_CHILD_SA response&lt;/span&gt; &lt;span class=&quot;ike-arrow&quot; aria-hidden=&quot;true&quot; data-astro-cid-57zxa6jv&gt;&lt;/span&gt; &lt;code class=&quot;ike-payload&quot; data-astro-cid-57zxa6jv&gt;HDR, SK{SA, Nr, [KEr,] [TSi, TSr]}&lt;/code&gt; &lt;/div&gt; &lt;/li&gt;&lt;li class=&quot;ike-row  first&quot; data-astro-cid-57zxa6jv&gt; &lt;span class=&quot;ike-ex&quot; data-astro-cid-57zxa6jv&gt;INFORMATIONAL&lt;/span&gt; &lt;div class=&quot;ike-msg dir-r&quot; tabindex=&quot;0&quot; data-tip=&quot;INFORMATIONAL request (encrypted)
HDR, SK{[N,] [D,] [CP,] …}
Notifications, deleting SAs, liveness checks (an empty message acts as a keep-alive). Either peer may send it.&quot; aria-label=&quot;INFORMATIONAL request: HDR, SK{[N,] [D,] [CP,] …}&quot; data-astro-cid-57zxa6jv&gt; &lt;span class=&quot;ike-label&quot; data-astro-cid-57zxa6jv&gt;INFORMATIONAL request&lt;/span&gt; &lt;span class=&quot;ike-arrow&quot; aria-hidden=&quot;true&quot; data-astro-cid-57zxa6jv&gt;&lt;/span&gt; &lt;code class=&quot;ike-payload&quot; data-astro-cid-57zxa6jv&gt;HDR, SK{[N,] [D,] [CP,] …}&lt;/code&gt; &lt;/div&gt; &lt;/li&gt;&lt;li class=&quot;ike-row  &quot; data-astro-cid-57zxa6jv&gt;  &lt;div class=&quot;ike-msg dir-l&quot; tabindex=&quot;0&quot; data-tip=&quot;INFORMATIONAL response (encrypted)
Every IKEv2 request gets a response; this is how reliability is achieved over UDP.&quot; aria-label=&quot;INFORMATIONAL response: HDR, SK{[N,] [D,] [CP,] …}&quot; data-astro-cid-57zxa6jv&gt; &lt;span class=&quot;ike-label&quot; data-astro-cid-57zxa6jv&gt;INFORMATIONAL response&lt;/span&gt; &lt;span class=&quot;ike-arrow&quot; aria-hidden=&quot;true&quot; data-astro-cid-57zxa6jv&gt;&lt;/span&gt; &lt;code class=&quot;ike-payload&quot; data-astro-cid-57zxa6jv&gt;HDR, SK{[N,] [D,] [CP,] …}&lt;/code&gt; &lt;/div&gt; &lt;/li&gt; &lt;/ol&gt; &lt;/figure&gt; 
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;IKE_SA_INIT&lt;/strong&gt; negotiates the cryptographic suite (encryption, PRF, integrity, key exchange method; see &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#21-transforms-transform-types-and-transform-ids&quot;&gt;transforms&lt;/a&gt;) and performs the key exchange in the KE payloads. From the shared secret and the nonces, both sides derive &lt;code&gt;SKEYSEED&lt;/code&gt;. From that come the keys &lt;code&gt;SK_d&lt;/code&gt; (for deriving Child SA keys), &lt;code&gt;SK_ai/ar&lt;/code&gt; (integrity), &lt;code&gt;SK_ei/er&lt;/code&gt; (encryption) and &lt;code&gt;SK_pi/pr&lt;/code&gt; (used in the AUTH computation). This exchange is &lt;strong&gt;not encrypted&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;IKE_INTERMEDIATE&lt;/strong&gt; (optional, RFC 9242) sits between IKE_SA_INIT and IKE_AUTH and is already encrypted. It is where additional, large PQ key exchanges go.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;IKE_AUTH&lt;/strong&gt; is encrypted with those keys. Each peer proves its identity in the AUTH payload, with a signature or a pre-shared key, and the first ESP Child SA is set up.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;CREATE_CHILD_SA&lt;/strong&gt; creates more Child SAs and rekeys existing ones. &lt;strong&gt;INFORMATIONAL&lt;/strong&gt; carries notifications, deletes and liveness checks.&lt;/li&gt;
&lt;/ul&gt;
&lt;h3 id=&quot;12-where-the-quantum-computer-hits&quot;&gt;1.2 Where the quantum computer hits&lt;/h3&gt;





























&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Phase / exchange&lt;/th&gt;&lt;th&gt;Component&lt;/th&gt;&lt;th&gt;Quantum impact&lt;/th&gt;&lt;th&gt;Urgency&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;IKE_SA_INIT, CREATE_CHILD_SA (rekey with PFS)&lt;/td&gt;&lt;td&gt;&lt;strong&gt;KE (DH/ECDH)&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;Broken by Shor; enables decryption of recorded traffic (&lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/#14-harvest-now-decrypt-later&quot;&gt;HNDL&lt;/a&gt;)&lt;/td&gt;&lt;td&gt;&lt;strong&gt;High (HNDL)&lt;/strong&gt;&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;IKE_AUTH&lt;/td&gt;&lt;td&gt;&lt;strong&gt;AUTH&lt;/strong&gt; (RSA, ECDSA, EdDSA signatures) and certificates&lt;/td&gt;&lt;td&gt;Broken by Shor; enables active impersonation once a CRQC exists&lt;/td&gt;&lt;td&gt;Medium, rising&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;All exchanges and ESP&lt;/td&gt;&lt;td&gt;&lt;strong&gt;PRF, integrity, encryption&lt;/strong&gt; (HMAC, AES)&lt;/td&gt;&lt;td&gt;Only Grover; fine with adequate key sizes&lt;/td&gt;&lt;td&gt;Low&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;h3 id=&quot;13-extensions-that-make-pqc-possible-in-ikev2&quot;&gt;1.3 Extensions that make PQC possible in IKEv2&lt;/h3&gt;
&lt;p&gt;Each of these RFCs fixes a specific limitation of the original protocol. &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#3-why-ikev2-needs-extensions-for-pure-pq-pqt-and-pqpq&quot;&gt;Section 3&lt;/a&gt; explains them in detail:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;RFC 7383&lt;/strong&gt;: IKEv2 message fragmentation. It only works on &lt;em&gt;encrypted&lt;/em&gt; messages, so &lt;strong&gt;not&lt;/strong&gt; on IKE_SA_INIT. See &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#32-assumption-2-ike_sa_init-messages-are-small&quot;&gt;3.2&lt;/a&gt;.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;RFC 9242&lt;/strong&gt;: the &lt;strong&gt;IKE_INTERMEDIATE&lt;/strong&gt; exchange, encrypted and therefore fragmentable; the place for large PQ public keys. See &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#32-assumption-2-ike_sa_init-messages-are-small&quot;&gt;3.2&lt;/a&gt;.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;RFC 9370&lt;/strong&gt;: &lt;strong&gt;multiple key exchanges&lt;/strong&gt;, up to seven additional ones (ADDKE1–7), each mixed into the keying material. See &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#31-assumption-1-exactly-one-key-exchange-per-sa&quot;&gt;3.1&lt;/a&gt;.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;RFC 7427&lt;/strong&gt;: generic Digital Signature authentication (method 14). See &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#34-assumption-4-one-signature-algorithm-per-authentication-method&quot;&gt;3.4&lt;/a&gt; and &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#23-authentication-method-14-how-it-was-done-before&quot;&gt;2.3&lt;/a&gt;.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;RFC 9593&lt;/strong&gt;: announcing supported authentication methods (SUPPORTED_AUTH_METHODS) before IKE_AUTH. See &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#34-assumption-4-one-signature-algorithm-per-authentication-method&quot;&gt;3.4&lt;/a&gt;.&lt;/li&gt;
&lt;/ul&gt;
&lt;h3 id=&quot;14-why-rfc-9395-ikev1-deprecation-matters-for-pqc&quot;&gt;1.4 Why RFC 9395 (IKEv1 deprecation) matters for PQC&lt;/h3&gt;
&lt;p&gt;&lt;strong&gt;RFC 9395&lt;/strong&gt; (April 2023) deprecated IKEv1 and moved RFCs 2407, 2408 and 2409 to Historic, together with a set of obsolete algorithms. That matters for PQC in three ways:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Every PQC mechanism is IKEv2-only.&lt;/strong&gt; IKE_INTERMEDIATE, multiple key exchanges, ML-KEM code points, PQ signature authentication and PPK mixing are all defined for IKEv2. IKEv1 has no standardised way to carry an additional key exchange.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;No new work for IKEv1.&lt;/strong&gt; Historic status means the IETF will not add new algorithms or extensions to IKEv1. An IKEv1 tunnel therefore cannot become quantum-resistant through standards.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;It is a good inventory trigger.&lt;/strong&gt; The same document deprecates old algorithms (DES variants, MD5- and SHA-1-based integrity, among others). Systems that still use them usually also lack the firmware headroom for PQC.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;If you still run IKEv1, migrating to IKEv2 is step zero of your PQC migration.&lt;/p&gt;
&lt;hr/&gt;
&lt;h2 id=&quot;2-iana-the-ikev2-registries&quot;&gt;2. IANA: the IKEv2 registries&lt;/h2&gt;
&lt;p&gt;The &lt;strong&gt;Internet Assigned Numbers Authority (IANA)&lt;/strong&gt; keeps the protocol registries (for the other standards bodies, see &lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/#2-who-sets-the-rules&quot;&gt;part 1&lt;/a&gt;). An IETF document defines &lt;em&gt;what&lt;/em&gt; an algorithm does in a protocol; IANA assigns the &lt;em&gt;number&lt;/em&gt; that goes on the wire. For IKEv2 everything lives in the &lt;em&gt;Internet Key Exchange Version 2 (IKEv2) Parameters&lt;/em&gt; registry. Most of its sub-registries use &lt;strong&gt;Expert Review&lt;/strong&gt; (current experts: Tero Kivinen and Valery Smyslov). The registry was last updated on 13 November 2026.&lt;/p&gt;
&lt;h3 id=&quot;21-transforms-transform-types-and-transform-ids&quot;&gt;2.1 Transforms, transform types and transform IDs&lt;/h3&gt;
&lt;p&gt;In IKE_SA_INIT the initiator offers one or more &lt;strong&gt;proposals&lt;/strong&gt; in the SA payload. Each proposal is a list of &lt;strong&gt;transforms&lt;/strong&gt;. Every transform has two numbers:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;a &lt;strong&gt;Transform Type&lt;/strong&gt;, saying &lt;em&gt;what role&lt;/em&gt; the algorithm plays: 1 = encryption (ENCR), 2 = pseudo-random function (PRF), 3 = integrity (INTEG), 4 = key exchange method (KE), 5 = sequence numbers, and since RFC 9370, 6–12 = additional key exchanges 1–7 (ADDKE1–ADDKE7);&lt;/li&gt;
&lt;li&gt;a &lt;strong&gt;Transform ID&lt;/strong&gt;, saying &lt;em&gt;which algorithm&lt;/em&gt; within that role. The numbering is separate per type.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;“Transform Type 4, ID 36” therefore means “key exchange method: ML-KEM-768”. The same number 36 under another transform type means something else entirely. Transform Type 4 IDs used to be called &lt;em&gt;Diffie-Hellman group numbers&lt;/em&gt;, and that is why many configuration guides still talk about “DH group 19” or “DH group 31”. A PQ/T hybrid proposal could look like this:&lt;/p&gt;
&lt;pre class=&quot;astro-code github-dark&quot; style=&quot;background-color:#24292e;color:#e1e4e8;overflow-x:auto&quot; tabindex=&quot;0&quot; data-language=&quot;text&quot;&gt;&lt;code&gt;&lt;span class=&quot;line&quot;&gt;&lt;span&gt;Proposal 1:&lt;/span&gt;&lt;/span&gt;
&lt;span class=&quot;line&quot;&gt;&lt;span&gt;  Type 1 (ENCR)    ID 20  ENCR_AES_GCM_16, 256-bit key&lt;/span&gt;&lt;/span&gt;
&lt;span class=&quot;line&quot;&gt;&lt;span&gt;  Type 2 (PRF)     ID 6   PRF_HMAC_SHA2_384&lt;/span&gt;&lt;/span&gt;
&lt;span class=&quot;line&quot;&gt;&lt;span&gt;  Type 4 (KE)      ID 20  384-bit random ECP group (P-384)&lt;/span&gt;&lt;/span&gt;
&lt;span class=&quot;line&quot;&gt;&lt;span&gt;  Type 6 (ADDKE1)  ID 37  ml-kem-1024&lt;/span&gt;&lt;/span&gt;&lt;/code&gt;&lt;/pre&gt;
&lt;p&gt;The responder picks one transform per type. Here that results in ECDH P-384 in IKE_SA_INIT, followed by ML-KEM-1024 in an IKE_INTERMEDIATE exchange.&lt;/p&gt;
&lt;h3 id=&quot;22-pqc-relevant-registry-entries&quot;&gt;2.2 PQC-relevant registry entries&lt;/h3&gt;


















































&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Registry&lt;/th&gt;&lt;th&gt;Entry&lt;/th&gt;&lt;th&gt;Defined by&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;Transform Types&lt;/td&gt;&lt;td&gt;Type 4 renamed from “Diffie-Hellman Group” to &lt;strong&gt;Key Exchange Method (KE)&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;RFC 9370&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Transform Types&lt;/td&gt;&lt;td&gt;Types 6–12: &lt;strong&gt;Additional Key Exchange 1–7&lt;/strong&gt; (ADDKE1…ADDKE7)&lt;/td&gt;&lt;td&gt;RFC 9370&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Transform Type 4 IDs (also valid for types 6–12)&lt;/td&gt;&lt;td&gt;&lt;strong&gt;35&lt;/strong&gt; ml-kem-512, &lt;strong&gt;36&lt;/strong&gt; ml-kem-768, &lt;strong&gt;37&lt;/strong&gt; ml-kem-1024&lt;/td&gt;&lt;td&gt;draft-ietf-ipsecme-ikev2-mlkem&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Authentication Method&lt;/td&gt;&lt;td&gt;&lt;strong&gt;14&lt;/strong&gt; Digital Signature (generic, see below)&lt;/td&gt;&lt;td&gt;RFC 7427&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Hash Algorithms (SIGNATURE_HASH_ALGORITHMS)&lt;/td&gt;&lt;td&gt;&lt;strong&gt;5&lt;/strong&gt; Identity (needed for ML-DSA/SLH-DSA, which hash internally)&lt;/td&gt;&lt;td&gt;RFC 8420&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Notify types&lt;/td&gt;&lt;td&gt;USE_PPK (16435), PPK_IDENTITY (16436), NO_PPK_AUTH (16437)&lt;/td&gt;&lt;td&gt;RFC 8784&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Notify types&lt;/td&gt;&lt;td&gt;USE_PPK_INT (16445), PPK_IDENTITY_KEY (16446)&lt;/td&gt;&lt;td&gt;RFC 9867&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Transform registries&lt;/td&gt;&lt;td&gt;“Status” column marking deprecated algorithms&lt;/td&gt;&lt;td&gt;RFC 9395&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;h3 id=&quot;23-authentication-method-14-how-it-was-done-before&quot;&gt;2.3 Authentication method 14: how it was done before&lt;/h3&gt;
&lt;p&gt;In the AUTH payload, the &lt;em&gt;Auth Method&lt;/em&gt; field tells the receiver how to verify the authentication data. Originally, &lt;strong&gt;each method number fixed the complete algorithm, including the hash&lt;/strong&gt;:&lt;/p&gt;









































&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Value&lt;/th&gt;&lt;th&gt;Method&lt;/th&gt;&lt;th&gt;Defined in&lt;/th&gt;&lt;th&gt;What was fixed&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;1&lt;/td&gt;&lt;td&gt;RSA Digital Signature&lt;/td&gt;&lt;td&gt;RFC 7296&lt;/td&gt;&lt;td&gt;RSASSA-PKCS1-v1_5; the hash was not negotiated&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;2&lt;/td&gt;&lt;td&gt;Shared Key Message Integrity Code&lt;/td&gt;&lt;td&gt;RFC 7296&lt;/td&gt;&lt;td&gt;PSK-based MAC with the negotiated PRF&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;3&lt;/td&gt;&lt;td&gt;DSS Digital Signature&lt;/td&gt;&lt;td&gt;RFC 7296&lt;/td&gt;&lt;td&gt;DSA with SHA-1&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;9 / 10 / 11&lt;/td&gt;&lt;td&gt;ECDSA with SHA-256 on P-256 / SHA-384 on P-384 / SHA-512 on P-521&lt;/td&gt;&lt;td&gt;RFC 4754&lt;/td&gt;&lt;td&gt;curve &lt;strong&gt;and&lt;/strong&gt; hash per number&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;14&lt;/td&gt;&lt;td&gt;&lt;strong&gt;Digital Signature&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;RFC 7427&lt;/td&gt;&lt;td&gt;nothing: the algorithm is named inside the AUTH data&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;p&gt;This model did not scale. Every new combination of signature scheme, parameters and hash, such as RSA-PSS or Ed25519, needed a new method number and a new RFC. &lt;strong&gt;RFC 7427 (2015)&lt;/strong&gt; solved this with the generic method 14. The AUTH data starts with an ASN.1 &lt;strong&gt;AlgorithmIdentifier&lt;/strong&gt;, the same object identifiers (OIDs) used in X.509 certificates, followed by the signature. The hash function is negotiated separately with the SIGNATURE_HASH_ALGORITHMS notification.&lt;/p&gt;
&lt;p&gt;For PQC this is very convenient. ML-DSA, SLH-DSA and composite ML-DSA need &lt;strong&gt;no new IKEv2 code points&lt;/strong&gt;: they reuse the OIDs from RFC 9881, RFC 9909 and RFC 10073. &lt;code&gt;draft-ietf-ipsecme-ikev2-pqc-auth&lt;/code&gt; makes no IANA requests at all. The only addition is the “Identity” hash (5), because ML-DSA and SLH-DSA sign the whole message themselves instead of a pre-computed hash (&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#34-assumption-4-one-signature-algorithm-per-authentication-method&quot;&gt;section 3.4&lt;/a&gt;).&lt;/p&gt;
&lt;h3 id=&quot;24-the-ppk-notifications&quot;&gt;2.4 The PPK notifications&lt;/h3&gt;
&lt;p&gt;Post-quantum preshared keys (PPKs) are negotiated with five Notify payloads:&lt;/p&gt;









































&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Notify (value)&lt;/th&gt;&lt;th&gt;RFC&lt;/th&gt;&lt;th&gt;Sent in&lt;/th&gt;&lt;th&gt;Purpose&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;USE_PPK&lt;/strong&gt; (16435)&lt;/td&gt;&lt;td&gt;RFC 8784&lt;/td&gt;&lt;td&gt;IKE_SA_INIT, both directions&lt;/td&gt;&lt;td&gt;“I support PPKs and want to use one.” No data.&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;PPK_IDENTITY&lt;/strong&gt; (16436)&lt;/td&gt;&lt;td&gt;RFC 8784&lt;/td&gt;&lt;td&gt;IKE_AUTH (request); reused by RFC 9867 in the IKE_INTERMEDIATE response&lt;/td&gt;&lt;td&gt;Identifies &lt;em&gt;which&lt;/em&gt; PPK is used (the PPK_ID). In RFC 9867 the responder uses it to confirm the PPK it selected.&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;NO_PPK_AUTH&lt;/strong&gt; (16437)&lt;/td&gt;&lt;td&gt;RFC 8784&lt;/td&gt;&lt;td&gt;IKE_AUTH request&lt;/td&gt;&lt;td&gt;Contains an alternative AUTH value computed &lt;strong&gt;without&lt;/strong&gt; the PPK, so that a responder without that PPK can still authenticate the initiator when PPK use is optional&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;USE_PPK_INT&lt;/strong&gt; (16445)&lt;/td&gt;&lt;td&gt;RFC 9867&lt;/td&gt;&lt;td&gt;IKE_SA_INIT, both directions, together with INTERMEDIATE_EXCHANGE_SUPPORTED&lt;/td&gt;&lt;td&gt;Negotiates the RFC 9867 variant, where the PPK is mixed in during IKE_INTERMEDIATE so that the initial IKE SA is protected too&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;PPK_IDENTITY_KEY&lt;/strong&gt; (16446)&lt;/td&gt;&lt;td&gt;RFC 9867&lt;/td&gt;&lt;td&gt;Last IKE_INTERMEDIATE request&lt;/td&gt;&lt;td&gt;One or more candidate PPKs, each as a PPK_ID plus an 8-byte &lt;em&gt;PPK Confirmation&lt;/em&gt; (first 8 bytes of prf(PPK, Ni | Nr | SPIi | SPIr)). The responder can check that it has the same PPK without the key being revealed.&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;p&gt;A responder chooses one mechanism: RFC 8784 (USE_PPK) or RFC 9867 (USE_PPK_INT), never both.&lt;/p&gt;
&lt;hr/&gt;
&lt;h2 id=&quot;3-why-ikev2-needs-extensions-for-pure-pq-pqt-and-pqpq&quot;&gt;3. Why IKEv2 needs extensions for pure PQ, PQ/T and PQ/PQ&lt;/h2&gt;
&lt;p&gt;IKEv2 as specified in RFC 7296 was designed around Diffie-Hellman and classical signatures. Several of its assumptions break as soon as post-quantum algorithms enter the picture. This section explains &lt;em&gt;which&lt;/em&gt; assumptions break and &lt;em&gt;which&lt;/em&gt; extension fixes each one.&lt;/p&gt;
&lt;h3 id=&quot;31-assumption-1-exactly-one-key-exchange-per-sa&quot;&gt;3.1 Assumption 1: exactly one key exchange per SA&lt;/h3&gt;
&lt;p&gt;In RFC 7296, the SA proposal carries a single Transform Type 4 (D-H group), and IKE_SA_INIT carries a single KE payload. There is no way to negotiate or perform two key exchanges. Every hybrid needs at least two, so this is the first blocker.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Fix: RFC 9370 (Multiple Key Exchanges).&lt;/strong&gt; It renames Transform Type 4 to &lt;em&gt;Key Exchange Method&lt;/em&gt; and adds Transform Types 6–12, &lt;em&gt;Additional Key Exchange 1–7&lt;/em&gt; (&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#21-transforms-transform-types-and-transform-ids&quot;&gt;example proposal&lt;/a&gt;). These carry up to seven extra key exchanges, run one by one in &lt;strong&gt;IKE_INTERMEDIATE&lt;/strong&gt; exchanges (during initial setup) or in &lt;strong&gt;IKE_FOLLOWUP_KE&lt;/strong&gt; exchanges (when creating or rekeying SAs). After each additional exchange, the new shared secret is chained into the keying material: &lt;code&gt;SKEYSEED(n) = prf(SK_d(n-1), SK(n) | Ni | Nr)&lt;/code&gt;. The final keys therefore depend on &lt;em&gt;all&lt;/em&gt; shared secrets, which is exactly the “secure if one component holds” property a hybrid needs.&lt;/p&gt;
&lt;p&gt;An ADDKE transform may include the value NONE, which makes that extra exchange optional. That is useful for interoperability, but it opens a downgrade path. A policy that &lt;em&gt;requires&lt;/em&gt; PQ must therefore not offer NONE.&lt;/p&gt;
&lt;h3 id=&quot;32-assumption-2-ike_sa_init-messages-are-small&quot;&gt;3.2 Assumption 2: IKE_SA_INIT messages are small&lt;/h3&gt;
&lt;p&gt;IKE_SA_INIT is sent before any keys exist, so it is neither encrypted nor integrity-protected. IKEv2 fragmentation (RFC 7383) only works inside the encrypted SK payload, which means &lt;strong&gt;IKE_SA_INIT cannot be fragmented by IKE&lt;/strong&gt;. Traditional KE payloads are small: 32 bytes for X25519, 96 for P-384, 384 for MODP-3072. With nonces, the SA proposal, NAT detection and possibly a cookie, IKE_SA_INIT fits comfortably in one packet.&lt;/p&gt;
&lt;p&gt;ML-KEM-768 adds about 1.2 KB on the initiator side, ML-KEM-1024 about 1.6 KB, and FrodoKEM roughly ten times as much (&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#6-key-exchange-and-signature-sizes-in-ikev2&quot;&gt;sizes in section 6&lt;/a&gt;). The result exceeds the 1,280-byte IPv6 minimum MTU and often the typical 1,500-byte Ethernet path. &lt;strong&gt;IP fragmentation&lt;/strong&gt; of UDP is then the only option, and fragments are frequently dropped by firewalls, NAT devices and carrier networks. Fragments are also a known denial-of-service vector.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Fix: RFC 9242 (IKE_INTERMEDIATE).&lt;/strong&gt; This new exchange sits between IKE_SA_INIT and IKE_AUTH. It is already protected with keys derived from IKE_SA_INIT, so &lt;strong&gt;it can be fragmented with RFC 7383&lt;/strong&gt;. Large PQ public keys and ciphertexts go there instead of into IKE_SA_INIT. RFC 9242 also adds the IKE_INTERMEDIATE content to the data signed in IKE_AUTH, so the intermediate exchanges are authenticated afterwards.&lt;/p&gt;
&lt;p&gt;The alternative is a transport without this limit: &lt;strong&gt;TCP encapsulation (RFC 9329)&lt;/strong&gt;, or the separate IKE and ESP transports in &lt;code&gt;draft-ietf-ipsecme-ikev2-reliable-transport&lt;/code&gt;.&lt;/p&gt;
&lt;h3 id=&quot;33-assumption-3-the-key-exchange-is-diffie-hellman&quot;&gt;3.3 Assumption 3: the key exchange is Diffie-Hellman&lt;/h3&gt;
&lt;p&gt;With Diffie-Hellman, both peers send a public value of the same kind, and either side could send first. A &lt;strong&gt;KEM&lt;/strong&gt; is asymmetric:&lt;/p&gt;
&lt;ol&gt;
&lt;li&gt;The initiator generates a key pair and sends the &lt;strong&gt;encapsulation key&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;The responder runs Encaps and sends back a &lt;strong&gt;ciphertext&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;The initiator runs Decaps. Both now share the same 32-byte secret.&lt;/li&gt;
&lt;/ol&gt;
&lt;p&gt;Luckily, this maps onto the IKEv2 request/response flow, because the initiator always speaks first. It still needs its own specification: what goes into the KE payload on each side, which validity checks must be done (FIPS 203 encapsulation key check and ciphertext type check), and the rule that &lt;strong&gt;key pairs must never be reused&lt;/strong&gt;. Some DH implementations reuse ephemeral keys for performance; for ML-KEM that is forbidden. This is what &lt;strong&gt;&lt;code&gt;draft-ietf-ipsecme-ikev2-mlkem&lt;/code&gt;&lt;/strong&gt; defines, together with the IANA code points 35–37. Every additional KEM, such as FrodoKEM or HQC, needs a similar document and code points of its own.&lt;/p&gt;
&lt;h3 id=&quot;34-assumption-4-one-signature-algorithm-per-authentication-method&quot;&gt;3.4 Assumption 4: one signature algorithm per authentication method&lt;/h3&gt;
&lt;p&gt;RFC 7296 ties authentication methods to specific algorithms (RSA = 1, DSS = 3), and RFC 4754 added ECDSA variants (9–11) with a fixed hash per curve (&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#23-authentication-method-14-how-it-was-done-before&quot;&gt;history&lt;/a&gt;). Adding a new method number for every PQ parameter set does not scale.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Fix: RFC 7427 (Digital Signature, method 14)&lt;/strong&gt; identifies the algorithm with an X.509 AlgorithmIdentifier, so the OIDs from RFC 9881 (ML-DSA), RFC 9909 (SLH-DSA) and RFC 10073 (composite) can be reused. RFC 7427 assumes &lt;em&gt;hash-then-sign&lt;/em&gt;, with the hash negotiated in SIGNATURE_HASH_ALGORITHMS. ML-DSA and SLH-DSA instead process the whole message internally, so peers must announce the &lt;strong&gt;“Identity” hash (RFC 8420)&lt;/strong&gt;. &lt;strong&gt;RFC 9593&lt;/strong&gt; (SUPPORTED_AUTH_METHODS) lets peers tell each other which signature algorithms they accept before IKE_AUTH. All of this is specified in &lt;strong&gt;&lt;code&gt;draft-ietf-ipsecme-ikev2-pqc-auth&lt;/code&gt;&lt;/strong&gt;.&lt;/p&gt;
&lt;p&gt;PQ signatures plus certificate chains add several kilobytes to IKE_AUTH. IKE_AUTH is encrypted, so RFC 7383 fragmentation can carry them, and the draft makes fragmentation support mandatory.&lt;/p&gt;
&lt;h3 id=&quot;35-assumption-5-one-auth-payload-per-peer&quot;&gt;3.5 Assumption 5: one AUTH payload per peer&lt;/h3&gt;
&lt;p&gt;A peer authenticates with one AUTH payload. A &lt;strong&gt;composite&lt;/strong&gt; signature (RFC 10073) is one algorithm, one key and one signature, so it fits without protocol changes; &lt;code&gt;draft-ietf-ipsecme-pq-hybrid-auth&lt;/code&gt; profiles it. &lt;strong&gt;Parallel&lt;/strong&gt; hybrid authentication with two independent certificates and signatures does &lt;em&gt;not&lt;/em&gt; fit and needs protocol changes. That is the subject of individual drafts such as &lt;code&gt;draft-hu-ipsecme-pqt-hybrid-auth-with-sep-cert&lt;/code&gt; and &lt;code&gt;draft-wang-ipsecme-multi-auth-ikev2-pq&lt;/code&gt;.&lt;/p&gt;
&lt;h3 id=&quot;36-what-each-mode-needs&quot;&gt;3.6 What each mode needs&lt;/h3&gt;





























&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Mode&lt;/th&gt;&lt;th&gt;Typical IKEv2 setup&lt;/th&gt;&lt;th&gt;Extensions required&lt;/th&gt;&lt;th&gt;Remaining constraints&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;Pure PQ&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;ML-KEM as the &lt;em&gt;only&lt;/em&gt; KE in IKE_SA_INIT; ML-DSA or SLH-DSA authentication&lt;/td&gt;&lt;td&gt;ML-KEM KE draft (KEM semantics, code points); PQC-auth draft + RFC 7427 + RFC 8420 + RFC 7383&lt;/td&gt;&lt;td&gt;Only &lt;strong&gt;ML-KEM-512&lt;/strong&gt; fits IKE_SA_INIT safely over UDP. ML-KEM-768/1024 need a known large PMTU or a reliable transport (RFC 9329). No fallback if ML-KEM is ever broken. This is why CNSA 2.0 does not simply put ML-KEM-1024 in IKE_SA_INIT.&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;PQ/T hybrid&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;(EC)DH in IKE_SA_INIT + ML-KEM as ADDKE1 in IKE_INTERMEDIATE; composite or PQ authentication&lt;/td&gt;&lt;td&gt;RFC 9370 + RFC 9242 + RFC 7383 + ML-KEM KE draft; for authentication, RFC 10073 + the hybrid-auth draft&lt;/td&gt;&lt;td&gt;One extra round trip. Policy must omit NONE in ADDKE1 to prevent downgrade.&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;PQ/PQ hybrid&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;Two PQ KEMs from different families, e.g. ML-KEM + FrodoKEM (or HQC)&lt;/td&gt;&lt;td&gt;Everything above + a KE specification and &lt;strong&gt;IANA code points&lt;/strong&gt; for the second KEM (FrodoKEM: WG draft; HQC: individual draft; no code points yet)&lt;/td&gt;&lt;td&gt;IKE_SA_INIT still needs a KE: a traditional group (giving a triple hybrid), ML-KEM-512, or a reliable transport. FrodoKEM’s size makes fragmentation unavoidable. Each extra KEM adds a round trip.&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;p&gt;For most deployments &lt;strong&gt;PQ/T hybrid is the sensible default&lt;/strong&gt;: it protects against a quantum computer &lt;em&gt;and&lt;/em&gt; against a flaw in the young PQ algorithm, as long as not both happen at once (&lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/#why-pqt-hybrid-rather-than-pq-only&quot;&gt;scenario table in part 1&lt;/a&gt;). PQ-only removes that safety net.&lt;/p&gt;
&lt;p&gt;Two newer ideas aim to reduce the cost of all this. A &lt;strong&gt;composite key exchange&lt;/strong&gt; carries a PQ and a traditional KEM in one exchange over a reliable transport (&lt;code&gt;draft-reddy-ipsecme-ikev2-hybrid-reliable&lt;/code&gt;, individual). &lt;strong&gt;Downgrade prevention&lt;/strong&gt; is being made generic in &lt;code&gt;draft-ietf-ipsecme-ikev2-downgrade-prevention&lt;/code&gt;, which is in the RFC Editor queue.&lt;/p&gt;
&lt;hr/&gt;
&lt;h2 id=&quot;4-the-ipsecme-work&quot;&gt;4. The ipsecme work&lt;/h2&gt;
&lt;p&gt;The &lt;strong&gt;IP Security Maintenance and Extensions&lt;/strong&gt; (ipsecme) working group maintains IKEv2 and ESP and specifies how PQC is used in them. Certificates (lamps) and general PQC guidance (pquip) are covered in &lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/#5-ietf-work-beyond-ipsec-lamps-and-pquip&quot;&gt;part 1&lt;/a&gt;.&lt;/p&gt;






































































&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Document&lt;/th&gt;&lt;th&gt;Subject&lt;/th&gt;&lt;th&gt;Status (Nov 2026)&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;RFC 8784&lt;/td&gt;&lt;td&gt;Mixing PPKs into IKEv2&lt;/td&gt;&lt;td&gt;Proposed Standard (Jun 2020)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;RFC 9242&lt;/td&gt;&lt;td&gt;IKE_INTERMEDIATE exchange&lt;/td&gt;&lt;td&gt;Proposed Standard (May 2022)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;RFC 9370&lt;/td&gt;&lt;td&gt;Multiple key exchanges in IKEv2&lt;/td&gt;&lt;td&gt;Proposed Standard (May 2023)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;RFC 9395&lt;/td&gt;&lt;td&gt;Deprecation of IKEv1 and obsolete algorithms&lt;/td&gt;&lt;td&gt;Proposed Standard (Apr 2023)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;RFC 9867&lt;/td&gt;&lt;td&gt;PPKs in IKE_INTERMEDIATE and CREATE_CHILD_SA&lt;/td&gt;&lt;td&gt;Proposed Standard (Nov 2025)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;code&gt;draft-ietf-ipsecme-ikev2-mlkem-09&lt;/code&gt;&lt;/td&gt;&lt;td&gt;ML-KEM key exchange in IKEv2 (alone or as additional exchange)&lt;/td&gt;&lt;td&gt;&lt;strong&gt;RFC Editor queue&lt;/strong&gt;, intended Proposed Standard; code points 35–37 assigned&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;code&gt;draft-ietf-ipsecme-ikev2-pqc-auth-12&lt;/code&gt;&lt;/td&gt;&lt;td&gt;ML-DSA and SLH-DSA authentication in IKEv2&lt;/td&gt;&lt;td&gt;&lt;strong&gt;RFC Editor queue&lt;/strong&gt;, intended Proposed Standard&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;code&gt;draft-ietf-ipsecme-ikev2-downgrade-prevention-08&lt;/code&gt;&lt;/td&gt;&lt;td&gt;Protection against downgrade attacks&lt;/td&gt;&lt;td&gt;RFC Editor queue&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;code&gt;draft-ietf-ipsecme-ikev2-reliable-transport-08&lt;/code&gt;&lt;/td&gt;&lt;td&gt;Separate transports for IKE and ESP&lt;/td&gt;&lt;td&gt;Submitted to IESG&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;code&gt;draft-ietf-ipsecme-pq-hybrid-auth-01&lt;/code&gt;&lt;/td&gt;&lt;td&gt;Composite signatures (RFC 10073) in IKEv2&lt;/td&gt;&lt;td&gt;WG document&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;code&gt;draft-ietf-ipsecme-hybrid-kem-ikev2-frodo-03&lt;/code&gt;&lt;/td&gt;&lt;td&gt;FrodoKEM key exchange in IKEv2&lt;/td&gt;&lt;td&gt;WG document&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;code&gt;draft-guthrie-cnsa2-ipsec-profile-05&lt;/code&gt;&lt;/td&gt;&lt;td&gt;CNSA 2.0 profile for IPsec (would obsolete RFC 9206)&lt;/td&gt;&lt;td&gt;Individual, intended Informational&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;p&gt;Key details of the two most important drafts:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;ML-KEM&lt;/strong&gt; (&lt;code&gt;draft-ietf-ipsecme-ikev2-mlkem-09&lt;/code&gt;): the KE payload carries the ML-KEM encapsulation key (initiator) or ciphertext (responder). A fresh key pair is required for every exchange, and keys must not be reused. Recipients must perform the input checks from FIPS 203 (encapsulation key check, ciphertext type check). The 32-byte shared secret is used just like the (EC)DH shared secret.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;PQC authentication&lt;/strong&gt; (&lt;code&gt;draft-ietf-ipsecme-ikev2-pqc-auth-12&lt;/code&gt;): uses the RFC 7427 Digital Signature method, with ML-DSA-44/65/87 and all twelve SLH-DSA parameter sets. Only &lt;em&gt;pure&lt;/em&gt; mode is used (no HashML-DSA). Peers must advertise the “Identity” hash in SIGNATURE_HASH_ALGORITHMS, and IKEv2 fragmentation is mandatory because of the signature sizes. Hedged signing is preferred over deterministic. Whether PQ authentication is &lt;em&gt;required&lt;/em&gt; is local policy; if it is optional, a downgrade is possible.&lt;/li&gt;
&lt;/ul&gt;
&lt;hr/&gt;
&lt;h2 id=&quot;5-algorithm-overview&quot;&gt;5. Algorithm overview&lt;/h2&gt;
&lt;p&gt;The table classifies the algorithms you will meet in IKEv2/IPsec by role. Security categories and sizes per algorithm are in &lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/#4-security-categories-parameter-sets-and-sizes&quot;&gt;part 1&lt;/a&gt;. Labels:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;T&lt;/strong&gt;: &lt;em&gt;traditional&lt;/em&gt;, a quantum-vulnerable asymmetric algorithm (RFC 9794);&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;PQ&lt;/strong&gt;: &lt;em&gt;post-quantum&lt;/em&gt;, an asymmetric algorithm designed to resist quantum attacks;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;QR&lt;/strong&gt;: &lt;em&gt;quantum-resistant&lt;/em&gt;, a symmetric or hash-based primitive that remains secure against known quantum attacks at the listed parameters;&lt;/li&gt;
&lt;li&gt;”–” means the algorithm does not play that role.&lt;/li&gt;
&lt;/ul&gt;





















































































































































































&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Algorithm (IANA ID)&lt;/th&gt;&lt;th&gt;Key exchange&lt;/th&gt;&lt;th&gt;PRF / key derivation&lt;/th&gt;&lt;th&gt;Integrity&lt;/th&gt;&lt;th&gt;Authentication&lt;/th&gt;&lt;th&gt;Encryption&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;MODP DH groups 14–18&lt;/td&gt;&lt;td&gt;T&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ECDH P-256 / P-384 / P-521 (19/20/21)&lt;/td&gt;&lt;td&gt;T&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;brainpoolP256r1 / P384r1 / P512r1 (28/29/30)&lt;/td&gt;&lt;td&gt;T&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Curve25519 / Curve448 (31/32)&lt;/td&gt;&lt;td&gt;T&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ML-KEM-512 / -768 / -1024 (35/36/37)&lt;/td&gt;&lt;td&gt;PQ&lt;sup&gt;&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fn-mlkem&quot; id=&quot;user-content-fnref-mlkem&quot; data-footnote-ref aria-describedby=&quot;footnote-label&quot;&gt;1&lt;/a&gt;&lt;/sup&gt;&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;FrodoKEM&lt;/td&gt;&lt;td&gt;PQ&lt;sup&gt;&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fn-frodo&quot; id=&quot;user-content-fnref-frodo&quot; data-footnote-ref aria-describedby=&quot;footnote-label&quot;&gt;2&lt;/a&gt;&lt;/sup&gt;&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;HQC&lt;/td&gt;&lt;td&gt;PQ&lt;sup&gt;&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fn-hqc&quot; id=&quot;user-content-fnref-hqc&quot; data-footnote-ref aria-describedby=&quot;footnote-label&quot;&gt;3&lt;/a&gt;&lt;/sup&gt;&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;PPK (RFC 8784, RFC 9867)&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;QR&lt;sup&gt;&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fn-ppk&quot; id=&quot;user-content-fnref-ppk&quot; data-footnote-ref aria-describedby=&quot;footnote-label&quot;&gt;4&lt;/a&gt;&lt;/sup&gt;&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;HMAC-SHA2-256 / -384 / -512 PRF (5/6/7)&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;QR&lt;sup&gt;&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fn-hmac&quot; id=&quot;user-content-fnref-hmac&quot; data-footnote-ref aria-describedby=&quot;footnote-label&quot;&gt;5&lt;/a&gt;&lt;/sup&gt;&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;HMAC-SHA2-256-128 / -384-192 / -512-256 (12/13/14)&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;QR&lt;sup&gt;&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fn-hmac&quot; id=&quot;user-content-fnref-hmac-2&quot; data-footnote-ref aria-describedby=&quot;footnote-label&quot;&gt;5&lt;/a&gt;&lt;/sup&gt;&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;AES-CBC-128 / -256 (12)&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;QR&lt;sup&gt;&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fn-aes&quot; id=&quot;user-content-fnref-aes&quot; data-footnote-ref aria-describedby=&quot;footnote-label&quot;&gt;6&lt;/a&gt;&lt;/sup&gt;&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;AES-GCM-128 / -256 (20)&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;QR&lt;sup&gt;&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fn-aes&quot; id=&quot;user-content-fnref-aes-2&quot; data-footnote-ref aria-describedby=&quot;footnote-label&quot;&gt;6&lt;/a&gt;&lt;/sup&gt;&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;QR&lt;sup&gt;&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fn-aes&quot; id=&quot;user-content-fnref-aes-3&quot; data-footnote-ref aria-describedby=&quot;footnote-label&quot;&gt;6&lt;/a&gt;&lt;/sup&gt;&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ChaCha20-Poly1305 (28)&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;QR&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;QR&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;RSA (PKCS#1 v1.5, PSS)&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;T&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ECDSA P-256 / P-384 / P-521&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;T&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Ed25519 / Ed448&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;T&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ML-DSA-44 / -65 / -87&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;PQ&lt;sup&gt;&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fn-pqauth&quot; id=&quot;user-content-fnref-pqauth&quot; data-footnote-ref aria-describedby=&quot;footnote-label&quot;&gt;7&lt;/a&gt;&lt;/sup&gt;&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;SLH-DSA (12 parameter sets)&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;PQ&lt;sup&gt;&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fn-pqauth&quot; id=&quot;user-content-fnref-pqauth-2&quot; data-footnote-ref aria-describedby=&quot;footnote-label&quot;&gt;7&lt;/a&gt;&lt;/sup&gt;&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Composite ML-DSA (RFC 10073)&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;T + PQ&lt;sup&gt;&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fn-composite&quot; id=&quot;user-content-fnref-composite&quot; data-footnote-ref aria-describedby=&quot;footnote-label&quot;&gt;8&lt;/a&gt;&lt;/sup&gt;&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;FN-DSA&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;PQ&lt;sup&gt;&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fn-fndsa&quot; id=&quot;user-content-fnref-fndsa&quot; data-footnote-ref aria-describedby=&quot;footnote-label&quot;&gt;9&lt;/a&gt;&lt;/sup&gt;&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;PSK authentication (RFC 7296)&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;QR&lt;sup&gt;&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fn-psk&quot; id=&quot;user-content-fnref-psk&quot; data-footnote-ref aria-describedby=&quot;footnote-label&quot;&gt;10&lt;/a&gt;&lt;/sup&gt;&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;hr/&gt;
&lt;h2 id=&quot;6-key-exchange-and-signature-sizes-in-ikev2&quot;&gt;6. Key exchange and signature sizes in IKEv2&lt;/h2&gt;
&lt;p&gt;PQ keys and signatures are one to three orders of magnitude larger than their traditional counterparts (&lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/#43-key-and-signature-sizes&quot;&gt;full size table in part 1&lt;/a&gt;). That size difference is the main engineering challenge in IKEv2, because IKE runs over UDP and IKE_SA_INIT cannot be fragmented (&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#32-assumption-2-ike_sa_init-messages-are-small&quot;&gt;section 3.2&lt;/a&gt;).&lt;/p&gt;
 &lt;figure class=&quot;viz&quot;&gt; &lt;figcaption&gt;Size of the initiator&apos;s key exchange data vs. packet size limits&lt;/figcaption&gt; &lt;p class=&quot;viz-sub&quot;&gt;Bytes, logarithmic scale. The rest of IKE_SA_INIT (headers, SA proposal, nonce, NAT detection) comes on top of this. Hover a bar for details.&lt;/p&gt; &lt;div class=&quot;legend&quot;&gt; &lt;span style=&quot;--sw: var(--series-1)&quot;&gt;Traditional (T)&lt;/span&gt; &lt;span style=&quot;--sw: var(--series-2)&quot;&gt;Post-quantum (PQ)&lt;/span&gt; &lt;/div&gt; &lt;svg viewBox=&quot;0 0 680 316&quot; role=&quot;img&quot; aria-label=&quot;Bar chart on a log scale of key exchange payload sizes against 1280 and 1500 byte MTU lines&quot;&gt; &lt;g&gt; &lt;line x1=&quot;120&quot; x2=&quot;120&quot; y1=&quot;16&quot; y2=&quot;292&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;120&quot; y=&quot;308&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;16&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;210.90909090909088&quot; x2=&quot;210.90909090909088&quot; y1=&quot;16&quot; y2=&quot;292&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;210.90909090909088&quot; y=&quot;308&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;64&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;301.8181818181818&quot; x2=&quot;301.8181818181818&quot; y1=&quot;16&quot; y2=&quot;292&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;301.8181818181818&quot; y=&quot;308&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;256&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;392.72727272727275&quot; x2=&quot;392.72727272727275&quot; y1=&quot;16&quot; y2=&quot;292&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;392.72727272727275&quot; y=&quot;308&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;1,024&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;483.63636363636357&quot; x2=&quot;483.63636363636357&quot; y1=&quot;16&quot; y2=&quot;292&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;483.63636363636357&quot; y=&quot;308&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;4,096&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;574.5454545454545&quot; x2=&quot;574.5454545454545&quot; y1=&quot;16&quot; y2=&quot;292&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;574.5454545454545&quot; y=&quot;308&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;16,384&lt;/text&gt; &lt;/g&gt; &lt;g&gt; &lt;line x1=&quot;407.3603679494256&quot; x2=&quot;407.3603679494256&quot; y1=&quot;8&quot; y2=&quot;292&quot; stroke=&quot;var(--text-primary)&quot; stroke-width=&quot;1.5&quot; stroke-dasharray=&quot;4 3&quot;&gt;&lt;/line&gt; &lt;text x=&quot;403.3603679494256&quot; y=&quot;12&quot; text-anchor=&quot;end&quot; font-size=&quot;10&quot; class=&quot;t-primary&quot;&gt;1280: IPv6 minimum MTU&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;417.76121751742016&quot; x2=&quot;417.76121751742016&quot; y1=&quot;8&quot; y2=&quot;292&quot; stroke=&quot;var(--text-primary)&quot; stroke-width=&quot;1.5&quot; stroke-dasharray=&quot;4 3&quot;&gt;&lt;/line&gt; &lt;text x=&quot;421.76121751742016&quot; y=&quot;12&quot; text-anchor=&quot;start&quot; font-size=&quot;10&quot; class=&quot;t-primary&quot;&gt;1500: Ethernet MTU&lt;/text&gt; &lt;/g&gt; &lt;g&gt; &lt;text x=&quot;112&quot; y=&quot;34&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;X25519&lt;/text&gt; &lt;rect x=&quot;120&quot; y=&quot;22&quot; width=&quot;45.45454545454547&quot; height=&quot;14&quot; rx=&quot;3&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;X25519: 32 bytes
Fits in an unfragmented IKE_SA_INIT&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;169.45454545454547&quot; y=&quot;33&quot; font-size=&quot;10&quot;&gt;32&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;112&quot; y=&quot;56&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;ECP P-256&lt;/text&gt; &lt;rect x=&quot;120&quot; y=&quot;44&quot; width=&quot;90.90909090909088&quot; height=&quot;14&quot; rx=&quot;3&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;ECP P-256: 64 bytes
Fits in an unfragmented IKE_SA_INIT&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;214.90909090909088&quot; y=&quot;55&quot; font-size=&quot;10&quot;&gt;64&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;112&quot; y=&quot;78&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;ECP P-384&lt;/text&gt; &lt;rect x=&quot;120&quot; y=&quot;66&quot; width=&quot;117.4982954873253&quot; height=&quot;14&quot; rx=&quot;3&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;ECP P-384: 96 bytes
Fits in an unfragmented IKE_SA_INIT&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;241.4982954873253&quot; y=&quot;77&quot; font-size=&quot;10&quot;&gt;96&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;112&quot; y=&quot;100&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;ECP P-521&lt;/text&gt; &lt;rect x=&quot;120&quot; y=&quot;88&quot; width=&quot;138.3815508799297&quot; height=&quot;14&quot; rx=&quot;3&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;ECP P-521: 132 bytes
Fits in an unfragmented IKE_SA_INIT&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;262.3815508799297&quot; y=&quot;99&quot; font-size=&quot;10&quot;&gt;132&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;112&quot; y=&quot;122&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;MODP-3072&lt;/text&gt; &lt;rect x=&quot;120&quot; y=&quot;110&quot; width=&quot;208.4073863964162&quot; height=&quot;14&quot; rx=&quot;3&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;MODP-3072: 384 bytes
Fits in an unfragmented IKE_SA_INIT&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;332.4073863964162&quot; y=&quot;121&quot; font-size=&quot;10&quot;&gt;384&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;112&quot; y=&quot;144&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;MODP-4096&lt;/text&gt; &lt;rect x=&quot;120&quot; y=&quot;132&quot; width=&quot;227.27272727272725&quot; height=&quot;14&quot; rx=&quot;3&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;MODP-4096: 512 bytes
Fits in an unfragmented IKE_SA_INIT&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;351.27272727272725&quot; y=&quot;143&quot; font-size=&quot;10&quot;&gt;512&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;112&quot; y=&quot;166&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;ML-KEM-512&lt;/text&gt; &lt;rect x=&quot;120&quot; y=&quot;154&quot; width=&quot;256.538917717033&quot; height=&quot;14&quot; rx=&quot;3&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;ML-KEM-512: 800 bytes
Fits in an unfragmented IKE_SA_INIT&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;380.538917717033&quot; y=&quot;165&quot; font-size=&quot;10&quot;&gt;800&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;112&quot; y=&quot;188&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;ML-KEM-768&lt;/text&gt; &lt;rect x=&quot;120&quot; y=&quot;176&quot; width=&quot;282.24788025586133&quot; height=&quot;14&quot; rx=&quot;3&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;ML-KEM-768: 1,184 bytes
Too large for a safe unfragmented IKE_SA_INIT over UDP: carry it in IKE_INTERMEDIATE (RFC 9242) or use a reliable transport&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;406.24788025586133&quot; y=&quot;187&quot; font-size=&quot;10&quot;&gt;1,184&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;112&quot; y=&quot;210&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;ML-KEM-1024&lt;/text&gt; &lt;rect x=&quot;120&quot; y=&quot;198&quot; width=&quot;300.66862927796404&quot; height=&quot;14&quot; rx=&quot;3&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;ML-KEM-1024: 1,568 bytes
Too large for a safe unfragmented IKE_SA_INIT over UDP: carry it in IKE_INTERMEDIATE (RFC 9242) or use a reliable transport&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;424.66862927796404&quot; y=&quot;209&quot; font-size=&quot;10&quot;&gt;1,568&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;112&quot; y=&quot;232&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;FrodoKEM-640&lt;/text&gt; &lt;rect x=&quot;120&quot; y=&quot;220&quot; width=&quot;419.60096275959927&quot; height=&quot;14&quot; rx=&quot;3&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;FrodoKEM-640: 9,616 bytes
Too large for a safe unfragmented IKE_SA_INIT over UDP: carry it in IKE_INTERMEDIATE (RFC 9242) or use a reliable transport&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;543.6009627595993&quot; y=&quot;231&quot; font-size=&quot;10&quot;&gt;9,616&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;112&quot; y=&quot;254&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;FrodoKEM-976&lt;/text&gt; &lt;rect x=&quot;120&quot; y=&quot;242&quot; width=&quot;451.46430690765385&quot; height=&quot;14&quot; rx=&quot;3&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;FrodoKEM-976: 15,632 bytes
Too large for a safe unfragmented IKE_SA_INIT over UDP: carry it in IKE_INTERMEDIATE (RFC 9242) or use a reliable transport&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;575.4643069076538&quot; y=&quot;253&quot; font-size=&quot;10&quot;&gt;15,632&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;112&quot; y=&quot;276&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;FrodoKEM-1344&lt;/text&gt; &lt;rect x=&quot;120&quot; y=&quot;264&quot; width=&quot;472.4268389747267&quot; height=&quot;14&quot; rx=&quot;3&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;FrodoKEM-1344: 21,520 bytes
Too large for a safe unfragmented IKE_SA_INIT over UDP: carry it in IKE_INTERMEDIATE (RFC 9242) or use a reliable transport&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;596.4268389747267&quot; y=&quot;275&quot; font-size=&quot;10&quot;&gt;21,520&lt;/text&gt; &lt;/g&gt; &lt;/svg&gt; &lt;details&gt; &lt;summary&gt;Show data as table&lt;/summary&gt; &lt;table&gt; &lt;thead&gt;&lt;tr&gt;&lt;th&gt;Key exchange&lt;/th&gt;&lt;th&gt;Type&lt;/th&gt;&lt;th&gt;Initiator KE data (bytes)&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt; &lt;tbody&gt;&lt;tr&gt;&lt;td&gt;X25519&lt;/td&gt;&lt;td&gt;T&lt;/td&gt;&lt;td&gt;32&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ECP P-256&lt;/td&gt;&lt;td&gt;T&lt;/td&gt;&lt;td&gt;64&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ECP P-384&lt;/td&gt;&lt;td&gt;T&lt;/td&gt;&lt;td&gt;96&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ECP P-521&lt;/td&gt;&lt;td&gt;T&lt;/td&gt;&lt;td&gt;132&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;MODP-3072&lt;/td&gt;&lt;td&gt;T&lt;/td&gt;&lt;td&gt;384&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;MODP-4096&lt;/td&gt;&lt;td&gt;T&lt;/td&gt;&lt;td&gt;512&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ML-KEM-512&lt;/td&gt;&lt;td&gt;PQ&lt;/td&gt;&lt;td&gt;800&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ML-KEM-768&lt;/td&gt;&lt;td&gt;PQ&lt;/td&gt;&lt;td&gt;1,184&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ML-KEM-1024&lt;/td&gt;&lt;td&gt;PQ&lt;/td&gt;&lt;td&gt;1,568&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;FrodoKEM-640&lt;/td&gt;&lt;td&gt;PQ&lt;/td&gt;&lt;td&gt;9,616&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;FrodoKEM-976&lt;/td&gt;&lt;td&gt;PQ&lt;/td&gt;&lt;td&gt;15,632&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;FrodoKEM-1344&lt;/td&gt;&lt;td&gt;PQ&lt;/td&gt;&lt;td&gt;21,520&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt; &lt;/table&gt; &lt;/details&gt; &lt;/figure&gt;

























































































&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Item&lt;/th&gt;&lt;th&gt;Size (bytes)&lt;/th&gt;&lt;th&gt;Carried in&lt;/th&gt;&lt;th&gt;Fits an unfragmented IKE_SA_INIT over UDP?&lt;/th&gt;&lt;th&gt;What makes it work&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;X25519 / ECP P-384 / ECP P-521&lt;/td&gt;&lt;td&gt;32 / 96 / 132&lt;/td&gt;&lt;td&gt;IKE_SA_INIT&lt;/td&gt;&lt;td&gt;Yes&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;MODP-3072 / MODP-4096&lt;/td&gt;&lt;td&gt;384 / 512&lt;/td&gt;&lt;td&gt;IKE_SA_INIT&lt;/td&gt;&lt;td&gt;Yes&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ML-KEM-512 (initiator / responder)&lt;/td&gt;&lt;td&gt;800 / 768&lt;/td&gt;&lt;td&gt;IKE_SA_INIT or IKE_INTERMEDIATE&lt;/td&gt;&lt;td&gt;Yes (the ML-KEM draft allows it)&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ML-KEM-768&lt;/td&gt;&lt;td&gt;1,184 / 1,088&lt;/td&gt;&lt;td&gt;IKE_INTERMEDIATE&lt;/td&gt;&lt;td&gt;&lt;strong&gt;No&lt;/strong&gt;, unless the path MTU is known to be large&lt;/td&gt;&lt;td&gt;RFC 9242 + RFC 9370 + RFC 7383, or TCP (RFC 9329)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ML-KEM-1024&lt;/td&gt;&lt;td&gt;1,568 / 1,568&lt;/td&gt;&lt;td&gt;IKE_INTERMEDIATE&lt;/td&gt;&lt;td&gt;&lt;strong&gt;No&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;RFC 9242 + RFC 9370 + RFC 7383, or TCP&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;FrodoKEM-640 … -1344&lt;/td&gt;&lt;td&gt;9,616 … 21,520 / 9,720 … 21,632&lt;/td&gt;&lt;td&gt;IKE_INTERMEDIATE&lt;/td&gt;&lt;td&gt;&lt;strong&gt;No&lt;/strong&gt; (several fragments)&lt;/td&gt;&lt;td&gt;RFC 9242 + RFC 9370 + RFC 7383&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ECDSA P-256 / Ed25519 signature&lt;/td&gt;&lt;td&gt;≈ 64–72 / 64&lt;/td&gt;&lt;td&gt;IKE_AUTH&lt;/td&gt;&lt;td&gt;n/a (IKE_AUTH is encrypted)&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ECDSA P-384 / P-521 signature&lt;/td&gt;&lt;td&gt;≈ 96–104 / ≈ 132–139&lt;/td&gt;&lt;td&gt;IKE_AUTH&lt;/td&gt;&lt;td&gt;n/a&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;RSA-3072 / RSA-4096 signature&lt;/td&gt;&lt;td&gt;384 / 512&lt;/td&gt;&lt;td&gt;IKE_AUTH&lt;/td&gt;&lt;td&gt;n/a&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ML-DSA-44 / -65 / -87 signature (public key)&lt;/td&gt;&lt;td&gt;2,420 / 3,309 / 4,627 (1,312 / 1,952 / 2,592)&lt;/td&gt;&lt;td&gt;IKE_AUTH (+ certificates)&lt;/td&gt;&lt;td&gt;n/a&lt;/td&gt;&lt;td&gt;RFC 7383 fragmentation (mandatory in the PQC-auth draft)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;SLH-DSA-128s … -256f signature&lt;/td&gt;&lt;td&gt;7,856 … 49,856&lt;/td&gt;&lt;td&gt;IKE_AUTH&lt;/td&gt;&lt;td&gt;n/a&lt;/td&gt;&lt;td&gt;RFC 7383 fragmentation&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;p&gt;Sizes are the raw key exchange data and signatures. The KE payload adds an 8-byte header, and IKE_SA_INIT also carries headers, the SA proposal, nonces and notifications. Each certificate in IKE_AUTH carries a public key &lt;strong&gt;and&lt;/strong&gt; a CA signature, so a PQ certificate chain easily reaches 10 KB or more.&lt;/p&gt;
&lt;hr/&gt;
&lt;h2 id=&quot;7-cnsa-20-and-ipsec&quot;&gt;7. CNSA 2.0 and IPsec&lt;/h2&gt;
&lt;p&gt;The NSA algorithm suites, their timeline and the comparison with NIST SP 800-227 are in &lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/#24-nsa-suite-b-cnsa-10-and-cnsa-20&quot;&gt;part 1&lt;/a&gt;. For IPsec specifically:&lt;/p&gt;

























&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Suite&lt;/th&gt;&lt;th&gt;IPsec profile&lt;/th&gt;&lt;th&gt;Key exchange in IKEv2&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;Suite B (2005)&lt;/td&gt;&lt;td&gt;RFC 6379 / RFC 6380 (2011), &lt;strong&gt;Historic&lt;/strong&gt; since RFC 8423 (2018)&lt;/td&gt;&lt;td&gt;ECDH P-256 or P-384&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;CNSA 1.0 (2015/16)&lt;/td&gt;&lt;td&gt;&lt;strong&gt;RFC 9206&lt;/strong&gt; (2022, Independent Submission)&lt;/td&gt;&lt;td&gt;ECDH P-384, DH-3072 or DH-4096&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;CNSA 2.0 (2022)&lt;/td&gt;&lt;td&gt;&lt;code&gt;draft-guthrie-cnsa2-ipsec-profile-05&lt;/code&gt; (Oct 2026, individual, would obsolete RFC 9206)&lt;/td&gt;&lt;td&gt;CNSA 1.0 key exchange &lt;strong&gt;plus ML-KEM-1024&lt;/strong&gt; as additional key exchange&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;p&gt;RFC 9206 defines three CNSA 1.0 IPsec suites: CNSA-GCM-256-ECDH-384, CNSA-GCM-256-DH-3072 and CNSA-GCM-256-DH-4096. They all use IKEv2, AES-256-GCM, and PRF_HMAC_SHA_512 (SHA-512 because of availability, although SHA-384 is the CNSA hash).&lt;/p&gt;
&lt;p&gt;In its CNSA 2.0 FAQ, NSA notes that ML-KEM-1024 cannot simply replace the CNSA 1.0 key exchange in IKEv2 because of message size (&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#6-key-exchange-and-signature-sizes-in-ikev2&quot;&gt;section 6&lt;/a&gt;). It expects to use the IETF approach: a smaller initial key exchange followed by a larger, encrypted one. Concretely, CNSA 1.0 key establishment in IKE_SA_INIT is fortified by ML-KEM-1024 in an additional exchange (RFC 9370). NSA does not &lt;em&gt;require&lt;/em&gt; hybrids. It rejects non-standard hybrids on NSS, but it does mention PPKs (RFC 8784) as possibly appropriate for specialised cases.&lt;/p&gt;
&lt;p&gt;For VPN products, the CNSA 2.0 category table sets &lt;strong&gt;support and prefer by 2026, exclusive use by 2030&lt;/strong&gt;, and from &lt;strong&gt;1 January 2027&lt;/strong&gt; all new acquisitions for US national security systems must be CNSA 2.0-compliant (&lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/#24-nsa-suite-b-cnsa-10-and-cnsa-20&quot;&gt;timeline in part 1&lt;/a&gt;).&lt;/p&gt;
&lt;hr/&gt;
&lt;h2 id=&quot;8-designing-for-several-regimes&quot;&gt;8. Designing for several regimes&lt;/h2&gt;
&lt;p&gt;European, Australian and US authorities differ on hybrids, parameter sizes and deadlines (&lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/#23-european-and-australian-authorities&quot;&gt;part 1, section 2.3&lt;/a&gt;). A VPN that must satisfy several regimes at once ends up with the &lt;strong&gt;union&lt;/strong&gt; of their requirements:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;PQ/T hybrid key exchange (for BSI and ANSSI);&lt;/li&gt;
&lt;li&gt;category 5 parameters, ML-KEM-1024 (for ASD and CNSA 2.0);&lt;/li&gt;
&lt;li&gt;a completion date of &lt;strong&gt;2030&lt;/strong&gt; for the most demanding customers.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;IKEv2 with RFC 9370 can satisfy all of them in one configuration, for example ECDH P-384 plus ML-KEM-1024. The flexibility to do that is the practical meaning of crypto agility.&lt;/p&gt;
&lt;hr/&gt;
&lt;h2 id=&quot;9-a-practical-migration-path&quot;&gt;9. A practical migration path&lt;/h2&gt;
&lt;ol&gt;
&lt;li&gt;&lt;strong&gt;Retire IKEv1&lt;/strong&gt; (RFC 9395). There is no PQC path for it (&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#14-why-rfc-9395-ikev1-deprecation-matters-for-pqc&quot;&gt;section 1.4&lt;/a&gt;).&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Determine which regimes apply&lt;/strong&gt; (EU/NIS2 sector, BSI, ANSSI, NCSC, ASD, CNSA 2.0; see &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#8-designing-for-several-regimes&quot;&gt;section 8&lt;/a&gt; and &lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/#23-european-and-australian-authorities&quot;&gt;part 1&lt;/a&gt;). They decide hybrid versus pure PQ, the parameter category, and your deadline.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Inventory&lt;/strong&gt; your gateways, clients, PKI and HSMs. Check firmware upgradability and vendor roadmaps for ML-KEM and ML-DSA.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Enable the plumbing&lt;/strong&gt;: IKEv2 fragmentation (RFC 7383), IKE_INTERMEDIATE (RFC 9242) and multiple key exchanges (RFC 9370). Consider TCP encapsulation (RFC 9329) where UDP fragmentation is a problem.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Deploy PQ/T hybrid key exchange now&lt;/strong&gt;: (EC)DH in IKE_SA_INIT plus ML-KEM-768, or ML-KEM-1024 for CNSA 2.0 and ASD, as ADDKE1. This addresses harvest-now-decrypt-later.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Where you control both ends and can distribute keys&lt;/strong&gt;, add a PPK (RFC 8784, or RFC 9867 to also protect the initial IKE SA).&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Strengthen the classical part&lt;/strong&gt;: AES-256-GCM, PRF and integrity with SHA-384 or SHA-512, ECDH groups 20/21/29/30/32.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Prepare PQ authentication&lt;/strong&gt;: an ML-DSA (RFC 9881) or composite ML-DSA (RFC 10073) PKI, with attention to certificate chain sizes. Track &lt;code&gt;draft-ietf-ipsecme-ikev2-pqc-auth&lt;/code&gt; until it becomes an RFC.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Close the downgrade door&lt;/strong&gt;: once all peers support ML-KEM, refuse proposals without it.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Build in agility&lt;/strong&gt;: algorithms in configuration, not code. Expect FN-DSA, HQC and new IETF RFCs in the coming years.&lt;/li&gt;
&lt;/ol&gt;
&lt;hr/&gt;
&lt;h2 id=&quot;10-sources&quot;&gt;10. Sources&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;IETF: RFCs&lt;/strong&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/rfc7296/&quot;&gt;RFC 7296: IKEv2 (STD 79)&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/rfc9395/&quot;&gt;RFC 9395: Deprecation of IKEv1&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.rfc-editor.org/info/rfc7383&quot;&gt;RFC 7383: IKEv2 Message Fragmentation&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.rfc-editor.org/info/rfc9242&quot;&gt;RFC 9242: Intermediate Exchange in IKEv2&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.rfc-editor.org/info/rfc9370&quot;&gt;RFC 9370: Multiple Key Exchanges in IKEv2&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.rfc-editor.org/info/rfc8784&quot;&gt;RFC 8784: Mixing Preshared Keys in IKEv2&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/rfc9867/&quot;&gt;RFC 9867: PPKs in IKE_INTERMEDIATE and CREATE_CHILD_SA&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.rfc-editor.org/info/rfc4754&quot;&gt;RFC 4754: IKE and IKEv2 Authentication Using ECDSA&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.rfc-editor.org/info/rfc7427&quot;&gt;RFC 7427: Signature Authentication in IKEv2&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.rfc-editor.org/info/rfc8420&quot;&gt;RFC 8420: Using EdDSA in IKEv2 (Identity hash)&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.rfc-editor.org/info/rfc9593&quot;&gt;RFC 9593: Announcing Supported Authentication Methods in IKEv2&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.rfc-editor.org/info/rfc9329&quot;&gt;RFC 9329: TCP Encapsulation of IKE and IPsec Packets&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://rfc-editor.org/info/rfc8247&quot;&gt;RFC 8247: Algorithm requirements for IKEv2&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.rfc-editor.org/info/rfc7696&quot;&gt;RFC 7696: Guidelines for Cryptographic Algorithm Agility (BCP 201)&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.rfc-editor.org/info/rfc2026&quot;&gt;RFC 2026&lt;/a&gt;, &lt;a href=&quot;https://www.rfc-editor.org/info/rfc6410&quot;&gt;RFC 6410&lt;/a&gt;, &lt;a href=&quot;https://www.rfc-editor.org/info/rfc7127&quot;&gt;RFC 7127&lt;/a&gt;: IETF standards process&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.rfc-editor.org/rfc/rfc9794.html&quot;&gt;RFC 9794: Terminology for PQ/T Hybrid Schemes&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/rfc9958/&quot;&gt;RFC 9958: Post-Quantum Cryptography for Engineers&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/rfc9955/&quot;&gt;RFC 9955: Hybrid Signature Spectrums&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.rfc-editor.org/info/rfc9881/&quot;&gt;RFC 9881: ML-DSA in X.509&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.rfc-editor.org/rfc/rfc9909.html&quot;&gt;RFC 9909: SLH-DSA in X.509&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/draft-ietf-lamps-pq-composite-sigs/&quot;&gt;RFC 10073: Composite ML-DSA&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.rfc-editor.org/info/rfc9206/&quot;&gt;RFC 9206: CNSA Suite for IPsec&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://rfc-editor.org/info/rfc6379&quot;&gt;RFC 6379&lt;/a&gt; and &lt;a href=&quot;https://www.rfc-editor.org/info/rfc6380&quot;&gt;RFC 6380&lt;/a&gt;: Suite B for IPsec (Historic)&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/rfc8423/&quot;&gt;RFC 8423: Reclassification of Suite B documents to Historic&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;&lt;strong&gt;IETF: Internet-Drafts and working groups&lt;/strong&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;ipsecme, in the RFC Editor queue or with the IESG:
&lt;ul&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/draft-ietf-ipsecme-ikev2-mlkem/&quot;&gt;draft-ietf-ipsecme-ikev2-mlkem&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/draft-ietf-ipsecme-ikev2-pqc-auth/&quot;&gt;draft-ietf-ipsecme-ikev2-pqc-auth&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/draft-ietf-ipsecme-ikev2-downgrade-prevention/&quot;&gt;draft-ietf-ipsecme-ikev2-downgrade-prevention&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/draft-ietf-ipsecme-ikev2-reliable-transport/&quot;&gt;draft-ietf-ipsecme-ikev2-reliable-transport&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/li&gt;
&lt;li&gt;ipsecme, WG documents:
&lt;ul&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/html/draft-ietf-ipsecme-pq-hybrid-auth-01&quot;&gt;draft-ietf-ipsecme-pq-hybrid-auth&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/draft-ietf-ipsecme-hybrid-kem-ikev2-frodo/&quot;&gt;draft-ietf-ipsecme-hybrid-kem-ikev2-frodo&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/li&gt;
&lt;li&gt;pquip, active with the IESG:
&lt;ul&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/draft-ietf-pquip-hbs-state/&quot;&gt;draft-ietf-pquip-hbs-state&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/draft-ietf-pquip-pqc-hsm-constrained/&quot;&gt;draft-ietf-pquip-pqc-hsm-constrained&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/li&gt;
&lt;li&gt;Individual drafts:
&lt;ul&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/html/draft-guthrie-cnsa2-ipsec-profile/&quot;&gt;draft-guthrie-cnsa2-ipsec-profile&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/draft-reddy-ipsecme-ikev2-hqc/&quot;&gt;draft-reddy-ipsecme-ikev2-hqc&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/draft-reddy-ipsecme-ikev2-hybrid-reliable/&quot;&gt;draft-reddy-ipsecme-ikev2-hybrid-reliable&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/draft-hu-ipsecme-pqt-hybrid-auth-with-sep-cert/&quot;&gt;draft-hu-ipsecme-pqt-hybrid-auth-with-sep-cert&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/draft-wang-ipsecme-multi-auth-ikev2-pq/&quot;&gt;draft-wang-ipsecme-multi-auth-ikev2-pq&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/draft-fu-ipsecme-ikev2-fn-dsa-pkr/&quot;&gt;draft-fu-ipsecme-ikev2-fn-dsa-pkr&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/draft-prabel-pquip-pqc-overview/&quot;&gt;draft-prabel-pquip-pqc-overview&lt;/a&gt; (expired)&lt;/li&gt;
&lt;/ul&gt;
&lt;/li&gt;
&lt;li&gt;Working group pages: &lt;a href=&quot;https://datatracker.ietf.org/group/ipsecme/documents/&quot;&gt;ipsecme&lt;/a&gt; · &lt;a href=&quot;https://datatracker.ietf.org/wg/pquip&quot;&gt;pquip&lt;/a&gt; · &lt;a href=&quot;https://datatracker.ietf.org/wg/lamps/documents/&quot;&gt;lamps&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;&lt;strong&gt;IANA&lt;/strong&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;a href=&quot;https://www.iana.org/assignments/ikev2-parameters/ikev2-parameters.xhtml&quot;&gt;IKEv2 Parameters registry&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;&lt;strong&gt;NIST&lt;/strong&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;a href=&quot;https://csrc.nist.gov/projects/post-quantum-cryptography&quot;&gt;NIST Post-Quantum Cryptography project&lt;/a&gt; and &lt;a href=&quot;https://csrc.nist.gov/Projects/post-quantum-cryptography/publications&quot;&gt;publications&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://csrc.nist.gov/pubs/sp/800/227/final&quot;&gt;NIST SP 800-227: Recommendations for Key-Encapsulation Mechanisms&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://csrc.nist.gov/pubs/ir/8547/final&quot;&gt;NIST IR 8547 (final)&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;&lt;strong&gt;NSA&lt;/strong&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;a href=&quot;https://media.defense.gov/2022/Sep/07/2003071836/-1/-1/0/CSI_CNSA_2.0_FAQ_.PDF&quot;&gt;CNSA 2.0 and Quantum Computing FAQ&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;&lt;strong&gt;Background (part 1)&lt;/strong&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/&quot;&gt;Post-Quantum Cryptography: Threat, Standards and Timelines&lt;/a&gt;: policy documents, NIST, EU/ASD/NSA guidance and research sources&lt;/li&gt;
&lt;/ul&gt;
&lt;aside class=&quot;status-note&quot;&gt;&lt;p&gt;&lt;strong&gt;Status note.&lt;/strong&gt; Every RFC, Internet-Draft and IANA entry in this article was checked against the IETF Datatracker, the RFC Editor and IANA in &lt;strong&gt;November 2026&lt;/strong&gt;. Several of the Internet-Drafts mentioned, including the ML-KEM and PQC-authentication drafts, are in their final stage before becoming an RFC (RFC Editor queue or IESG), so their numbers, versions and details can change quickly. Always check the Datatracker for the latest version before you rely on one.&lt;/p&gt;&lt;/aside&gt;
&lt;section data-footnotes class=&quot;footnotes&quot;&gt;&lt;h2 class=&quot;sr-only&quot; id=&quot;footnote-label&quot;&gt;Footnotes&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;user-content-fn-mlkem&quot;&gt;
&lt;p&gt;&lt;code&gt;draft-ietf-ipsecme-ikev2-mlkem-09&lt;/code&gt; (RFC Editor queue). Code points assigned. ML-KEM-768 and -1024 should run in IKE_INTERMEDIATE or IKE_FOLLOWUP_KE (RFC 9370) unless the path MTU is known or a reliable transport is used. &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fnref-mlkem&quot; data-footnote-backref aria-label=&quot;Back to reference 1&quot; class=&quot;data-footnote-backref&quot;&gt;↩&lt;/a&gt;&lt;/p&gt;
&lt;/li&gt;
&lt;li id=&quot;user-content-fn-frodo&quot;&gt;
&lt;p&gt;Conservative unstructured-lattice KEM, not a NIST standard. Specified for IKEv2 in &lt;code&gt;draft-ietf-ipsecme-hybrid-kem-ikev2-frodo-03&lt;/code&gt; (WG document). No IANA code point yet. &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fnref-frodo&quot; data-footnote-backref aria-label=&quot;Back to reference 2&quot; class=&quot;data-footnote-backref&quot;&gt;↩&lt;/a&gt;&lt;/p&gt;
&lt;/li&gt;
&lt;li id=&quot;user-content-fn-hqc&quot;&gt;
&lt;p&gt;Code-based KEM selected by NIST in March 2025 as a backup to ML-KEM. The FIPS (expected FIPS 207) is not yet final. IKEv2 usage is only an individual draft (&lt;code&gt;draft-reddy-ipsecme-ikev2-hqc-00&lt;/code&gt;). &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fnref-hqc&quot; data-footnote-backref aria-label=&quot;Back to reference 3&quot; class=&quot;data-footnote-backref&quot;&gt;↩&lt;/a&gt;&lt;/p&gt;
&lt;/li&gt;
&lt;li id=&quot;user-content-fn-ppk&quot;&gt;
&lt;p&gt;Not a key exchange but a symmetric secret mixed into key derivation. It is only as quantum-resistant as the PPK’s entropy and out-of-band distribution. RFC 8784 protects Child SAs only; RFC 9867 also protects the initial IKE SA. &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fnref-ppk&quot; data-footnote-backref aria-label=&quot;Back to reference 4&quot; class=&quot;data-footnote-backref&quot;&gt;↩&lt;/a&gt;&lt;/p&gt;
&lt;/li&gt;
&lt;li id=&quot;user-content-fn-hmac&quot;&gt;
&lt;p&gt;HMAC with SHA-2 is not meaningfully weakened by known quantum algorithms. CNSA 2.0 requires SHA-384 or SHA-512. &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fnref-hmac&quot; data-footnote-backref aria-label=&quot;Back to reference 5&quot; class=&quot;data-footnote-backref&quot;&gt;↩&lt;/a&gt; &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fnref-hmac-2&quot; data-footnote-backref aria-label=&quot;Back to reference 5-2&quot; class=&quot;data-footnote-backref&quot;&gt;↩&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/li&gt;
&lt;li id=&quot;user-content-fn-aes&quot;&gt;
&lt;p&gt;Grover’s algorithm offers at most a quadratic speed-up and parallelises poorly. NIST uses AES-128 key search as the benchmark for its lowest PQ security category. Policy profiles such as CNSA 2.0 nevertheless require AES-256. &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fnref-aes&quot; data-footnote-backref aria-label=&quot;Back to reference 6&quot; class=&quot;data-footnote-backref&quot;&gt;↩&lt;/a&gt; &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fnref-aes-2&quot; data-footnote-backref aria-label=&quot;Back to reference 6-2&quot; class=&quot;data-footnote-backref&quot;&gt;↩&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fnref-aes-3&quot; data-footnote-backref aria-label=&quot;Back to reference 6-3&quot; class=&quot;data-footnote-backref&quot;&gt;↩&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/li&gt;
&lt;li id=&quot;user-content-fn-pqauth&quot;&gt;
&lt;p&gt;&lt;code&gt;draft-ietf-ipsecme-ikev2-pqc-auth-12&lt;/code&gt; (RFC Editor queue). Uses RFC 7427 with OIDs from RFC 9881 / RFC 9909, pure mode only, requires the “Identity” hash and IKEv2 fragmentation. &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fnref-pqauth&quot; data-footnote-backref aria-label=&quot;Back to reference 7&quot; class=&quot;data-footnote-backref&quot;&gt;↩&lt;/a&gt; &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fnref-pqauth-2&quot; data-footnote-backref aria-label=&quot;Back to reference 7-2&quot; class=&quot;data-footnote-backref&quot;&gt;↩&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/li&gt;
&lt;li id=&quot;user-content-fn-composite&quot;&gt;
&lt;p&gt;PQ/T hybrid: one composite key and one composite signature. Defined in RFC 10073 (Proposed Standard, Nov 2026). IKEv2 usage in &lt;code&gt;draft-ietf-ipsecme-pq-hybrid-auth-01&lt;/code&gt; (WG document). &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fnref-composite&quot; data-footnote-backref aria-label=&quot;Back to reference 8&quot; class=&quot;data-footnote-backref&quot;&gt;↩&lt;/a&gt;&lt;/p&gt;
&lt;/li&gt;
&lt;li id=&quot;user-content-fn-fndsa&quot;&gt;
&lt;p&gt;Falcon-based. FIPS 206 is not yet final. IKEv2 usage only in an individual draft (&lt;code&gt;draft-fu-ipsecme-ikev2-fn-dsa-pkr-00&lt;/code&gt;). &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fnref-fndsa&quot; data-footnote-backref aria-label=&quot;Back to reference 9&quot; class=&quot;data-footnote-backref&quot;&gt;↩&lt;/a&gt;&lt;/p&gt;
&lt;/li&gt;
&lt;li id=&quot;user-content-fn-psk&quot;&gt;
&lt;p&gt;Symmetric, so not affected by Shor, but only secure with a high-entropy key. Low-entropy (password-like) PSKs are vulnerable to offline dictionary attacks even without a quantum computer. &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#user-content-fnref-psk&quot; data-footnote-backref aria-label=&quot;Back to reference 10&quot; class=&quot;data-footnote-backref&quot;&gt;↩&lt;/a&gt;&lt;/p&gt;
&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;</content:encoded><category>PQC</category><category>IPsec</category><category>VPN</category></item><item><title>Post-Quantum Cryptography: Threat, Standards and Timelines</title><link>https://www.itv6.be/blog/post-quantum-cryptography/</link><guid isPermaLink="true">https://www.itv6.be/blog/post-quantum-cryptography/</guid><description>Why post-quantum cryptography is urgent, who sets the rules (IETF, NIST, EU, ASD, NSA), the key concepts, and the algorithms with their security categories and sizes. Status as of November 2026.</description><pubDate>Sat, 21 Nov 2026 00:00:00 GMT</pubDate><content:encoded>&lt;p&gt;This is part 1 of a two-part series. It covers post-quantum cryptography (PQC) in general: the threat, the organisations that set the rules, the concepts and the algorithms. Part 2, &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/&quot;&gt;&lt;strong&gt;Post-Quantum Cryptography in IPsec&lt;/strong&gt;&lt;/a&gt;, applies all of this to IKEv2 and IPsec VPNs.&lt;/p&gt;
&lt;h2 id=&quot;1-why-pqc-and-how-urgent-is-it&quot;&gt;1. Why PQC, and how urgent is it?&lt;/h2&gt;
&lt;h3 id=&quot;11-the-threat&quot;&gt;1.1 The threat&lt;/h3&gt;
&lt;p&gt;All public-key cryptography in IPsec today rests on two mathematical problems: integer factorisation (RSA) and discrete logarithms in finite fields or on elliptic curves (MODP Diffie-Hellman, ECDH, ECDSA, EdDSA). &lt;strong&gt;Shor’s algorithm&lt;/strong&gt; solves both efficiently on a large, fault-tolerant quantum computer, usually called a &lt;em&gt;cryptographically relevant quantum computer&lt;/em&gt; (CRQC). Once a CRQC exists, every IKEv2 key exchange and signature that uses these algorithms is broken.&lt;/p&gt;
&lt;p&gt;Symmetric cryptography is affected far less. &lt;strong&gt;Grover’s algorithm&lt;/strong&gt; gives at most a quadratic speed-up for brute-force key search, and it parallelises poorly. AES, SHA-2 and HMAC therefore stay usable. Larger parameters (AES-256, SHA-384) give a comfortable margin.&lt;/p&gt;
&lt;h3 id=&quot;12-security-strength-classical-vs-quantum&quot;&gt;1.2 Security strength: classical vs. quantum&lt;/h3&gt;
&lt;p&gt;The difference is easiest to see in bits of security. Against a classical computer, RSA-3072, P-256 and AES-128 all offer about 128 bits. Against a CRQC, RSA and elliptic curves drop to effectively zero, while AES keeps half of its nominal strength at worst.&lt;/p&gt;
 &lt;figure class=&quot;viz&quot;&gt; &lt;figcaption&gt;Security strength: classical computer vs. CRQC&lt;/figcaption&gt; &lt;p class=&quot;viz-sub&quot;&gt;Bits of security (higher is stronger). The solid line is the 128-bit classical baseline. The dashed line is AES-128 under Grover, the floor NIST accepts for Category 1. Hover a bar for details.&lt;/p&gt; &lt;div class=&quot;legend&quot;&gt; &lt;span style=&quot;--sw: var(--series-1)&quot;&gt;Against a classical computer&lt;/span&gt; &lt;span style=&quot;--sw: var(--series-2)&quot;&gt;Against a CRQC&lt;/span&gt; &lt;/div&gt; &lt;svg viewBox=&quot;0 0 680 568&quot; role=&quot;img&quot; aria-label=&quot;Grouped bar chart of security strength in bits, classical versus quantum, for traditional, symmetric and post-quantum algorithms&quot;&gt; &lt;g&gt; &lt;line x1=&quot;150&quot; x2=&quot;150&quot; y1=&quot;14&quot; y2=&quot;534&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;150&quot; y=&quot;548&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;0&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;262.94117647058823&quot; x2=&quot;262.94117647058823&quot; y1=&quot;14&quot; y2=&quot;534&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;262.94117647058823&quot; y=&quot;548&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;64&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;375.88235294117646&quot; x2=&quot;375.88235294117646&quot; y1=&quot;14&quot; y2=&quot;534&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;375.88235294117646&quot; y=&quot;548&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;128&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;488.82352941176475&quot; x2=&quot;488.82352941176475&quot; y1=&quot;14&quot; y2=&quot;534&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;488.82352941176475&quot; y=&quot;548&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;192&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;601.7647058823529&quot; x2=&quot;601.7647058823529&quot; y1=&quot;14&quot; y2=&quot;534&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;601.7647058823529&quot; y=&quot;548&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;256&lt;/text&gt; &lt;/g&gt; &lt;line x1=&quot;375.88235294117646&quot; x2=&quot;375.88235294117646&quot; y1=&quot;6&quot; y2=&quot;534&quot; stroke=&quot;var(--text-primary)&quot; stroke-width=&quot;1.5&quot;&gt;&lt;/line&gt; &lt;text x=&quot;379.88235294117646&quot; y=&quot;12&quot; font-size=&quot;10&quot; class=&quot;t-primary&quot;&gt;128-bit baseline&lt;/text&gt; &lt;line x1=&quot;262.94117647058823&quot; x2=&quot;262.94117647058823&quot; y1=&quot;6&quot; y2=&quot;534&quot; stroke=&quot;var(--text-primary)&quot; stroke-width=&quot;1.2&quot; stroke-dasharray=&quot;4 3&quot;&gt;&lt;/line&gt; &lt;text x=&quot;258.94117647058823&quot; y=&quot;12&quot; font-size=&quot;10&quot; text-anchor=&quot;end&quot; class=&quot;t-primary&quot;&gt;Cat. 1 floor (quantum)&lt;/text&gt; &lt;g&gt; &lt;text x=&quot;0&quot; y=&quot;35&quot; font-size=&quot;11&quot; font-weight=&quot;700&quot; class=&quot;t-primary&quot;&gt;Traditional public-key&lt;/text&gt; &lt;g&gt; &lt;text x=&quot;142&quot; y=&quot;53&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;RSA-2048&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;42&quot; height=&quot;8&quot; width=&quot;197.6470588235294&quot; rx=&quot;2&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;RSA-2048
Classical: 112 bits
&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;351.6470588235294&quot; y=&quot;49&quot; font-size=&quot;10&quot;&gt;112&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;52&quot; height=&quot;8&quot; width=&quot;3&quot; rx=&quot;2&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;RSA-2048
Against a CRQC: ≈0 bits
Factoring: broken by Shor&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;157&quot; y=&quot;59&quot; font-size=&quot;10&quot;&gt; broken &lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;142&quot; y=&quot;81&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;RSA-3072&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;70&quot; height=&quot;8&quot; width=&quot;225.88235294117646&quot; rx=&quot;2&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;RSA-3072
Classical: 128 bits
&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;379.88235294117646&quot; y=&quot;77&quot; font-size=&quot;10&quot;&gt;128&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;80&quot; height=&quot;8&quot; width=&quot;3&quot; rx=&quot;2&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;RSA-3072
Against a CRQC: ≈0 bits
Factoring: broken by Shor&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;157&quot; y=&quot;87&quot; font-size=&quot;10&quot;&gt; broken &lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;142&quot; y=&quot;109&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;DH MODP-3072&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;98&quot; height=&quot;8&quot; width=&quot;225.88235294117646&quot; rx=&quot;2&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;DH MODP-3072
Classical: 128 bits
&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;379.88235294117646&quot; y=&quot;105&quot; font-size=&quot;10&quot;&gt;128&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;108&quot; height=&quot;8&quot; width=&quot;3&quot; rx=&quot;2&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;DH MODP-3072
Against a CRQC: ≈0 bits
Discrete log: broken by Shor&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;157&quot; y=&quot;115&quot; font-size=&quot;10&quot;&gt; broken &lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;142&quot; y=&quot;137&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;ECDH/ECDSA P-256&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;126&quot; height=&quot;8&quot; width=&quot;225.88235294117646&quot; rx=&quot;2&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;ECDH/ECDSA P-256
Classical: 128 bits
&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;379.88235294117646&quot; y=&quot;133&quot; font-size=&quot;10&quot;&gt;128&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;136&quot; height=&quot;8&quot; width=&quot;3&quot; rx=&quot;2&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;ECDH/ECDSA P-256
Against a CRQC: ≈0 bits
Elliptic-curve discrete log: broken by Shor&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;157&quot; y=&quot;143&quot; font-size=&quot;10&quot;&gt; broken &lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;142&quot; y=&quot;165&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;X25519 / Ed25519&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;154&quot; height=&quot;8&quot; width=&quot;225.88235294117646&quot; rx=&quot;2&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;X25519 / Ed25519
Classical: 128 bits
&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;379.88235294117646&quot; y=&quot;161&quot; font-size=&quot;10&quot;&gt;128&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;164&quot; height=&quot;8&quot; width=&quot;3&quot; rx=&quot;2&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;X25519 / Ed25519
Against a CRQC: ≈0 bits
Elliptic-curve discrete log: broken by Shor&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;157&quot; y=&quot;171&quot; font-size=&quot;10&quot;&gt; broken &lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;142&quot; y=&quot;193&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;ECDH/ECDSA P-384&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;182&quot; height=&quot;8&quot; width=&quot;338.82352941176475&quot; rx=&quot;2&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;ECDH/ECDSA P-384
Classical: 192 bits
&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;492.82352941176475&quot; y=&quot;189&quot; font-size=&quot;10&quot;&gt;192&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;192&quot; height=&quot;8&quot; width=&quot;3&quot; rx=&quot;2&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;ECDH/ECDSA P-384
Against a CRQC: ≈0 bits
Elliptic-curve discrete log: broken by Shor&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;157&quot; y=&quot;199&quot; font-size=&quot;10&quot;&gt; broken &lt;/text&gt; &lt;/g&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;0&quot; y=&quot;225&quot; font-size=&quot;11&quot; font-weight=&quot;700&quot; class=&quot;t-primary&quot;&gt;Symmetric&lt;/text&gt; &lt;g&gt; &lt;text x=&quot;142&quot; y=&quot;243&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;AES-128&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;232&quot; height=&quot;8&quot; width=&quot;225.88235294117646&quot; rx=&quot;2&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;AES-128
Classical: 128 bits
&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;379.88235294117646&quot; y=&quot;239&quot; font-size=&quot;10&quot;&gt;128&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;242&quot; height=&quot;8&quot; width=&quot;112.94117647058823&quot; rx=&quot;2&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;AES-128
Against a CRQC: ≈64 bits (same as AES under Grover)
Grover: nominal halving is an upper bound. Grover parallelises poorly; a 2026 analysis (Valsorda) puts a realistic attack at about 2^104 operations. NIST still uses AES-128 as the Category 1 yardstick&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;266.94117647058823&quot; y=&quot;249&quot; font-size=&quot;10&quot;&gt; ≈64 &lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;142&quot; y=&quot;271&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;AES-192&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;260&quot; height=&quot;8&quot; width=&quot;338.82352941176475&quot; rx=&quot;2&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;AES-192
Classical: 192 bits
&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;492.82352941176475&quot; y=&quot;267&quot; font-size=&quot;10&quot;&gt;192&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;270&quot; height=&quot;8&quot; width=&quot;169.41176470588238&quot; rx=&quot;2&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;AES-192
Against a CRQC: ≈96 bits (same as AES under Grover)
Grover: nominal halving&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;323.4117647058824&quot; y=&quot;277&quot; font-size=&quot;10&quot;&gt; ≈96 &lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;142&quot; y=&quot;299&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;AES-256&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;288&quot; height=&quot;8&quot; width=&quot;451.7647058823529&quot; rx=&quot;2&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;AES-256
Classical: 256 bits
&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;605.7647058823529&quot; y=&quot;295&quot; font-size=&quot;10&quot;&gt;256&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;298&quot; height=&quot;8&quot; width=&quot;225.88235294117646&quot; rx=&quot;2&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;AES-256
Against a CRQC: ≈128 bits (same as AES under Grover)
Grover: nominal halving; comfortable margin&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;379.88235294117646&quot; y=&quot;305&quot; font-size=&quot;10&quot;&gt; ≈128 &lt;/text&gt; &lt;/g&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;0&quot; y=&quot;331&quot; font-size=&quot;11&quot; font-weight=&quot;700&quot; class=&quot;t-primary&quot;&gt;Post-quantum (by NIST category)&lt;/text&gt; &lt;g&gt; &lt;text x=&quot;142&quot; y=&quot;349&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;ML-KEM-512&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;338&quot; height=&quot;8&quot; width=&quot;225.88235294117646&quot; rx=&quot;2&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;ML-KEM-512
Classical: ≥128 bits
Category 1: at least as hard as AES-128 key search, classical and quantum&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;379.88235294117646&quot; y=&quot;345&quot; font-size=&quot;10&quot;&gt;≥128&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;348&quot; height=&quot;8&quot; width=&quot;112.94117647058823&quot; rx=&quot;2&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;ML-KEM-512
Against a CRQC: ≥ 64 bits (same as AES under Grover)
Category 1: at least as hard as AES-128 key search, classical and quantum&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;266.94117647058823&quot; y=&quot;355&quot; font-size=&quot;10&quot;&gt; ≥64 &lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;142&quot; y=&quot;377&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;ML-KEM-768&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;366&quot; height=&quot;8&quot; width=&quot;338.82352941176475&quot; rx=&quot;2&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;ML-KEM-768
Classical: ≥192 bits
Category 3: at least as hard as AES-192 key search&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;492.82352941176475&quot; y=&quot;373&quot; font-size=&quot;10&quot;&gt;≥192&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;376&quot; height=&quot;8&quot; width=&quot;169.41176470588238&quot; rx=&quot;2&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;ML-KEM-768
Against a CRQC: ≥ 96 bits (same as AES under Grover)
Category 3: at least as hard as AES-192 key search&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;323.4117647058824&quot; y=&quot;383&quot; font-size=&quot;10&quot;&gt; ≥96 &lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;142&quot; y=&quot;405&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;ML-KEM-1024&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;394&quot; height=&quot;8&quot; width=&quot;451.7647058823529&quot; rx=&quot;2&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;ML-KEM-1024
Classical: ≥256 bits
Category 5: at least as hard as AES-256 key search&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;605.7647058823529&quot; y=&quot;401&quot; font-size=&quot;10&quot;&gt;≥256&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;404&quot; height=&quot;8&quot; width=&quot;225.88235294117646&quot; rx=&quot;2&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;ML-KEM-1024
Against a CRQC: ≥ 128 bits (same as AES under Grover)
Category 5: at least as hard as AES-256 key search&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;379.88235294117646&quot; y=&quot;411&quot; font-size=&quot;10&quot;&gt; ≥128 &lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;142&quot; y=&quot;433&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;ML-DSA-65&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;422&quot; height=&quot;8&quot; width=&quot;338.82352941176475&quot; rx=&quot;2&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;ML-DSA-65
Classical: ≥192 bits
Category 3: at least as hard as AES-192 key search&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;492.82352941176475&quot; y=&quot;429&quot; font-size=&quot;10&quot;&gt;≥192&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;432&quot; height=&quot;8&quot; width=&quot;169.41176470588238&quot; rx=&quot;2&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;ML-DSA-65
Against a CRQC: ≥ 96 bits (same as AES under Grover)
Category 3: at least as hard as AES-192 key search&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;323.4117647058824&quot; y=&quot;439&quot; font-size=&quot;10&quot;&gt; ≥96 &lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;142&quot; y=&quot;461&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;ML-DSA-87&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;450&quot; height=&quot;8&quot; width=&quot;451.7647058823529&quot; rx=&quot;2&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;ML-DSA-87
Classical: ≥256 bits
Category 5: at least as hard as AES-256 key search&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;605.7647058823529&quot; y=&quot;457&quot; font-size=&quot;10&quot;&gt;≥256&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;460&quot; height=&quot;8&quot; width=&quot;225.88235294117646&quot; rx=&quot;2&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;ML-DSA-87
Against a CRQC: ≥ 128 bits (same as AES under Grover)
Category 5: at least as hard as AES-256 key search&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;379.88235294117646&quot; y=&quot;467&quot; font-size=&quot;10&quot;&gt; ≥128 &lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;142&quot; y=&quot;489&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;SLH-DSA-128s/f&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;478&quot; height=&quot;8&quot; width=&quot;225.88235294117646&quot; rx=&quot;2&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;SLH-DSA-128s/f
Classical: ≥128 bits
Category 1: at least as hard as AES-128 key search&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;379.88235294117646&quot; y=&quot;485&quot; font-size=&quot;10&quot;&gt;≥128&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;488&quot; height=&quot;8&quot; width=&quot;112.94117647058823&quot; rx=&quot;2&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;SLH-DSA-128s/f
Against a CRQC: ≥ 64 bits (same as AES under Grover)
Category 1: at least as hard as AES-128 key search&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;266.94117647058823&quot; y=&quot;495&quot; font-size=&quot;10&quot;&gt; ≥64 &lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;142&quot; y=&quot;517&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;SLH-DSA-256s/f&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;506&quot; height=&quot;8&quot; width=&quot;451.7647058823529&quot; rx=&quot;2&quot; fill=&quot;var(--series-1)&quot; tabindex=&quot;0&quot; data-tip=&quot;SLH-DSA-256s/f
Classical: ≥256 bits
Category 5: at least as hard as AES-256 key search&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;605.7647058823529&quot; y=&quot;513&quot; font-size=&quot;10&quot;&gt;≥256&lt;/text&gt; &lt;rect x=&quot;150&quot; y=&quot;516&quot; height=&quot;8&quot; width=&quot;225.88235294117646&quot; rx=&quot;2&quot; fill=&quot;var(--series-2)&quot; tabindex=&quot;0&quot; data-tip=&quot;SLH-DSA-256s/f
Against a CRQC: ≥ 128 bits (same as AES under Grover)
Category 5: at least as hard as AES-256 key search&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;379.88235294117646&quot; y=&quot;523&quot; font-size=&quot;10&quot;&gt; ≥128 &lt;/text&gt; &lt;/g&gt; &lt;/g&gt; &lt;text x=&quot;390&quot; y=&quot;564&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;security strength (bits)&lt;/text&gt; &lt;/svg&gt; &lt;details&gt; &lt;summary&gt;Show data as table&lt;/summary&gt; &lt;table&gt; &lt;thead&gt;&lt;tr&gt;&lt;th&gt;Algorithm&lt;/th&gt;&lt;th&gt;Classical&lt;/th&gt;&lt;th&gt;vs. CRQC&lt;/th&gt;&lt;th&gt;Why&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt; &lt;tbody&gt; &lt;tr&gt;&lt;td&gt;RSA-2048&lt;/td&gt;&lt;td&gt;112&lt;/td&gt;&lt;td&gt;broken&lt;/td&gt;&lt;td&gt;Factoring: broken by Shor&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;RSA-3072&lt;/td&gt;&lt;td&gt;128&lt;/td&gt;&lt;td&gt;broken&lt;/td&gt;&lt;td&gt;Factoring: broken by Shor&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;DH MODP-3072&lt;/td&gt;&lt;td&gt;128&lt;/td&gt;&lt;td&gt;broken&lt;/td&gt;&lt;td&gt;Discrete log: broken by Shor&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ECDH/ECDSA P-256&lt;/td&gt;&lt;td&gt;128&lt;/td&gt;&lt;td&gt;broken&lt;/td&gt;&lt;td&gt;Elliptic-curve discrete log: broken by Shor&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;X25519 / Ed25519&lt;/td&gt;&lt;td&gt;128&lt;/td&gt;&lt;td&gt;broken&lt;/td&gt;&lt;td&gt;Elliptic-curve discrete log: broken by Shor&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ECDH/ECDSA P-384&lt;/td&gt;&lt;td&gt;192&lt;/td&gt;&lt;td&gt;broken&lt;/td&gt;&lt;td&gt;Elliptic-curve discrete log: broken by Shor&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;AES-128&lt;/td&gt;&lt;td&gt;128&lt;/td&gt;&lt;td&gt;≈64&lt;/td&gt;&lt;td&gt;Grover: nominal halving is an upper bound. Grover parallelises poorly; a 2026 analysis (Valsorda) puts a realistic attack at about 2^104 operations. NIST still uses AES-128 as the Category 1 yardstick&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;AES-192&lt;/td&gt;&lt;td&gt;192&lt;/td&gt;&lt;td&gt;≈96&lt;/td&gt;&lt;td&gt;Grover: nominal halving&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;AES-256&lt;/td&gt;&lt;td&gt;256&lt;/td&gt;&lt;td&gt;≈128&lt;/td&gt;&lt;td&gt;Grover: nominal halving; comfortable margin&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ML-KEM-512&lt;/td&gt;&lt;td&gt;≥128&lt;/td&gt;&lt;td&gt;≥64&lt;/td&gt;&lt;td&gt;Category 1: at least as hard as AES-128 key search, classical and quantum&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ML-KEM-768&lt;/td&gt;&lt;td&gt;≥192&lt;/td&gt;&lt;td&gt;≥96&lt;/td&gt;&lt;td&gt;Category 3: at least as hard as AES-192 key search&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ML-KEM-1024&lt;/td&gt;&lt;td&gt;≥256&lt;/td&gt;&lt;td&gt;≥128&lt;/td&gt;&lt;td&gt;Category 5: at least as hard as AES-256 key search&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ML-DSA-65&lt;/td&gt;&lt;td&gt;≥192&lt;/td&gt;&lt;td&gt;≥96&lt;/td&gt;&lt;td&gt;Category 3: at least as hard as AES-192 key search&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ML-DSA-87&lt;/td&gt;&lt;td&gt;≥256&lt;/td&gt;&lt;td&gt;≥128&lt;/td&gt;&lt;td&gt;Category 5: at least as hard as AES-256 key search&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;SLH-DSA-128s/f&lt;/td&gt;&lt;td&gt;≥128&lt;/td&gt;&lt;td&gt;≥64&lt;/td&gt;&lt;td&gt;Category 1: at least as hard as AES-128 key search&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;SLH-DSA-256s/f&lt;/td&gt;&lt;td&gt;≥256&lt;/td&gt;&lt;td&gt;≥128&lt;/td&gt;&lt;td&gt;Category 5: at least as hard as AES-256 key search&lt;/td&gt;&lt;/tr&gt; &lt;/tbody&gt; &lt;/table&gt; &lt;/details&gt; &lt;/figure&gt;
&lt;p&gt;Post-quantum algorithms are not in this chart, because they are not rated in bits but in NIST &lt;em&gt;security categories&lt;/em&gt;. Each category is defined as “at least as hard to break as” a reference symmetric attack, under classical &lt;strong&gt;and&lt;/strong&gt; quantum attack (&lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/#4-security-categories-parameter-sets-and-sizes&quot;&gt;section 4&lt;/a&gt;).&lt;/p&gt;
&lt;h3 id=&quot;13-how-far-away-is-a-crqc&quot;&gt;1.3 How far away is a CRQC?&lt;/h3&gt;
&lt;p&gt;Nobody knows, but the estimated resources needed keep falling, and faster than expected:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;2019:&lt;/strong&gt; Gidney and Ekerå estimated that RSA-2048 could be factored in about eight hours with some &lt;strong&gt;20 million&lt;/strong&gt; noisy physical qubits.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;May 2025:&lt;/strong&gt; Craig Gidney (Google Quantum AI) cut that to &lt;strong&gt;fewer than one million&lt;/strong&gt; noisy qubits running for under a week, under the same hardware assumptions.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;March 2026:&lt;/strong&gt; a Caltech and Oratomic preprint (arXiv:2603.28627) proposed &lt;em&gt;high-rate&lt;/em&gt; error-correcting codes for neutral-atom machines. Today about 1,000 physical qubits are often needed per logical qubit; this approach could need as few as five. That leads to estimates of roughly &lt;strong&gt;10,000–26,000 physical qubits&lt;/strong&gt; to break 256-bit elliptic-curve cryptography in about ten days, and similar numbers for RSA-2048 with runtimes of months to years. The work is theoretical, not yet peer-reviewed, and assumes machines that still have to be built. Neutral-atom arrays with more than 6,000 atoms have been demonstrated.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Spring 2026:&lt;/strong&gt; Google researchers described a more efficient version of Shor’s algorithm against &lt;strong&gt;elliptic curves&lt;/strong&gt;. They published it as a zero-knowledge proof, showing feasibility without revealing the method.&lt;/li&gt;
&lt;/ul&gt;
 &lt;figure class=&quot;viz&quot;&gt; &lt;figcaption&gt;Estimated physical qubits to break RSA-2048&lt;/figcaption&gt; &lt;p class=&quot;viz-sub&quot;&gt;Logarithmic scale; each point is a published resource estimate, not a machine that exists. Assumptions differ (superconducting vs. neutral-atom hardware, runtime). Hover a point for details.&lt;/p&gt; &lt;svg viewBox=&quot;0 0 680 276&quot; role=&quot;img&quot; aria-label=&quot;Log-scale chart of published estimates of physical qubits needed to factor RSA-2048, falling from about a billion in 2012 to about ten thousand in 2026&quot;&gt; &lt;g&gt; &lt;line x1=&quot;70&quot; x2=&quot;650&quot; y1=&quot;236&quot; y2=&quot;236&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;62&quot; y=&quot;240&quot; text-anchor=&quot;end&quot; font-size=&quot;11&quot;&gt;1k&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;70&quot; x2=&quot;650&quot; y1=&quot;202.15384615384616&quot; y2=&quot;202.15384615384616&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;62&quot; y=&quot;206.15384615384616&quot; text-anchor=&quot;end&quot; font-size=&quot;11&quot;&gt;10k&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;70&quot; x2=&quot;650&quot; y1=&quot;168.3076923076923&quot; y2=&quot;168.3076923076923&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;62&quot; y=&quot;172.3076923076923&quot; text-anchor=&quot;end&quot; font-size=&quot;11&quot;&gt;100k&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;70&quot; x2=&quot;650&quot; y1=&quot;134.46153846153845&quot; y2=&quot;134.46153846153845&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;62&quot; y=&quot;138.46153846153845&quot; text-anchor=&quot;end&quot; font-size=&quot;11&quot;&gt;1M&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;70&quot; x2=&quot;650&quot; y1=&quot;100.61538461538461&quot; y2=&quot;100.61538461538461&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;62&quot; y=&quot;104.61538461538461&quot; text-anchor=&quot;end&quot; font-size=&quot;11&quot;&gt;10M&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;70&quot; x2=&quot;650&quot; y1=&quot;66.76923076923076&quot; y2=&quot;66.76923076923076&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;62&quot; y=&quot;70.76923076923076&quot; text-anchor=&quot;end&quot; font-size=&quot;11&quot;&gt;100M&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;70&quot; x2=&quot;650&quot; y1=&quot;32.92307692307691&quot; y2=&quot;32.92307692307691&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;62&quot; y=&quot;36.92307692307691&quot; text-anchor=&quot;end&quot; font-size=&quot;11&quot;&gt;1B&lt;/text&gt; &lt;/g&gt; &lt;text x=&quot;106.25&quot; y=&quot;254&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2012&lt;/text&gt;&lt;text x=&quot;178.75&quot; y=&quot;254&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2014&lt;/text&gt;&lt;text x=&quot;251.25&quot; y=&quot;254&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2016&lt;/text&gt;&lt;text x=&quot;323.75&quot; y=&quot;254&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2018&lt;/text&gt;&lt;text x=&quot;396.25&quot; y=&quot;254&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2020&lt;/text&gt;&lt;text x=&quot;468.75&quot; y=&quot;254&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2022&lt;/text&gt;&lt;text x=&quot;541.25&quot; y=&quot;254&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2024&lt;/text&gt;&lt;text x=&quot;613.75&quot; y=&quot;254&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2026&lt;/text&gt; &lt;polyline points=&quot;106.25,32.92307692307691 360,90.42667706983448 577.5,134.46153846153845 613.75,194.43072867012387&quot; fill=&quot;none&quot; stroke=&quot;var(--series-1)&quot; stroke-width=&quot;2&quot;&gt;&lt;/polyline&gt; &lt;g&gt;  &lt;circle cx=&quot;106.25&quot; cy=&quot;32.92307692307691&quot; r=&quot;6&quot; fill=&quot;var(--series-1)&quot; stroke=&quot;var(--surface-1)&quot; stroke-width=&quot;2&quot; tabindex=&quot;0&quot; data-tip=&quot;2012 estimate cited by Google Quantum AI: about a billion physical qubits&quot;&gt;&lt;/circle&gt; &lt;text x=&quot;116.25&quot; y=&quot;24.923076923076913&quot; text-anchor=&quot;start&quot; font-size=&quot;11&quot; class=&quot;t-primary&quot;&gt;≈1 billion&lt;/text&gt; &lt;/g&gt;&lt;g&gt;  &lt;circle cx=&quot;360&quot; cy=&quot;90.42667706983448&quot; r=&quot;6&quot; fill=&quot;var(--series-1)&quot; stroke=&quot;var(--surface-1)&quot; stroke-width=&quot;2&quot; tabindex=&quot;0&quot; data-tip=&quot;Gidney &amp;#38; Ekerå (2019): about 20 million noisy qubits, ~8 hours (superconducting assumptions)&quot;&gt;&lt;/circle&gt; &lt;text x=&quot;370&quot; y=&quot;82.42667706983448&quot; text-anchor=&quot;start&quot; font-size=&quot;11&quot; class=&quot;t-primary&quot;&gt;≈20 million&lt;/text&gt; &lt;/g&gt;&lt;g&gt;  &lt;circle cx=&quot;577.5&quot; cy=&quot;134.46153846153845&quot; r=&quot;6&quot; fill=&quot;var(--series-1)&quot; stroke=&quot;var(--surface-1)&quot; stroke-width=&quot;2&quot; tabindex=&quot;0&quot; data-tip=&quot;Gidney (May 2025): fewer than one million noisy qubits, under a week (same hardware assumptions as 2019)&quot;&gt;&lt;/circle&gt; &lt;text x=&quot;587.5&quot; y=&quot;126.46153846153845&quot; text-anchor=&quot;start&quot; font-size=&quot;11&quot; class=&quot;t-primary&quot;&gt;&amp;lt;1 million&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;613.75&quot; x2=&quot;613.75&quot; y1=&quot;200.75286296387546&quot; y2=&quot;188.1085943763723&quot; stroke=&quot;var(--series-1)&quot; stroke-width=&quot;6&quot; stroke-linecap=&quot;round&quot; opacity=&quot;0.5&quot;&gt;&lt;/line&gt; &lt;circle cx=&quot;613.75&quot; cy=&quot;194.43072867012387&quot; r=&quot;6&quot; fill=&quot;var(--series-1)&quot; stroke=&quot;var(--surface-1)&quot; stroke-width=&quot;2&quot; tabindex=&quot;0&quot; data-tip=&quot;Caltech &amp;#38; Oratomic preprint (Mar 2026, arXiv:2603.28627): ~11,000–26,000 qubits on neutral-atom hardware with high-rate error-correcting codes, months to years of runtime. Theoretical, not peer-reviewed.&quot;&gt;&lt;/circle&gt; &lt;text x=&quot;603.75&quot; y=&quot;186.43072867012387&quot; text-anchor=&quot;end&quot; font-size=&quot;11&quot; class=&quot;t-primary&quot;&gt;≈11k–26k&lt;/text&gt; &lt;/g&gt; &lt;text x=&quot;14&quot; y=&quot;126&quot; font-size=&quot;11&quot; transform=&quot;rotate(-90 14 126)&quot; text-anchor=&quot;middle&quot;&gt;physical qubits&lt;/text&gt; &lt;/svg&gt; &lt;details&gt; &lt;summary&gt;Show data as table&lt;/summary&gt; &lt;table&gt; &lt;thead&gt;&lt;tr&gt;&lt;th&gt;Year&lt;/th&gt;&lt;th&gt;Estimate&lt;/th&gt;&lt;th&gt;Source and assumptions&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt; &lt;tbody&gt;&lt;tr&gt;&lt;td&gt;2012&lt;/td&gt;&lt;td&gt;≈1 billion&lt;/td&gt;&lt;td&gt;2012 estimate cited by Google Quantum AI: about a billion physical qubits&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;2019&lt;/td&gt;&lt;td&gt;≈20 million&lt;/td&gt;&lt;td&gt;Gidney &amp;amp; Ekerå (2019): about 20 million noisy qubits, ~8 hours (superconducting assumptions)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;2025&lt;/td&gt;&lt;td&gt;&amp;lt;1 million&lt;/td&gt;&lt;td&gt;Gidney (May 2025): fewer than one million noisy qubits, under a week (same hardware assumptions as 2019)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;2026&lt;/td&gt;&lt;td&gt;≈11k–26k&lt;/td&gt;&lt;td&gt;Caltech &amp;amp; Oratomic preprint (Mar 2026, arXiv:2603.28627): ~11,000–26,000 qubits on neutral-atom hardware with high-rate error-correcting codes, months to years of runtime. Theoretical, not peer-reviewed.&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt; &lt;/table&gt; &lt;/details&gt; &lt;/figure&gt;
&lt;p&gt;Real demonstrations remain tiny. In April 2026 an independent researcher broke a &lt;strong&gt;15-bit&lt;/strong&gt; elliptic-curve key on a publicly accessible quantum computer (Project Eleven’s “Q-Day Prize”). The previous public record, in September 2025, was 6 bits. Hardware progress is real but incremental: Google showed error correction &lt;em&gt;below threshold&lt;/em&gt; on its 105-qubit Willow chip (late 2024), and IBM reported record logical-qubit fidelity in February 2026, albeit for only 55 microseconds.&lt;/p&gt;
&lt;p&gt;Two conclusions matter for planning:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Elliptic curves may fall first.&lt;/strong&gt; ECDH and ECDSA (used everywhere in TLS and IKEv2) need fewer logical qubits to break than RSA of comparable classical strength.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;The estimates dropped by about two orders of magnitude in seven years.&lt;/strong&gt; A planning horizon that assumes “decades” is no longer defensible.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;&lt;strong&gt;Does that mean you should switch to large MODP groups (DH-4096) instead?&lt;/strong&gt; No. Finite-field Diffie-Hellman is broken by the same Shor algorithm. A larger group raises the attacker’s cost only &lt;em&gt;polynomially&lt;/em&gt;, roughly in proportion to the key size, not exponentially as for classical attacks. It buys a little time at best. Meanwhile it costs CPU time and packet space (512 bytes per DH-4096 public value), and it does nothing against harvest-now-decrypt-later once a CRQC exists. The way out is a post-quantum algorithm, ideally combined with a fast elliptic curve in a PQ/T hybrid (&lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/#32-pure-pq-pqt-hybrid-and-pqpq-hybrid&quot;&gt;section 3.2&lt;/a&gt;).&lt;/p&gt;
&lt;p&gt;The symmetric side remains calm. A 2026 analysis by cryptographer Filippo Valsorda argues that because Grover’s algorithm parallelises so poorly, attacking AES-128 would need about 2⁴⁷ parallel quantum machines and some 2¹⁰⁴·⁵ operations in total. That is in line with NIST and BSI, which keep AES-128 acceptable.&lt;/p&gt;
&lt;h3 id=&quot;14-harvest-now-decrypt-later&quot;&gt;1.4 Harvest now, decrypt later&lt;/h3&gt;
&lt;p&gt;For IPsec the decisive argument is not &lt;em&gt;when&lt;/em&gt; a CRQC will exist, but what an adversary can do &lt;strong&gt;today&lt;/strong&gt;:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Confidentiality is at risk now.&lt;/strong&gt; An attacker can record encrypted traffic (TLS, SSH, IPsec VPNs) today and decrypt it later, once a CRQC is available (“harvest now, decrypt later”, HNDL). Long-lived data such as health records, intellectual property and state secrets is worth storing.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Authentication is at risk later.&lt;/strong&gt; A signature can only be forged &lt;em&gt;after&lt;/em&gt; a CRQC exists; a handshake that is already finished cannot be impersonated retroactively. Long-lived trust anchors, certificates and firmware-signing keys still need a head start.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;This leads to the core priority for every protocol: &lt;strong&gt;make the key exchange quantum-resistant first, then the authentication.&lt;/strong&gt;&lt;/p&gt;
&lt;p&gt;A useful rule of thumb is &lt;em&gt;Mosca’s inequality&lt;/em&gt;. If &lt;em&gt;x&lt;/em&gt; is how long your data must stay confidential, &lt;em&gt;y&lt;/em&gt; is how long your migration takes, and &lt;em&gt;z&lt;/em&gt; is the time until a CRQC exists, then you are already too late when &lt;em&gt;x + y &amp;gt; z&lt;/em&gt;. For network infrastructure with hardware refresh cycles of 5–10 years, &lt;em&gt;y&lt;/em&gt; alone is large.&lt;/p&gt;
&lt;h3 id=&quot;15-policy-deadlines&quot;&gt;1.5 Policy deadlines&lt;/h3&gt;













































&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Who&lt;/th&gt;&lt;th&gt;Document&lt;/th&gt;&lt;th&gt;Key dates&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;NIST (US)&lt;/td&gt;&lt;td&gt;NIST IR 8547 (final, 27 Oct 2026)&lt;/td&gt;&lt;td&gt;Quantum-vulnerable public-key algorithms deprecated after 2030 and disallowed after 2035&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;EU Member States&lt;/td&gt;&lt;td&gt;Coordinated Implementation Roadmap (June 2025)&lt;/td&gt;&lt;td&gt;Start transition by end of 2026; high-risk use cases migrated by end of 2030; as many systems as feasible by 2035&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;UK NCSC&lt;/td&gt;&lt;td&gt;Timelines for migration to PQC (Mar 2025)&lt;/td&gt;&lt;td&gt;Plan by 2028; highest-priority migrations by 2031; complete by 2035&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Germany (BSI)&lt;/td&gt;&lt;td&gt;TR-02102-1, version 2026-01&lt;/td&gt;&lt;td&gt;Classical-only key agreement no longer recommended after 2031 (2030 for the most sensitive uses)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;France (ANSSI)&lt;/td&gt;&lt;td&gt;Announcement, June 2026&lt;/td&gt;&lt;td&gt;No security certification for products without PQC from 2027&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Australia (ASD)&lt;/td&gt;&lt;td&gt;ISM / &lt;em&gt;Planning for PQC&lt;/em&gt; (Sep 2025)&lt;/td&gt;&lt;td&gt;Cease traditional asymmetric cryptography by &lt;strong&gt;end of 2030&lt;/strong&gt;&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;NSA (US NSS)&lt;/td&gt;&lt;td&gt;CNSA 2.0, made binding by &lt;strong&gt;CNSSP 15&lt;/strong&gt; (&lt;em&gt;Committee on National Security Systems Policy 15&lt;/em&gt;)&lt;/td&gt;&lt;td&gt;New acquisitions CNSA 2.0-compliant from 1 Jan 2027; see &lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/#24-nsa-suite-b-cnsa-10-and-cnsa-20&quot;&gt;section 2.4&lt;/a&gt;&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
 &lt;figure class=&quot;viz&quot;&gt; &lt;figcaption&gt;PQC policy timelines by authority&lt;/figcaption&gt; &lt;p class=&quot;viz-sub&quot;&gt;Coloured segments show what each authority expects in that period. Grey hatching marks the past. Hover or tab to a segment for the exact requirement.&lt;/p&gt; &lt;div class=&quot;legend&quot;&gt; &lt;span style=&quot;--sw: var(--series-1)&quot;&gt;Prepare &amp;amp; plan&lt;/span&gt;&lt;span style=&quot;--sw: var(--series-2)&quot;&gt;Migrate (priority / high-risk first)&lt;/span&gt;&lt;span style=&quot;--sw: var(--series-3)&quot;&gt;Complete migration; classical deprecated&lt;/span&gt;&lt;span style=&quot;--sw: #6250d6&quot;&gt;Classical-only no longer acceptable&lt;/span&gt; &lt;/div&gt; &lt;svg viewBox=&quot;0 0 680 250&quot; role=&quot;img&quot; aria-label=&quot;Timeline bars of PQC migration phases for EU, NIST, NSA, UK, Germany, France and Australia&quot;&gt; &lt;defs&gt; &lt;pattern id=&quot;hatch-policy&quot; patternUnits=&quot;userSpaceOnUse&quot; width=&quot;6&quot; height=&quot;6&quot; patternTransform=&quot;rotate(45)&quot;&gt; &lt;rect width=&quot;6&quot; height=&quot;6&quot; fill=&quot;#f1f0ec&quot;&gt;&lt;/rect&gt; &lt;line x1=&quot;0&quot; y1=&quot;0&quot; x2=&quot;0&quot; y2=&quot;6&quot; stroke=&quot;#cfcdc6&quot; stroke-width=&quot;2&quot;&gt;&lt;/line&gt; &lt;/pattern&gt; &lt;/defs&gt; &lt;rect x=&quot;130&quot; y=&quot;8&quot; width=&quot;87.87657784011674&quot; height=&quot;214&quot; fill=&quot;url(#hatch-policy)&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;134&quot; y=&quot;218&quot; font-size=&quot;10&quot;&gt;past&lt;/text&gt; &lt;g&gt; &lt;line x1=&quot;130&quot; x2=&quot;130&quot; y1=&quot;8&quot; y2=&quot;228&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;130&quot; y=&quot;244&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2025&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;176.43478260869566&quot; x2=&quot;176.43478260869566&quot; y1=&quot;8&quot; y2=&quot;228&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;176.43478260869566&quot; y=&quot;244&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2026&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;222.8695652173913&quot; x2=&quot;222.8695652173913&quot; y1=&quot;8&quot; y2=&quot;228&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;222.8695652173913&quot; y=&quot;244&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2027&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;269.30434782608694&quot; x2=&quot;269.30434782608694&quot; y1=&quot;8&quot; y2=&quot;228&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;269.30434782608694&quot; y=&quot;244&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2028&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;315.7391304347826&quot; x2=&quot;315.7391304347826&quot; y1=&quot;8&quot; y2=&quot;228&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;315.7391304347826&quot; y=&quot;244&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2029&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;362.17391304347825&quot; x2=&quot;362.17391304347825&quot; y1=&quot;8&quot; y2=&quot;228&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;362.17391304347825&quot; y=&quot;244&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2030&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;408.6086956521739&quot; x2=&quot;408.6086956521739&quot; y1=&quot;8&quot; y2=&quot;228&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;408.6086956521739&quot; y=&quot;244&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2031&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;455.04347826086956&quot; x2=&quot;455.04347826086956&quot; y1=&quot;8&quot; y2=&quot;228&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;455.04347826086956&quot; y=&quot;244&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2032&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;501.4782608695652&quot; x2=&quot;501.4782608695652&quot; y1=&quot;8&quot; y2=&quot;228&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;501.4782608695652&quot; y=&quot;244&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2033&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;547.9130434782609&quot; x2=&quot;547.9130434782609&quot; y1=&quot;8&quot; y2=&quot;228&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;547.9130434782609&quot; y=&quot;244&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2034&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;594.3478260869565&quot; x2=&quot;594.3478260869565&quot; y1=&quot;8&quot; y2=&quot;228&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;594.3478260869565&quot; y=&quot;244&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2035&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;640.7826086956522&quot; x2=&quot;640.7826086956522&quot; y1=&quot;8&quot; y2=&quot;228&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;640.7826086956522&quot; y=&quot;244&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2036&lt;/text&gt; &lt;/g&gt; &lt;g&gt; &lt;text x=&quot;122&quot; y=&quot;29&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;EU (NIS CG roadmap)&lt;/text&gt; &lt;rect x=&quot;131&quot; y=&quot;18&quot; width=&quot;90.86956521739131&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;var(--series-1)&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;EU: national roadmaps and first steps by end of 2026
(Prepare &amp;#38; plan)&quot;&gt;&lt;/rect&gt;&lt;rect x=&quot;223.8695652173913&quot; y=&quot;18&quot; width=&quot;183.73913043478257&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;var(--series-2)&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;EU: high-risk use cases migrated by end of 2030; no stand-alone quantum-vulnerable public-key crypto there after 2030
(Migrate (priority / high-risk first))&quot;&gt;&lt;/rect&gt;&lt;rect x=&quot;409.6086956521739&quot; y=&quot;18&quot; width=&quot;230.17391304347836&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;var(--series-3)&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;EU: as many systems as feasible migrated by 2035
(Complete migration; classical deprecated)&quot;&gt;&lt;/rect&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;122&quot; y=&quot;59&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;NIST (IR 8547)&lt;/text&gt; &lt;rect x=&quot;131&quot; y=&quot;48&quot; width=&quot;276.6086956521739&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;var(--series-2)&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;NIST: transition period; quantum-vulnerable algorithms still allowed until end of 2030
(Migrate (priority / high-risk first))&quot;&gt;&lt;/rect&gt;&lt;rect x=&quot;409.6086956521739&quot; y=&quot;48&quot; width=&quot;230.17391304347836&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;var(--series-3)&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;NIST: quantum-vulnerable public-key algorithms deprecated after 2030
(Complete migration; classical deprecated)&quot;&gt;&lt;/rect&gt;&lt;rect x=&quot;641.7826086956522&quot; y=&quot;48&quot; width=&quot;21.217391304347757&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;#6250d6&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;NIST: quantum-vulnerable public-key algorithms disallowed after 2035
(Classical-only no longer acceptable)&quot;&gt;&lt;/rect&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;122&quot; y=&quot;89&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;NSA (CNSSP 15)&lt;/text&gt; &lt;rect x=&quot;131&quot; y=&quot;78&quot; width=&quot;90.86956521739131&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;var(--series-1)&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;NSS: prepare; vendors add CNSA 2.0 support
(Prepare &amp;#38; plan)&quot;&gt;&lt;/rect&gt;&lt;rect x=&quot;223.8695652173913&quot; y=&quot;78&quot; width=&quot;183.73913043478257&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;var(--series-2)&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;NSS: new acquisitions CNSA 2.0-compliant from 1 Jan 2027; non-supporting equipment phased out by 31 Dec 2030
(Migrate (priority / high-risk first))&quot;&gt;&lt;/rect&gt;&lt;rect x=&quot;409.6086956521739&quot; y=&quot;78&quot; width=&quot;44.434782608695684&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;var(--series-3)&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;NSS: CNSA 2.0 algorithms mandated by 31 Dec 2031
(Complete migration; classical deprecated)&quot;&gt;&lt;/rect&gt;&lt;rect x=&quot;456.04347826086956&quot; y=&quot;78&quot; width=&quot;206.95652173913044&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;#6250d6&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;NSS: CNSA 2.0 mandatory; all NSS quantum-resistant by 2035 (NSM-10)
(Classical-only no longer acceptable)&quot;&gt;&lt;/rect&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;122&quot; y=&quot;119&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;UK (NCSC)&lt;/text&gt; &lt;rect x=&quot;131&quot; y=&quot;108&quot; width=&quot;137.30434782608694&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;var(--series-1)&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;UK: discovery and migration plan by 2028
(Prepare &amp;#38; plan)&quot;&gt;&lt;/rect&gt;&lt;rect x=&quot;270.30434782608694&quot; y=&quot;108&quot; width=&quot;137.30434782608694&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;var(--series-2)&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;UK: highest-priority migrations by 2031
(Migrate (priority / high-risk first))&quot;&gt;&lt;/rect&gt;&lt;rect x=&quot;409.6086956521739&quot; y=&quot;108&quot; width=&quot;183.73913043478262&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;var(--series-3)&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;UK: complete migration of all systems by 2035
(Complete migration; classical deprecated)&quot;&gt;&lt;/rect&gt;&lt;rect x=&quot;595.3478260869565&quot; y=&quot;108&quot; width=&quot;67.6521739130435&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;#6250d6&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;UK: migration to PQC complete
(Classical-only no longer acceptable)&quot;&gt;&lt;/rect&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;122&quot; y=&quot;149&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;Germany (BSI)&lt;/text&gt; &lt;rect x=&quot;131&quot; y=&quot;138&quot; width=&quot;276.6086956521739&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;var(--series-2)&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;BSI TR-02102-1: hybrid PQC key establishment recommended now
(Migrate (priority / high-risk first))&quot;&gt;&lt;/rect&gt;&lt;rect x=&quot;409.6086956521739&quot; y=&quot;138&quot; width=&quot;44.434782608695684&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;var(--series-3)&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;BSI: classical-only key agreement no longer recommended after 2030 for the most sensitive uses
(Complete migration; classical deprecated)&quot;&gt;&lt;/rect&gt;&lt;rect x=&quot;456.04347826086956&quot; y=&quot;138&quot; width=&quot;206.95652173913044&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;#6250d6&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;BSI: classical-only key agreement no longer recommended after 2031
(Classical-only no longer acceptable)&quot;&gt;&lt;/rect&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;122&quot; y=&quot;179&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;France (ANSSI)&lt;/text&gt; &lt;rect x=&quot;131&quot; y=&quot;168&quot; width=&quot;90.86956521739131&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;var(--series-2)&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;ANSSI: hybrid PQC strongly recommended
(Migrate (priority / high-risk first))&quot;&gt;&lt;/rect&gt;&lt;rect x=&quot;223.8695652173913&quot; y=&quot;168&quot; width=&quot;439.1304347826087&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;#6250d6&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;ANSSI (announced June 2026): no security certification for products without PQC from 2027
(Classical-only no longer acceptable)&quot;&gt;&lt;/rect&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;122&quot; y=&quot;209&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot;&gt;Australia (ASD)&lt;/text&gt; &lt;rect x=&quot;131&quot; y=&quot;198&quot; width=&quot;90.86956521739131&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;var(--series-1)&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;ASD: refined transition plan by end of 2026
(Prepare &amp;#38; plan)&quot;&gt;&lt;/rect&gt;&lt;rect x=&quot;223.8695652173913&quot; y=&quot;198&quot; width=&quot;90.86956521739131&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;var(--series-2)&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;ASD: migration of critical systems started by end of 2028
(Migrate (priority / high-risk first))&quot;&gt;&lt;/rect&gt;&lt;rect x=&quot;316.7391304347826&quot; y=&quot;198&quot; width=&quot;90.86956521739125&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;var(--series-3)&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;ASD: transition complete by end of 2030
(Complete migration; classical deprecated)&quot;&gt;&lt;/rect&gt;&lt;rect x=&quot;409.6086956521739&quot; y=&quot;198&quot; width=&quot;253.39130434782612&quot; height=&quot;16&quot; rx=&quot;3&quot; fill=&quot;#6250d6&quot; fill-opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;ASD: traditional asymmetric cryptography not approved beyond 2030
(Classical-only no longer acceptable)&quot;&gt;&lt;/rect&gt; &lt;/g&gt; &lt;line x1=&quot;217.87657784011674&quot; x2=&quot;217.87657784011674&quot; y1=&quot;4&quot; y2=&quot;228&quot; stroke=&quot;var(--text-primary)&quot; stroke-width=&quot;1.5&quot;&gt;&lt;/line&gt; &lt;text x=&quot;221.87657784011674&quot; y=&quot;12&quot; font-size=&quot;10&quot; class=&quot;t-primary&quot;&gt;today&lt;/text&gt; &lt;/svg&gt; &lt;details&gt; &lt;summary&gt;Show data as table&lt;/summary&gt; &lt;table&gt; &lt;thead&gt;&lt;tr&gt;&lt;th&gt;Authority&lt;/th&gt;&lt;th&gt;Period&lt;/th&gt;&lt;th&gt;Requirement&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt; &lt;tbody&gt; &lt;tr&gt;&lt;td&gt;EU (NIS CG roadmap)&lt;/td&gt;&lt;td&gt;2025–2027&lt;/td&gt;&lt;td&gt;EU: national roadmaps and first steps by end of 2026&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;EU (NIS CG roadmap)&lt;/td&gt;&lt;td&gt;2027–2031&lt;/td&gt;&lt;td&gt;EU: high-risk use cases migrated by end of 2030; no stand-alone quantum-vulnerable public-key crypto there after 2030&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;EU (NIS CG roadmap)&lt;/td&gt;&lt;td&gt;2031–2036&lt;/td&gt;&lt;td&gt;EU: as many systems as feasible migrated by 2035&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;NIST (IR 8547)&lt;/td&gt;&lt;td&gt;2025–2031&lt;/td&gt;&lt;td&gt;NIST: transition period; quantum-vulnerable algorithms still allowed until end of 2030&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;NIST (IR 8547)&lt;/td&gt;&lt;td&gt;2031–2036&lt;/td&gt;&lt;td&gt;NIST: quantum-vulnerable public-key algorithms deprecated after 2030&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;NIST (IR 8547)&lt;/td&gt;&lt;td&gt;2036–…&lt;/td&gt;&lt;td&gt;NIST: quantum-vulnerable public-key algorithms disallowed after 2035&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;NSA (CNSSP 15)&lt;/td&gt;&lt;td&gt;2025–2027&lt;/td&gt;&lt;td&gt;NSS: prepare; vendors add CNSA 2.0 support&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;NSA (CNSSP 15)&lt;/td&gt;&lt;td&gt;2027–2031&lt;/td&gt;&lt;td&gt;NSS: new acquisitions CNSA 2.0-compliant from 1 Jan 2027; non-supporting equipment phased out by 31 Dec 2030&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;NSA (CNSSP 15)&lt;/td&gt;&lt;td&gt;2031–2032&lt;/td&gt;&lt;td&gt;NSS: CNSA 2.0 algorithms mandated by 31 Dec 2031&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;NSA (CNSSP 15)&lt;/td&gt;&lt;td&gt;2032–…&lt;/td&gt;&lt;td&gt;NSS: CNSA 2.0 mandatory; all NSS quantum-resistant by 2035 (NSM-10)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;UK (NCSC)&lt;/td&gt;&lt;td&gt;2025–2028&lt;/td&gt;&lt;td&gt;UK: discovery and migration plan by 2028&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;UK (NCSC)&lt;/td&gt;&lt;td&gt;2028–2031&lt;/td&gt;&lt;td&gt;UK: highest-priority migrations by 2031&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;UK (NCSC)&lt;/td&gt;&lt;td&gt;2031–2035&lt;/td&gt;&lt;td&gt;UK: complete migration of all systems by 2035&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;UK (NCSC)&lt;/td&gt;&lt;td&gt;2035–…&lt;/td&gt;&lt;td&gt;UK: migration to PQC complete&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Germany (BSI)&lt;/td&gt;&lt;td&gt;2025–2031&lt;/td&gt;&lt;td&gt;BSI TR-02102-1: hybrid PQC key establishment recommended now&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Germany (BSI)&lt;/td&gt;&lt;td&gt;2031–2032&lt;/td&gt;&lt;td&gt;BSI: classical-only key agreement no longer recommended after 2030 for the most sensitive uses&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Germany (BSI)&lt;/td&gt;&lt;td&gt;2032–…&lt;/td&gt;&lt;td&gt;BSI: classical-only key agreement no longer recommended after 2031&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;France (ANSSI)&lt;/td&gt;&lt;td&gt;2025–2027&lt;/td&gt;&lt;td&gt;ANSSI: hybrid PQC strongly recommended&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;France (ANSSI)&lt;/td&gt;&lt;td&gt;2027–…&lt;/td&gt;&lt;td&gt;ANSSI (announced June 2026): no security certification for products without PQC from 2027&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Australia (ASD)&lt;/td&gt;&lt;td&gt;2025–2027&lt;/td&gt;&lt;td&gt;ASD: refined transition plan by end of 2026&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Australia (ASD)&lt;/td&gt;&lt;td&gt;2027–2029&lt;/td&gt;&lt;td&gt;ASD: migration of critical systems started by end of 2028&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Australia (ASD)&lt;/td&gt;&lt;td&gt;2029–2031&lt;/td&gt;&lt;td&gt;ASD: transition complete by end of 2030&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Australia (ASD)&lt;/td&gt;&lt;td&gt;2031–…&lt;/td&gt;&lt;td&gt;ASD: traditional asymmetric cryptography not approved beyond 2030&lt;/td&gt;&lt;/tr&gt; &lt;/tbody&gt; &lt;/table&gt; &lt;/details&gt; &lt;/figure&gt;
&lt;p&gt;&lt;strong&gt;How urgent is it?&lt;/strong&gt; For any system that protects data with a confidentiality requirement beyond roughly 2030, PQ/T hybrid key exchange should be deployed &lt;strong&gt;now&lt;/strong&gt;. For IPsec, the standards and code points are in place; see the companion article &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/&quot;&gt;&lt;em&gt;Post-Quantum Cryptography in IPsec&lt;/em&gt;&lt;/a&gt;.&lt;/p&gt;
&lt;hr/&gt;
&lt;h2 id=&quot;2-who-sets-the-rules&quot;&gt;2. Who sets the rules?&lt;/h2&gt;
&lt;p&gt;Several kinds of organisations shape PQC. The &lt;strong&gt;IETF&lt;/strong&gt; writes the protocol specifications. &lt;strong&gt;NIST&lt;/strong&gt; standardises the algorithms. &lt;strong&gt;National and regional authorities&lt;/strong&gt; decide which algorithms you may use, in which mode, and by when; for US national security systems, that is the &lt;strong&gt;NSA&lt;/strong&gt;. (For IPsec, &lt;strong&gt;IANA&lt;/strong&gt; also matters: it assigns the numbers that go on the wire, see the &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#2-iana-the-ikev2-registries&quot;&gt;IPsec article&lt;/a&gt;.)&lt;/p&gt;
&lt;h3 id=&quot;21-the-ietf-and-its-rfcs&quot;&gt;2.1 The IETF and its RFCs&lt;/h3&gt;
&lt;h4 id=&quot;ietf-internet-drafts-and-rfcs&quot;&gt;IETF, Internet-Drafts and RFCs&lt;/h4&gt;
&lt;p&gt;The &lt;strong&gt;Internet Engineering Task Force (IETF)&lt;/strong&gt; develops Internet protocols in open &lt;strong&gt;working groups&lt;/strong&gt; (WGs), grouped into areas. Security protocols such as IPsec live in the Security Area. The &lt;strong&gt;IESG&lt;/strong&gt; (Internet Engineering Steering Group) approves documents for publication.&lt;/p&gt;
&lt;p&gt;Work starts as an &lt;strong&gt;Internet-Draft&lt;/strong&gt; (I-D). A draft named &lt;code&gt;draft-ietf-&amp;lt;wg&amp;gt;-…&lt;/code&gt; has been adopted by a working group. A draft named &lt;code&gt;draft-&amp;lt;author&amp;gt;-…&lt;/code&gt; is an individual submission with no formal standing. Drafts expire after six months unless they are updated, and they should be cited only as “work in progress.” Typical states you will see on the Datatracker are &lt;em&gt;WG Document&lt;/em&gt;, &lt;em&gt;Submitted to IESG for Publication&lt;/em&gt;, &lt;em&gt;RFC Ed Queue&lt;/em&gt;, and finally &lt;em&gt;RFC&lt;/em&gt;.&lt;/p&gt;
&lt;p&gt;A &lt;strong&gt;Request for Comments (RFC)&lt;/strong&gt; is never changed after publication. Corrections go into errata, and new versions get a new number that &lt;em&gt;obsoletes&lt;/em&gt; or &lt;em&gt;updates&lt;/em&gt; the older one. RFCs come from different &lt;strong&gt;streams&lt;/strong&gt;: IETF, IRTF, IAB, and the Independent Submission stream. Independent RFCs are not IETF standards. RFC 9206, the CNSA 1.0 profile for IPsec, is an example.&lt;/p&gt;
&lt;h4 id=&quot;rfc-statuses&quot;&gt;RFC statuses&lt;/h4&gt;

































&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Status&lt;/th&gt;&lt;th&gt;Meaning&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;Proposed Standard&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;First level of the Standards Track&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;Internet Standard&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;Second (and final) level; gets an STD number. IKEv2 itself is STD 79 (RFC 7296).&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Best Current Practice (BCP)&lt;/td&gt;&lt;td&gt;Process or operational guidance&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Informational&lt;/td&gt;&lt;td&gt;Not a standard: terminology, overviews, guidance (e.g. RFC 9794, RFC 9958)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Experimental&lt;/td&gt;&lt;td&gt;Specifications for experimentation&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Historic&lt;/td&gt;&lt;td&gt;Superseded or obsolete (e.g. IKEv1 since RFC 9395)&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;h4 id=&quot;proposed-standard-is-in-practice-the-standard&quot;&gt;“Proposed Standard” is, in practice, &lt;em&gt;the&lt;/em&gt; standard&lt;/h4&gt;
&lt;p&gt;RFC 2026 originally defined three maturity levels: Proposed Standard, Draft Standard and Internet Standard. &lt;strong&gt;RFC 6410&lt;/strong&gt; (2011) cut this to two by merging Draft Standard into Internet Standard. &lt;strong&gt;RFC 7127&lt;/strong&gt; (2014) then rewrote the characterisation of Proposed Standard: such a specification is stable, has been thoroughly reviewed, and is deployed and relied upon on the Internet. Very few documents are ever advanced to Internet Standard.&lt;/p&gt;
&lt;p&gt;So when you read that a specification is “only” a Proposed Standard: that &lt;strong&gt;is&lt;/strong&gt; the IETF standard. RFC 9370 (multiple key exchanges), RFC 9242 (IKE_INTERMEDIATE), RFC 8784 (PPK) and the upcoming ML-KEM and ML-DSA specifications for IKEv2 are all on the Standards Track at Proposed Standard level.&lt;/p&gt;
&lt;h3 id=&quot;22-nist-and-the-fips-standards&quot;&gt;2.2 NIST and the FIPS standards&lt;/h3&gt;
&lt;p&gt;The US &lt;strong&gt;National Institute of Standards and Technology (NIST)&lt;/strong&gt; ran the public PQC competition that started in 2016. The resulting &lt;strong&gt;Federal Information Processing Standards (FIPS)&lt;/strong&gt; are binding for US federal systems. They are also the de-facto global reference, adopted by the IETF and by most national agencies, with some local preferences.&lt;/p&gt;























































&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Publication&lt;/th&gt;&lt;th&gt;Content&lt;/th&gt;&lt;th&gt;Status (Nov 2026)&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;FIPS 203&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;ML-KEM (from CRYSTALS-Kyber), key encapsulation&lt;/td&gt;&lt;td&gt;Final, 13 Aug 2024&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;FIPS 204&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;ML-DSA (from CRYSTALS-Dilithium), signatures&lt;/td&gt;&lt;td&gt;Final, 13 Aug 2024&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;FIPS 205&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;SLH-DSA (from SPHINCS+), stateless hash-based signatures&lt;/td&gt;&lt;td&gt;Final, 13 Aug 2024&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;FIPS 206&lt;/td&gt;&lt;td&gt;FN-DSA (from Falcon), compact lattice signatures&lt;/td&gt;&lt;td&gt;In development; not yet published as a final standard&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;HQC standard&lt;/td&gt;&lt;td&gt;HQC, code-based backup KEM (expected as FIPS 207)&lt;/td&gt;&lt;td&gt;Selected 11 Mar 2025 (NIST IR 8545); standard not yet final&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;SP 800-208&lt;/td&gt;&lt;td&gt;LMS and XMSS, stateful hash-based signatures&lt;/td&gt;&lt;td&gt;Final (2020)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;SP 800-227&lt;/td&gt;&lt;td&gt;Recommendations for KEMs, including combining them&lt;/td&gt;&lt;td&gt;Final, 18 Sep 2025&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;NIST IR 8547&lt;/td&gt;&lt;td&gt;Transition to PQC standards (timeline)&lt;/td&gt;&lt;td&gt;Final, 27 Oct 2026&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;NIST IR 8610&lt;/td&gt;&lt;td&gt;Status report, round 2 of the additional signature schemes&lt;/td&gt;&lt;td&gt;Final, 14 May 2026&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;p&gt;What this means in practice:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;ML-KEM&lt;/strong&gt; is the PQ key exchange used by the IETF for TLS, SSH and IKEv2.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;ML-DSA&lt;/strong&gt; and &lt;strong&gt;SLH-DSA&lt;/strong&gt; are the PQ signature algorithms for authentication and certificates.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;HQC&lt;/strong&gt; matters for PQ/PQ hybrids. It uses different mathematics (codes rather than lattices) as a hedge against a break of ML-KEM.&lt;/li&gt;
&lt;li&gt;FIPS validation of cryptographic modules (CMVP, FIPS 140-3) determines when PQC is usable in regulated environments.&lt;/li&gt;
&lt;/ul&gt;
&lt;h3 id=&quot;23-european-and-australian-authorities&quot;&gt;2.3 European and Australian authorities&lt;/h3&gt;
&lt;p&gt;NIST defines the algorithms, but &lt;strong&gt;national and regional authorities&lt;/strong&gt; decide which ones you may use, in which mode (hybrid or not), and by when. Their positions differ, especially on hybrids, so check the guidance for the country and sector your IPsec deployment serves.&lt;/p&gt;
&lt;h4 id=&quot;european-union&quot;&gt;European Union&lt;/h4&gt;






























&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Body&lt;/th&gt;&lt;th&gt;Role&lt;/th&gt;&lt;th&gt;Key PQC output&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;European Commission&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;Policy&lt;/td&gt;&lt;td&gt;&lt;em&gt;Recommendation on a Coordinated Implementation Roadmap for the transition to PQC&lt;/em&gt; (11 April 2024)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;NIS Cooperation Group&lt;/strong&gt;, PQC work stream (co-chaired by Germany, France and the Netherlands)&lt;/td&gt;&lt;td&gt;Coordination between Member States&lt;/td&gt;&lt;td&gt;&lt;em&gt;Coordinated Implementation Roadmap for the Transition to PQC&lt;/em&gt; (23 June 2025): start by end of 2026, high-risk use cases by end of 2030, as many systems as feasible by 2035. A FAQ followed in September 2026; a second deliverable is in preparation.&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;ENISA&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;EU cybersecurity agency&lt;/td&gt;&lt;td&gt;Studies on PQC and migration; participates in the roadmap and G7 work&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;ETSI&lt;/strong&gt; TC CYBER, Quantum-Safe Cryptography WG&lt;/td&gt;&lt;td&gt;European standards body&lt;/td&gt;&lt;td&gt;&lt;strong&gt;ETSI TS 103 744&lt;/strong&gt; &lt;em&gt;Quantum-safe Hybrid Key Establishment&lt;/em&gt; (V1.2.2, Feb 2026): ECDH + ML-KEM combiners (CatKDF, CasKDF). A European Norm on authenticated hybrid key establishment is in development.&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;p&gt;The roadmap’s high-risk category explicitly includes critical infrastructure such as energy, water, health, finance and transport. For those, the guidance is that quantum-vulnerable public-key mechanisms should no longer be used &lt;strong&gt;on their own&lt;/strong&gt; after 2030. In practice that means hybrid or pure PQ.&lt;/p&gt;
&lt;h4 id=&quot;national-authorities-in-europe&quot;&gt;National authorities in Europe&lt;/h4&gt;








































&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Country&lt;/th&gt;&lt;th&gt;Authority&lt;/th&gt;&lt;th&gt;Position and key documents&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;Belgium&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;&lt;strong&gt;CCB&lt;/strong&gt; (Centre for Cybersecurity Belgium)&lt;/td&gt;&lt;td&gt;Co-authored the Cyber Security Coalition white paper series &lt;em&gt;Preparing for the Quantum Era&lt;/em&gt; (2025)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;/td&gt;&lt;td&gt;&lt;strong&gt;BIPT&lt;/strong&gt; (telecom regulator)&lt;/td&gt;&lt;td&gt;&lt;em&gt;Communication of 13 January 2026 on minimum requirements and roadmap for the post-quantum transition in the telecommunications sector&lt;/em&gt;: inventory, risk analysis and a phased roadmap for Belgian operators&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;Germany&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;&lt;strong&gt;BSI&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;Technical Guideline &lt;strong&gt;TR-02102-1&lt;/strong&gt; (version 2026-01) recommends &lt;strong&gt;hybrid&lt;/strong&gt; key establishment with ML-KEM-768/1024 or FrodoKEM-976/1344. Classical-only key agreement is no longer recommended after end of 2031 (end of 2030 for the most sensitive applications). In 2026 BSI advised against Classic McEliece for new developments. IPsec specifics are in TR-02102-3.&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;France&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;&lt;strong&gt;ANSSI&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;Insists on &lt;strong&gt;hybridation&lt;/strong&gt; wherever quantum protection is needed (position papers 2022, 2023; FAQ Oct 2025). Published the technical guide &lt;em&gt;&lt;strong&gt;Transition post-quantique d’IPsec&lt;/strong&gt;&lt;/em&gt; (2 Feb 2026), alongside guides for TLS 1.3 and SSHv2, and &lt;em&gt;ANSSI views on crypto agility&lt;/em&gt; (Jan 2026). Announced in June 2026 that from 2027 it will no longer certify products without PQC.&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;Netherlands&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;&lt;strong&gt;AIVD&lt;/strong&gt;, with CWI and TNO&lt;/td&gt;&lt;td&gt;&lt;em&gt;The PQC Migration Handbook&lt;/em&gt;, 2nd edition (Dec 2024), plus the PQChoiceAssistant tool&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;United Kingdom&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;&lt;strong&gt;NCSC&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;&lt;em&gt;Timelines for migration to PQC&lt;/em&gt;: discovery and plan &lt;strong&gt;by 2028&lt;/strong&gt;, highest-priority migrations &lt;strong&gt;by 2031&lt;/strong&gt;, full migration &lt;strong&gt;by 2035&lt;/strong&gt;. Treats hybrids as an interim measure; ML-KEM-768 as the general-purpose choice.&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;p&gt;At G7 level, the Cybersecurity Working Group (Canadian presidency 2025, French presidency 2026, with the European Commission and ENISA as guests) issued a joint statement on PQC migration in 2025 and a call to action in September 2026.&lt;/p&gt;
&lt;h4 id=&quot;australia&quot;&gt;Australia&lt;/h4&gt;
&lt;p&gt;The &lt;strong&gt;Australian Signals Directorate (ASD)&lt;/strong&gt;, through its Australian Cyber Security Centre, publishes the &lt;strong&gt;Information Security Manual (ISM)&lt;/strong&gt;. The ISM’s &lt;em&gt;Guidelines for Cryptography&lt;/em&gt; define the ASD-approved algorithms.&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Deadline:&lt;/strong&gt; ASD recommends ceasing the use of traditional asymmetric cryptography (RSA, DH, ECDH, ECDSA) by the &lt;strong&gt;end of 2030&lt;/strong&gt;, five years earlier than NIST and the UK. Traditional algorithms are not approved beyond 2030.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Milestones&lt;/strong&gt; (&lt;em&gt;Planning for post-quantum cryptography&lt;/em&gt;, last updated September 2025): a refined transition plan by end of 2026, migration of critical systems started by end of 2028, and the transition complete by end of 2030.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Algorithms:&lt;/strong&gt; the December 2024 ISM update approved ML-KEM and ML-DSA at the highest parameter sets: &lt;strong&gt;ML-KEM-1024&lt;/strong&gt; and &lt;strong&gt;ML-DSA-87&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Hybrids:&lt;/strong&gt; ASD &lt;strong&gt;does not recommend, but does not prohibit&lt;/strong&gt;, PQ/T hybrid schemes. This is a notable contrast with BSI and ANSSI.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Method:&lt;/strong&gt; the &lt;em&gt;LATICE&lt;/em&gt; framework, five high-level phases that run from locating and inventorying traditional asymmetric cryptography to implementing PQC and engaging vendors and stakeholders.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;What this means for an IPsec design, where several regimes must be met at once, is covered in the &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#8-designing-for-several-regimes&quot;&gt;IPsec article&lt;/a&gt;.&lt;/p&gt;
&lt;h3 id=&quot;24-nsa-suite-b-cnsa-10-and-cnsa-20&quot;&gt;2.4 NSA: Suite B, CNSA 1.0 and CNSA 2.0&lt;/h3&gt;
&lt;p&gt;NSA publishes the commercial algorithm suites that US &lt;strong&gt;National Security Systems&lt;/strong&gt; (NSS) must use; CNSSP 15 makes them binding. They matter far beyond the US government, because vendors build their products to them.&lt;/p&gt;
&lt;h4 id=&quot;algorithm-suites&quot;&gt;Algorithm suites&lt;/h4&gt;















































&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Function&lt;/th&gt;&lt;th&gt;Suite B (2005)&lt;/th&gt;&lt;th&gt;CNSA 1.0 (2015/16)&lt;/th&gt;&lt;th&gt;CNSA 2.0 (2022)&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;Encryption&lt;/td&gt;&lt;td&gt;AES-128, AES-256&lt;/td&gt;&lt;td&gt;AES-256&lt;/td&gt;&lt;td&gt;AES-256&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Key establishment&lt;/td&gt;&lt;td&gt;ECDH P-256, P-384&lt;/td&gt;&lt;td&gt;ECDH P-384; DH ≥ 3072; RSA ≥ 3072&lt;/td&gt;&lt;td&gt;&lt;strong&gt;ML-KEM-1024&lt;/strong&gt;&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Signatures&lt;/td&gt;&lt;td&gt;ECDSA P-256, P-384&lt;/td&gt;&lt;td&gt;ECDSA P-384; RSA ≥ 3072&lt;/td&gt;&lt;td&gt;&lt;strong&gt;ML-DSA-87&lt;/strong&gt;&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Software / firmware signing&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;(as above)&lt;/td&gt;&lt;td&gt;&lt;strong&gt;LMS, XMSS&lt;/strong&gt; (SP 800-208; LMS SHA-256/192 recommended)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Hashing&lt;/td&gt;&lt;td&gt;SHA-256, SHA-384&lt;/td&gt;&lt;td&gt;SHA-384&lt;/td&gt;&lt;td&gt;SHA-384 or SHA-512 (SHA3-384/512 only for internal hardware functions)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;IPsec profile&lt;/td&gt;&lt;td&gt;RFC 6379 / RFC 6380 (2011), &lt;strong&gt;Historic&lt;/strong&gt; since RFC 8423 (2018)&lt;/td&gt;&lt;td&gt;&lt;strong&gt;RFC 9206&lt;/strong&gt; (2022, Independent Submission)&lt;/td&gt;&lt;td&gt;&lt;code&gt;draft-guthrie-cnsa2-ipsec-profile-05&lt;/code&gt; (Oct 2026, individual)&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;p&gt;RFC 9206 defines three CNSA 1.0 IPsec suites: CNSA-GCM-256-ECDH-384, CNSA-GCM-256-DH-3072 and CNSA-GCM-256-DH-4096. They all use IKEv2, AES-256-GCM, and PRF_HMAC_SHA_512 (SHA-512 because of availability, although SHA-384 is the CNSA hash).&lt;/p&gt;
&lt;h4 id=&quot;cnsa-20-and-ikev2&quot;&gt;CNSA 2.0 and IKEv2&lt;/h4&gt;
&lt;p&gt;(Details for IPsec, including the CNSA IPsec profiles, are in the &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#7-cnsa-20-and-ipsec&quot;&gt;IPsec article&lt;/a&gt;.) In its CNSA 2.0 FAQ, NSA notes that ML-KEM-1024 cannot simply replace the CNSA 1.0 key exchange in IKEv2 because of message size. It expects to use the IETF approach: a smaller initial key exchange followed by a larger, encrypted one. Concretely, CNSA 1.0 key establishment in IKE_SA_INIT is fortified by ML-KEM-1024 in an additional exchange (RFC 9370). NSA does not &lt;em&gt;require&lt;/em&gt; hybrids. It rejects non-standard hybrids on NSS, but it does mention PPKs (RFC 8784) as possibly appropriate for specialised cases.&lt;/p&gt;
&lt;h4 id=&quot;timeline&quot;&gt;Timeline&lt;/h4&gt;
 &lt;figure class=&quot;viz&quot;&gt; &lt;figcaption&gt;NSA CNSA 2.0 timeline&lt;/figcaption&gt; &lt;p class=&quot;viz-sub&quot;&gt;Bars: from &quot;support &amp;amp; prefer CNSA 2.0&quot; to &quot;use exclusively&quot; (2022 category table). Dashed lines: CNSSP 15 milestones used by NSA since FAQ v2.1. Grey hatching marks the past. Hover for details.&lt;/p&gt; &lt;svg viewBox=&quot;0 0 680 240&quot; role=&quot;img&quot; aria-label=&quot;Gantt chart of CNSA 2.0 transition by product category with CNSSP 15 milestones&quot;&gt; &lt;defs&gt; &lt;pattern id=&quot;hatch-cnsa&quot; patternUnits=&quot;userSpaceOnUse&quot; width=&quot;6&quot; height=&quot;6&quot; patternTransform=&quot;rotate(45)&quot;&gt; &lt;rect width=&quot;6&quot; height=&quot;6&quot; fill=&quot;#f1f0ec&quot;&gt;&lt;/rect&gt; &lt;line x1=&quot;0&quot; y1=&quot;0&quot; x2=&quot;0&quot; y2=&quot;6&quot; stroke=&quot;#cfcdc6&quot; stroke-width=&quot;2&quot;&gt;&lt;/line&gt; &lt;/pattern&gt; &lt;/defs&gt; &lt;rect x=&quot;190&quot; y=&quot;28&quot; width=&quot;100.35683699686433&quot; height=&quot;190&quot; fill=&quot;url(#hatch-cnsa)&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;194&quot; y=&quot;214&quot; font-size=&quot;10&quot;&gt;past&lt;/text&gt; &lt;g&gt; &lt;line x1=&quot;210.9734513274337&quot; x2=&quot;210.9734513274337&quot; y1=&quot;28&quot; y2=&quot;218&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;210.9734513274337&quot; y=&quot;234&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2025&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;252.92035398230115&quot; x2=&quot;252.92035398230115&quot; y1=&quot;28&quot; y2=&quot;218&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;252.92035398230115&quot; y=&quot;234&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2026&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;294.86725663716857&quot; x2=&quot;294.86725663716857&quot; y1=&quot;28&quot; y2=&quot;218&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;294.86725663716857&quot; y=&quot;234&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2027&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;336.814159292036&quot; x2=&quot;336.814159292036&quot; y1=&quot;28&quot; y2=&quot;218&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;336.814159292036&quot; y=&quot;234&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2028&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;378.7610619469034&quot; x2=&quot;378.7610619469034&quot; y1=&quot;28&quot; y2=&quot;218&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;378.7610619469034&quot; y=&quot;234&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2029&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;420.70796460177087&quot; x2=&quot;420.70796460177087&quot; y1=&quot;28&quot; y2=&quot;218&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;420.70796460177087&quot; y=&quot;234&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2030&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;462.6548672566383&quot; x2=&quot;462.6548672566383&quot; y1=&quot;28&quot; y2=&quot;218&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;462.6548672566383&quot; y=&quot;234&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2031&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;504.6017699115057&quot; x2=&quot;504.6017699115057&quot; y1=&quot;28&quot; y2=&quot;218&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;504.6017699115057&quot; y=&quot;234&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2032&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;546.5486725663732&quot; x2=&quot;546.5486725663732&quot; y1=&quot;28&quot; y2=&quot;218&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;546.5486725663732&quot; y=&quot;234&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2033&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;588.4955752212406&quot; x2=&quot;588.4955752212406&quot; y1=&quot;28&quot; y2=&quot;218&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;588.4955752212406&quot; y=&quot;234&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2034&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;630.442477876108&quot; x2=&quot;630.442477876108&quot; y1=&quot;28&quot; y2=&quot;218&quot; stroke=&quot;var(--grid)&quot; stroke-width=&quot;1&quot;&gt;&lt;/line&gt; &lt;text x=&quot;630.442477876108&quot; y=&quot;234&quot; text-anchor=&quot;middle&quot; font-size=&quot;11&quot;&gt;2035&lt;/text&gt; &lt;/g&gt; &lt;g&gt; &lt;line x1=&quot;294.86725663716857&quot; x2=&quot;294.86725663716857&quot; y1=&quot;22&quot; y2=&quot;218&quot; stroke=&quot;var(--series-2)&quot; stroke-width=&quot;2&quot; stroke-dasharray=&quot;4 3&quot;&gt;&lt;/line&gt; &lt;rect x=&quot;287.86725663716857&quot; y=&quot;22&quot; width=&quot;14&quot; height=&quot;196&quot; fill=&quot;transparent&quot; tabindex=&quot;0&quot; data-tip=&quot;CNSSP 15: all new NSS acquisitions CNSA 2.0-compliant from 1 Jan 2027&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;294.86725663716857&quot; y=&quot;16&quot; text-anchor=&quot;middle&quot; font-size=&quot;10&quot; class=&quot;t-primary&quot;&gt;New acquisitions&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;462.23539823008997&quot; x2=&quot;462.23539823008997&quot; y1=&quot;22&quot; y2=&quot;218&quot; stroke=&quot;var(--series-2)&quot; stroke-width=&quot;2&quot; stroke-dasharray=&quot;4 3&quot;&gt;&lt;/line&gt; &lt;rect x=&quot;455.23539823008997&quot; y=&quot;22&quot; width=&quot;14&quot; height=&quot;196&quot; fill=&quot;transparent&quot; tabindex=&quot;0&quot; data-tip=&quot;CNSSP 15: equipment that cannot support CNSA 2.0 phased out by 31 Dec 2030&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;462.23539823008997&quot; y=&quot;28&quot; text-anchor=&quot;middle&quot; font-size=&quot;10&quot; class=&quot;t-primary&quot;&gt;Phase-out&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;504.1823008849574&quot; x2=&quot;504.1823008849574&quot; y1=&quot;22&quot; y2=&quot;218&quot; stroke=&quot;var(--series-2)&quot; stroke-width=&quot;2&quot; stroke-dasharray=&quot;4 3&quot;&gt;&lt;/line&gt; &lt;rect x=&quot;497.1823008849574&quot; y=&quot;22&quot; width=&quot;14&quot; height=&quot;196&quot; fill=&quot;transparent&quot; tabindex=&quot;0&quot; data-tip=&quot;CNSSP 15: CNSA 2.0 algorithms mandated by 31 Dec 2031&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;504.1823008849574&quot; y=&quot;16&quot; text-anchor=&quot;middle&quot; font-size=&quot;10&quot; class=&quot;t-primary&quot;&gt;Mandated&lt;/text&gt; &lt;/g&gt;&lt;g&gt; &lt;line x1=&quot;630.442477876108&quot; x2=&quot;630.442477876108&quot; y1=&quot;22&quot; y2=&quot;218&quot; stroke=&quot;var(--series-2)&quot; stroke-width=&quot;2&quot; stroke-dasharray=&quot;4 3&quot;&gt;&lt;/line&gt; &lt;rect x=&quot;623.442477876108&quot; y=&quot;22&quot; width=&quot;14&quot; height=&quot;196&quot; fill=&quot;transparent&quot; tabindex=&quot;0&quot; data-tip=&quot;NSM-10 goal: all NSS quantum-resistant by 2035&quot;&gt;&lt;/rect&gt; &lt;text x=&quot;630.442477876108&quot; y=&quot;28&quot; text-anchor=&quot;middle&quot; font-size=&quot;10&quot; class=&quot;t-primary&quot;&gt;All NSS QR&lt;/text&gt; &lt;/g&gt; &lt;g&gt; &lt;text x=&quot;182&quot; y=&quot;51&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot; font-weight=&quot;400&quot;&gt;Software / firmware signing&lt;/text&gt; &lt;rect x=&quot;210.9734513274337&quot; y=&quot;41&quot; width=&quot;209.73451327433716&quot; height=&quot;14&quot; rx=&quot;4&quot; fill=&quot;var(--series-1)&quot; opacity=&quot;0.55&quot; tabindex=&quot;0&quot; data-tip=&quot;Software / firmware signing
Support &amp;#38; prefer by 2025
Exclusively by 2030&quot;&gt;&lt;/rect&gt; &lt;circle cx=&quot;210.9734513274337&quot; cy=&quot;48&quot; r=&quot;4&quot; fill=&quot;var(--surface-1)&quot; stroke=&quot;var(--series-1)&quot; stroke-width=&quot;2&quot;&gt;&lt;/circle&gt; &lt;circle cx=&quot;420.70796460177087&quot; cy=&quot;48&quot; r=&quot;4.5&quot; fill=&quot;var(--series-1)&quot; stroke=&quot;var(--surface-1)&quot; stroke-width=&quot;1.5&quot;&gt;&lt;/circle&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;182&quot; y=&quot;81&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot; font-weight=&quot;400&quot;&gt;Browsers, servers, cloud&lt;/text&gt; &lt;rect x=&quot;210.9734513274337&quot; y=&quot;71&quot; width=&quot;335.57522123893943&quot; height=&quot;14&quot; rx=&quot;4&quot; fill=&quot;var(--series-1)&quot; opacity=&quot;0.55&quot; tabindex=&quot;0&quot; data-tip=&quot;Browsers, servers, cloud
Support &amp;#38; prefer by 2025
Exclusively by 2033&quot;&gt;&lt;/rect&gt; &lt;circle cx=&quot;210.9734513274337&quot; cy=&quot;78&quot; r=&quot;4&quot; fill=&quot;var(--surface-1)&quot; stroke=&quot;var(--series-1)&quot; stroke-width=&quot;2&quot;&gt;&lt;/circle&gt; &lt;circle cx=&quot;546.5486725663732&quot; cy=&quot;78&quot; r=&quot;4.5&quot; fill=&quot;var(--series-1)&quot; stroke=&quot;var(--surface-1)&quot; stroke-width=&quot;1.5&quot;&gt;&lt;/circle&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;182&quot; y=&quot;111&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot; font-weight=&quot;700&quot;&gt;Networking (VPNs, routers)&lt;/text&gt; &lt;rect x=&quot;252.92035398230115&quot; y=&quot;101&quot; width=&quot;167.78761061946972&quot; height=&quot;14&quot; rx=&quot;4&quot; fill=&quot;var(--series-1)&quot; opacity=&quot;1&quot; tabindex=&quot;0&quot; data-tip=&quot;Networking (VPNs, routers)
Support &amp;#38; prefer by 2026
Exclusively by 2030&quot;&gt;&lt;/rect&gt; &lt;circle cx=&quot;252.92035398230115&quot; cy=&quot;108&quot; r=&quot;4&quot; fill=&quot;var(--surface-1)&quot; stroke=&quot;var(--series-1)&quot; stroke-width=&quot;2&quot;&gt;&lt;/circle&gt; &lt;circle cx=&quot;420.70796460177087&quot; cy=&quot;108&quot; r=&quot;4.5&quot; fill=&quot;var(--series-1)&quot; stroke=&quot;var(--surface-1)&quot; stroke-width=&quot;1.5&quot;&gt;&lt;/circle&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;182&quot; y=&quot;141&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot; font-weight=&quot;400&quot;&gt;Operating systems&lt;/text&gt; &lt;rect x=&quot;294.86725663716857&quot; y=&quot;131&quot; width=&quot;251.6814159292046&quot; height=&quot;14&quot; rx=&quot;4&quot; fill=&quot;var(--series-1)&quot; opacity=&quot;0.55&quot; tabindex=&quot;0&quot; data-tip=&quot;Operating systems
Support &amp;#38; prefer by 2027
Exclusively by 2033&quot;&gt;&lt;/rect&gt; &lt;circle cx=&quot;294.86725663716857&quot; cy=&quot;138&quot; r=&quot;4&quot; fill=&quot;var(--surface-1)&quot; stroke=&quot;var(--series-1)&quot; stroke-width=&quot;2&quot;&gt;&lt;/circle&gt; &lt;circle cx=&quot;546.5486725663732&quot; cy=&quot;138&quot; r=&quot;4.5&quot; fill=&quot;var(--series-1)&quot; stroke=&quot;var(--surface-1)&quot; stroke-width=&quot;1.5&quot;&gt;&lt;/circle&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;182&quot; y=&quot;171&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot; font-weight=&quot;400&quot;&gt;Niche equipment&lt;/text&gt; &lt;rect x=&quot;420.70796460177087&quot; y=&quot;161&quot; width=&quot;125.8407079646023&quot; height=&quot;14&quot; rx=&quot;4&quot; fill=&quot;var(--series-1)&quot; opacity=&quot;0.55&quot; tabindex=&quot;0&quot; data-tip=&quot;Niche equipment
Support &amp;#38; prefer by 2030
Exclusively by 2033&quot;&gt;&lt;/rect&gt; &lt;circle cx=&quot;420.70796460177087&quot; cy=&quot;168&quot; r=&quot;4&quot; fill=&quot;var(--surface-1)&quot; stroke=&quot;var(--series-1)&quot; stroke-width=&quot;2&quot;&gt;&lt;/circle&gt; &lt;circle cx=&quot;546.5486725663732&quot; cy=&quot;168&quot; r=&quot;4.5&quot; fill=&quot;var(--series-1)&quot; stroke=&quot;var(--surface-1)&quot; stroke-width=&quot;1.5&quot;&gt;&lt;/circle&gt; &lt;/g&gt;&lt;g&gt; &lt;text x=&quot;182&quot; y=&quot;201&quot; text-anchor=&quot;end&quot; font-size=&quot;12&quot; class=&quot;t-primary&quot; font-weight=&quot;400&quot;&gt;Custom &amp;amp; legacy&lt;/text&gt; &lt;rect x=&quot;540.2566371681392&quot; y=&quot;191&quot; width=&quot;6.29203539823402&quot; height=&quot;14&quot; rx=&quot;4&quot; fill=&quot;var(--series-1)&quot; opacity=&quot;0.55&quot; tabindex=&quot;0&quot; data-tip=&quot;Custom &amp;#38; legacy
Update or replace by 2033&quot;&gt;&lt;/rect&gt;  &lt;circle cx=&quot;546.5486725663732&quot; cy=&quot;198&quot; r=&quot;4.5&quot; fill=&quot;var(--series-1)&quot; stroke=&quot;var(--surface-1)&quot; stroke-width=&quot;1.5&quot;&gt;&lt;/circle&gt; &lt;/g&gt; &lt;line x1=&quot;290.35683699686433&quot; x2=&quot;290.35683699686433&quot; y1=&quot;28&quot; y2=&quot;218&quot; stroke=&quot;var(--text-primary)&quot; stroke-width=&quot;1.5&quot;&gt;&lt;/line&gt; &lt;text x=&quot;286.35683699686433&quot; y=&quot;214&quot; text-anchor=&quot;end&quot; font-size=&quot;10&quot; class=&quot;t-primary&quot;&gt;today&lt;/text&gt; &lt;/svg&gt; &lt;details&gt; &lt;summary&gt;Show data as table&lt;/summary&gt; &lt;table&gt; &lt;thead&gt;&lt;tr&gt;&lt;th&gt;Category&lt;/th&gt;&lt;th&gt;Support &amp;amp; prefer by&lt;/th&gt;&lt;th&gt;Exclusively by&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt; &lt;tbody&gt; &lt;tr&gt;&lt;td&gt;Software / firmware signing&lt;/td&gt;&lt;td&gt;2025&lt;/td&gt;&lt;td&gt;2030&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Browsers, servers, cloud&lt;/td&gt;&lt;td&gt;2025&lt;/td&gt;&lt;td&gt;2033&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Networking (VPNs, routers)&lt;/td&gt;&lt;td&gt;2026&lt;/td&gt;&lt;td&gt;2030&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Operating systems&lt;/td&gt;&lt;td&gt;2027&lt;/td&gt;&lt;td&gt;2033&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Niche equipment&lt;/td&gt;&lt;td&gt;2030&lt;/td&gt;&lt;td&gt;2033&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Custom &amp;amp; legacy&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;update or replace by 2033&lt;/td&gt;&lt;/tr&gt; &lt;/tbody&gt; &lt;/table&gt; &lt;/details&gt; &lt;/figure&gt;
&lt;p&gt;In 2022 NSA published the category table shown as bars above. &lt;strong&gt;Traditional networking equipment (VPNs, routers)&lt;/strong&gt; has one of the tightest schedules: support and prefer CNSA 2.0 by 2026, use it exclusively by 2030. Since the December 2024 version of the FAQ (v2.1), NSA has framed the transition around the CNSSP 15 milestones (the dashed lines) rather than the category table. The latest FAQ is v2.2 (October 2026). For IPsec vendors and VPN operators in that ecosystem, &lt;strong&gt;1 January 2027 is the procurement gate&lt;/strong&gt;. VPN equipment without ML-KEM-1024 support will be increasingly hard to sell or deploy there.&lt;/p&gt;
&lt;h4 id=&quot;why-cnsa-20-accepts-only-ml-kem-1024&quot;&gt;Why CNSA 2.0 accepts only ML-KEM-1024&lt;/h4&gt;
&lt;p&gt;The CNSA 2.0 FAQ lists &lt;strong&gt;ML-KEM-1024 “for all classification levels”&lt;/strong&gt; (and likewise ML-DSA-87, AES-256 and SHA-384/512). NSA does not publish a detailed rationale beyond that, but the choice follows the same logic as its earlier suites and its other choices:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;One parameter set for everything up to TOP SECRET.&lt;/strong&gt; CNSA 1.0 already dropped Suite B’s two levels (P-256/AES-128 for SECRET, P-384/AES-256 for TOP SECRET) in favour of the strongest option only. A single set simplifies product certification, configuration and interoperability across all national security systems.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Maximum security margin for long-lived secrets.&lt;/strong&gt; Classified information can need protection for decades. Category 5 leaves the most room if cryptanalysis of lattices improves.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;No hybrid required, so the PQ part must stand alone.&lt;/strong&gt; Because NSA does not require a traditional component, ML-KEM alone has to carry the security. That argues for the largest parameter set.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;The cost is acceptable.&lt;/strong&gt; ML-KEM-1024 is still fast; its main cost is size (1,568 bytes), which IKEv2 handles with RFC 9242 and RFC 7383.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;Australia’s ASD made the same choice (ML-KEM-1024, ML-DSA-87). For most other uses, NIST, the IETF and the UK NCSC consider ML-KEM-768 (category 3) sufficient.&lt;/p&gt;
&lt;h4 id=&quot;cnsa-20-vs-nist-sp-800-227-on-pqt-hybrids&quot;&gt;CNSA 2.0 vs. NIST SP 800-227 on PQ/T hybrids&lt;/h4&gt;
&lt;p&gt;Is there a conflict between NSA’s position and NIST’s KEM guidance? In short: &lt;strong&gt;no direct contradiction, but different emphasis.&lt;/strong&gt;&lt;/p&gt;






























&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Aspect&lt;/th&gt;&lt;th&gt;NIST SP 800-227 (Sep 2025)&lt;/th&gt;&lt;th&gt;NSA CNSA 2.0 (FAQ v2.1/v2.2)&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;Hybrids&lt;/td&gt;&lt;td&gt;Accommodates them. It gives guidance on combining the shared secrets of several KEMs or key-establishment schemes into one key with an approved key-derivation method.&lt;/td&gt;&lt;td&gt;Does &lt;strong&gt;not require&lt;/strong&gt; them. Accepts standards-based hybrids (and plans one for IKEv2), rejects non-standard hybrids on NSS.&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Which PQ component counts&lt;/td&gt;&lt;td&gt;Any approved KEM parameter set (ML-KEM-512, -768, -1024)&lt;/td&gt;&lt;td&gt;Only &lt;strong&gt;ML-KEM-1024&lt;/strong&gt;&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Role of the traditional component&lt;/td&gt;&lt;td&gt;Part of the combined key; the result is approved as long as an approved component is combined in an approved way&lt;/td&gt;&lt;td&gt;Adds no CNSA 2.0 credit; in IKEv2 it is kept mainly for practical reasons (message size, interoperability)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;Long term&lt;/td&gt;&lt;td&gt;NIST IR 8547: quantum-vulnerable algorithms disallowed after 2035&lt;/td&gt;&lt;td&gt;CNSA 2.0 algorithms mandated by end of 2031&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;p&gt;For an IPsec designer the practical outcome is the same: &lt;strong&gt;ECDH P-384 + ML-KEM-1024 via RFC 9370&lt;/strong&gt; satisfies SP 800-227, CNSA 2.0, BSI and ANSSI at once. One open point remains: how long a traditional component may stay inside a hybrid after the 2031 and 2035 dates. Check the final CNSA 2.0 IPsec profile (&lt;code&gt;draft-guthrie-cnsa2-ipsec-profile&lt;/code&gt;) when it is published.&lt;/p&gt;
&lt;hr/&gt;
&lt;h2 id=&quot;3-concepts-pure-pq-hybrids-hybrid-signatures-and-crypto-agility&quot;&gt;3. Concepts: pure PQ, hybrids, hybrid signatures and crypto agility&lt;/h2&gt;
&lt;h3 id=&quot;31-terminology&quot;&gt;3.1 Terminology&lt;/h3&gt;
&lt;p&gt;The IETF vocabulary is defined in &lt;strong&gt;RFC 9794&lt;/strong&gt; (&lt;em&gt;Terminology for Post-Quantum Traditional Hybrid Schemes&lt;/em&gt;, June 2025):&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Traditional algorithm&lt;/strong&gt;: an asymmetric algorithm based on factorisation or (elliptic-curve) discrete logarithms (RSA, DH, ECDH, ECDSA, EdDSA).&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Post-quantum algorithm&lt;/strong&gt;: an asymmetric algorithm designed to be secure against both classical and quantum computers (ML-KEM, ML-DSA, SLH-DSA, …). RFC 9794 prefers “post-quantum” over “quantum-resistant” or “quantum-safe”, because the label describes a design goal, not a guarantee.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Multi-algorithm scheme&lt;/strong&gt;: a scheme with more than one component algorithm serving the same purpose.&lt;/li&gt;
&lt;/ul&gt;
&lt;h3 id=&quot;32-pure-pq-pqt-hybrid-and-pqpq-hybrid&quot;&gt;3.2 Pure PQ, PQ/T hybrid and PQ/PQ hybrid&lt;/h3&gt;

























&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Approach&lt;/th&gt;&lt;th&gt;Definition&lt;/th&gt;&lt;th&gt;IKEv2 example (&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#36-what-each-mode-needs&quot;&gt;details&lt;/a&gt;)&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;Pure PQ&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;Only post-quantum components&lt;/td&gt;&lt;td&gt;ML-KEM alone in IKE_SA_INIT. &lt;code&gt;draft-ietf-ipsecme-ikev2-mlkem&lt;/code&gt; allows ML-KEM-512 there; ML-KEM-768/1024 SHOULD NOT be used in IKE_SA_INIT over UDP when the path MTU is unknown, because IKE_SA_INIT cannot be fragmented.&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;PQ/T hybrid&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;At least one traditional and at least one PQ component; designed so that an attacker must break &lt;em&gt;all&lt;/em&gt; components&lt;/td&gt;&lt;td&gt;ECDH (P-384 or Curve25519) in IKE_SA_INIT plus ML-KEM in IKE_INTERMEDIATE via RFC 9370&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;PQ/PQ hybrid&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;Only PQ components, preferably based on different hard problems&lt;/td&gt;&lt;td&gt;ML-KEM plus FrodoKEM (&lt;code&gt;draft-ietf-ipsecme-hybrid-kem-ikev2-frodo&lt;/code&gt;) or plus HQC (&lt;code&gt;draft-reddy-ipsecme-ikev2-hqc&lt;/code&gt;, individual)&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;p&gt;RFC 9794 also separates &lt;strong&gt;composite&lt;/strong&gt; from &lt;strong&gt;non-composite&lt;/strong&gt; constructions:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Composite&lt;/strong&gt;: the protocol stays the same and the cryptographic element changes. One “algorithm” internally contains several components, for example one composite public key and one composite signature.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Non-composite&lt;/strong&gt;: each component keeps its own format, and the protocol changes to carry several of them. &lt;strong&gt;IKEv2 with RFC 9370 is RFC 9794’s own example of a non-composite construction.&lt;/strong&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;h4 id=&quot;why-pqt-hybrid-rather-than-pq-only&quot;&gt;Why PQ/T hybrid rather than PQ-only?&lt;/h4&gt;
&lt;p&gt;ML-KEM and ML-DSA are young compared with ECDH and ECDSA. The underlying lattice problems have been studied for far less time, and implementations of new algorithms have bugs (timing side channels such as KyberSlash were found and fixed in early Kyber/ML-KEM implementations in 2024). A hybrid stays secure as long as &lt;strong&gt;one&lt;/strong&gt; component holds:&lt;/p&gt;



































&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Scenario&lt;/th&gt;&lt;th&gt;PQ-only&lt;/th&gt;&lt;th&gt;PQ/T hybrid&lt;/th&gt;&lt;th&gt;Traditional only&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;No CRQC, PQ algorithm sound&lt;/td&gt;&lt;td&gt;Secure&lt;/td&gt;&lt;td&gt;Secure&lt;/td&gt;&lt;td&gt;Secure&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;No CRQC, &lt;strong&gt;PQ algorithm or implementation broken&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;&lt;strong&gt;Broken&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;Secure (traditional part holds)&lt;/td&gt;&lt;td&gt;Secure&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;CRQC exists&lt;/strong&gt;, PQ algorithm sound&lt;/td&gt;&lt;td&gt;Secure&lt;/td&gt;&lt;td&gt;Secure (PQ part holds)&lt;/td&gt;&lt;td&gt;&lt;strong&gt;Broken&lt;/strong&gt;&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;CRQC exists &lt;strong&gt;and&lt;/strong&gt; PQ algorithm broken&lt;/td&gt;&lt;td&gt;Broken&lt;/td&gt;&lt;td&gt;Broken&lt;/td&gt;&lt;td&gt;Broken&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;p&gt;A hybrid therefore protects against the two &lt;em&gt;separate&lt;/em&gt; risks: a quantum computer, and a flaw in the new algorithm. Once a CRQC exists, the traditional component no longer adds anything, and only the PQ component protects the session. That is why hybrids are seen as a transition measure, and why the PQ component must be strong in its own right. Combined with crypto agility (&lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/#34-crypto-agility&quot;&gt;section 3.4&lt;/a&gt;), a hybrid lets you replace a broken PQ algorithm without losing protection in the meantime.&lt;/p&gt;
&lt;p&gt;Several European agencies (e.g. BSI and ANSSI) recommend PQ/T hybrids for the transition (&lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/#23-european-and-australian-authorities&quot;&gt;section 2.3&lt;/a&gt;). Australia’s ASD does not recommend them, but does not prohibit them either. NSA does not require hybrids for CNSA 2.0, but anticipates a hybrid-style construction in IKEv2 for practical reasons (&lt;a href=&quot;https://www.itv6.be/blog/post-quantum-cryptography/#cnsa-20-vs-nist-sp-800-227-on-pqt-hybrids&quot;&gt;section 2.4&lt;/a&gt;).&lt;/p&gt;
&lt;p&gt;A note on &lt;strong&gt;pre-shared keys&lt;/strong&gt;: RFC 8784 and RFC 9867 mix a &lt;strong&gt;Post-quantum Preshared Key (PPK)&lt;/strong&gt; into the IKEv2 key derivation (&lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#24-the-ppk-notifications&quot;&gt;PPK notifications&lt;/a&gt;). This is a &lt;em&gt;symmetric&lt;/em&gt; hedge against quantum attacks, not a PQ algorithm. It works today with any key exchange, provided you can distribute high-entropy PPKs out of band. RFC 8784 protects the Child SAs but not the initial IKE SA. RFC 9867 (November 2025) moves the PPK into IKE_INTERMEDIATE so that the initial IKE SA, including identities, is protected too. It also allows PPKs that change frequently, for example keys delivered by QKD.&lt;/p&gt;
&lt;h3 id=&quot;33-hybrid-signatures-parallel-serial-and-composite&quot;&gt;3.3 Hybrid signatures: parallel, serial and composite&lt;/h3&gt;
&lt;p&gt;For authentication there are three main ways to combine a traditional and a PQ signature. RFC 9794 names them, and &lt;strong&gt;RFC 9955&lt;/strong&gt; (&lt;em&gt;Hybrid Signature Spectrums&lt;/em&gt;, July 2026) analyses their security properties.&lt;/p&gt;





























&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Type&lt;/th&gt;&lt;th&gt;How it works&lt;/th&gt;&lt;th&gt;Properties&lt;/th&gt;&lt;th&gt;IKEv2 work&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;Parallel signatures&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;Two independent signatures over the same data, each with its own key and usually its own certificate (“multi-certificate authentication”, a “PQ/T parallel PKI” in RFC 9794 terms)&lt;/td&gt;&lt;td&gt;Simple and backward compatible; the components are &lt;strong&gt;separable&lt;/strong&gt;, so an attacker can strip one and present the other as a valid single signature&lt;/td&gt;&lt;td&gt;Individual drafts, e.g. &lt;code&gt;draft-hu-ipsecme-pqt-hybrid-auth-with-sep-cert&lt;/code&gt;, &lt;code&gt;draft-wang-ipsecme-multi-auth-ikev2-pq&lt;/code&gt;&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;Serial (nested) signatures&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;The second signature also covers the first: sig₂ = Sign₂(m ‖ sig₁)&lt;/td&gt;&lt;td&gt;Stripping the outer signature is detectable, stripping the inner one is not; order and verification logic matter&lt;/td&gt;&lt;td&gt;No dedicated IKEv2 specification&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;Composite signatures&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;One composite key, one composite signature and one algorithm identifier; internally ML-DSA plus a traditional algorithm, bound together by a domain separator&lt;/td&gt;&lt;td&gt;Looks like a single algorithm to the protocol; designed to resist stripping&lt;/td&gt;&lt;td&gt;&lt;strong&gt;RFC 10073&lt;/strong&gt; (composite ML-DSA, Nov 2026) + &lt;code&gt;draft-ietf-ipsecme-pq-hybrid-auth&lt;/code&gt;&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;p&gt;RFC 9955’s central concept is &lt;strong&gt;separability&lt;/strong&gt;: can an attacker turn a hybrid signature into a valid signature for only one component? Its &lt;em&gt;non-separability spectrum&lt;/em&gt; runs from fully separable (parallel) to weak and strong non-separability.&lt;/p&gt;
&lt;h3 id=&quot;34-crypto-agility&quot;&gt;3.4 Crypto agility&lt;/h3&gt;
&lt;p&gt;Crypto agility is the ability to change algorithms without redesigning the protocol or replacing the equipment (see RFC 7696, BCP 201). IKEv2 is agile by design: every algorithm is a negotiated transform with an IANA number, which is why ML-KEM fits in with only three new code points. In practice, agility also requires:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;an &lt;strong&gt;inventory&lt;/strong&gt; of where which algorithms and keys are used (VPN gateways, clients, PKI, HSMs);&lt;/li&gt;
&lt;li&gt;algorithms that are &lt;strong&gt;configurable&lt;/strong&gt; rather than hard-coded, and firmware that can be upgraded;&lt;/li&gt;
&lt;li&gt;room for &lt;strong&gt;larger messages&lt;/strong&gt;: IKE fragmentation (RFC 7383), IKE_INTERMEDIATE (RFC 9242), or TCP encapsulation (RFC 9329);&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;downgrade protection&lt;/strong&gt;. As long as traditional-only proposals are accepted, an active attacker with a CRQC could force them. Once all peers support ML-KEM, policy should refuse proposals without it. A general fix is in progress (&lt;code&gt;draft-ietf-ipsecme-ikev2-downgrade-prevention&lt;/code&gt;).&lt;/li&gt;
&lt;/ul&gt;
&lt;hr/&gt;
&lt;h2 id=&quot;4-security-categories-parameter-sets-and-sizes&quot;&gt;4. Security categories, parameter sets and sizes&lt;/h2&gt;
&lt;h3 id=&quot;41-the-five-categories&quot;&gt;4.1 The five categories&lt;/h3&gt;
&lt;p&gt;NIST defines five &lt;strong&gt;post-quantum security strength categories&lt;/strong&gt;. Each one is tied to the cost of attacking a well-understood symmetric primitive, under classical and quantum attack:&lt;/p&gt;





























&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Category&lt;/th&gt;&lt;th&gt;At least as hard to break as …&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;1&lt;/td&gt;&lt;td&gt;Exhaustive key search on AES-128&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;2&lt;/td&gt;&lt;td&gt;Collision search on SHA-256&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;3&lt;/td&gt;&lt;td&gt;Exhaustive key search on AES-192&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;4&lt;/td&gt;&lt;td&gt;Collision search on SHA-384&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;5&lt;/td&gt;&lt;td&gt;Exhaustive key search on AES-256&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;h3 id=&quot;42-parameter-sets-per-category&quot;&gt;4.2 Parameter sets per category&lt;/h3&gt;





























































&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Algorithm (standard)&lt;/th&gt;&lt;th&gt;Cat. 1&lt;/th&gt;&lt;th&gt;Cat. 2&lt;/th&gt;&lt;th&gt;Cat. 3&lt;/th&gt;&lt;th&gt;Cat. 4&lt;/th&gt;&lt;th&gt;Cat. 5&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;ML-KEM&lt;/strong&gt; (FIPS 203)&lt;/td&gt;&lt;td&gt;ML-KEM-512&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;ML-KEM-768&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;ML-KEM-1024&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;ML-DSA&lt;/strong&gt; (FIPS 204)&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;ML-DSA-44&lt;/td&gt;&lt;td&gt;ML-DSA-65&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;ML-DSA-87&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;SLH-DSA&lt;/strong&gt; (FIPS 205)&lt;/td&gt;&lt;td&gt;128s, 128f&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;192s, 192f&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;256s, 256f&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;FN-DSA&lt;/strong&gt; (FIPS 206, not final)&lt;/td&gt;&lt;td&gt;FN-DSA-512&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;FN-DSA-1024&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;HQC&lt;/strong&gt; (expected FIPS 207, not final)&lt;/td&gt;&lt;td&gt;HQC-1&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;HQC-3&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;HQC-5&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;FrodoKEM&lt;/strong&gt; (not NIST; IETF draft)&lt;/td&gt;&lt;td&gt;FrodoKEM-640&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;FrodoKEM-976&lt;/td&gt;&lt;td&gt;–&lt;/td&gt;&lt;td&gt;FrodoKEM-1344&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;p&gt;No current algorithm targets category 4. SLH-DSA comes in SHA2 and SHAKE variants for each parameter set; the &lt;strong&gt;s&lt;/strong&gt; variants have smaller signatures, the &lt;strong&gt;f&lt;/strong&gt; variants sign faster. In practice, &lt;strong&gt;ML-KEM-768&lt;/strong&gt; (category 3) is the common default, while CNSA 2.0 and the Australian ISM require category 5 (&lt;strong&gt;ML-KEM-1024&lt;/strong&gt;, &lt;strong&gt;ML-DSA-87&lt;/strong&gt;).&lt;/p&gt;
&lt;h3 id=&quot;43-key-and-signature-sizes&quot;&gt;4.3 Key and signature sizes&lt;/h3&gt;











































































&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Algorithm&lt;/th&gt;&lt;th&gt;Public key (bytes)&lt;/th&gt;&lt;th&gt;Ciphertext / signature (bytes)&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;X25519 (traditional, for comparison)&lt;/td&gt;&lt;td&gt;32&lt;/td&gt;&lt;td&gt;32&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ECDSA P-256 (traditional, for comparison)&lt;/td&gt;&lt;td&gt;64&lt;/td&gt;&lt;td&gt;≈ 64–72&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ECDSA P-384 (traditional, for comparison)&lt;/td&gt;&lt;td&gt;96&lt;/td&gt;&lt;td&gt;≈ 96–104&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ECDSA P-521 (traditional, for comparison)&lt;/td&gt;&lt;td&gt;132&lt;/td&gt;&lt;td&gt;≈ 132–139&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;RSA-3072 (traditional, for comparison)&lt;/td&gt;&lt;td&gt;384&lt;/td&gt;&lt;td&gt;384&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;RSA-4096 (traditional, for comparison)&lt;/td&gt;&lt;td&gt;512&lt;/td&gt;&lt;td&gt;512&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ML-KEM-512 / -768 / -1024&lt;/td&gt;&lt;td&gt;800 / 1,184 / 1,568&lt;/td&gt;&lt;td&gt;768 / 1,088 / 1,568&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;ML-DSA-44 / -65 / -87&lt;/td&gt;&lt;td&gt;1,312 / 1,952 / 2,592&lt;/td&gt;&lt;td&gt;2,420 / 3,309 / 4,627&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;SLH-DSA-128s / -128f&lt;/td&gt;&lt;td&gt;32&lt;/td&gt;&lt;td&gt;7,856 / 17,088&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;SLH-DSA-192s / -192f&lt;/td&gt;&lt;td&gt;48&lt;/td&gt;&lt;td&gt;16,224 / 35,664&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;SLH-DSA-256s / -256f&lt;/td&gt;&lt;td&gt;64&lt;/td&gt;&lt;td&gt;29,792 / 49,856&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;FN-DSA-512 / -1024 (not final)&lt;/td&gt;&lt;td&gt;897 / 1,793&lt;/td&gt;&lt;td&gt;≈ 666 / ≈ 1,280&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;FrodoKEM-640 / -976 / -1344&lt;/td&gt;&lt;td&gt;9,616 / 15,632 / 21,520&lt;/td&gt;&lt;td&gt;9,720 / 15,744 / 21,632&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;p&gt;PQ keys and signatures are one to three orders of magnitude larger than traditional ones. For protocols that run over UDP, like IKEv2, that is the main engineering challenge; see &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#6-key-exchange-and-signature-sizes-in-ikev2&quot;&gt;key exchange and signature sizes in IKEv2&lt;/a&gt;.&lt;/p&gt;
&lt;hr/&gt;
&lt;h2 id=&quot;5-ietf-work-beyond-ipsec-lamps-and-pquip&quot;&gt;5. IETF work beyond IPsec: lamps and pquip&lt;/h2&gt;
&lt;p&gt;The IPsec-specific work happens in the &lt;strong&gt;ipsecme&lt;/strong&gt; working group; see the &lt;a href=&quot;https://www.itv6.be/blog/pqc-in-ipsec/#4-the-ipsecme-work&quot;&gt;IPsec article&lt;/a&gt;. Two other working groups produce results that every protocol, including IPsec, depends on.&lt;/p&gt;
&lt;h3 id=&quot;51-lamps-limited-additional-mechanisms-for-pkix-and-smime&quot;&gt;5.1 lamps: Limited Additional Mechanisms for PKIX and SMIME&lt;/h3&gt;
&lt;p&gt;IKEv2 authentication relies on X.509 certificates, and LAMPS defines how PQ keys and signatures go into them.&lt;/p&gt;

























&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Document&lt;/th&gt;&lt;th&gt;Subject&lt;/th&gt;&lt;th&gt;Status&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;RFC 9881&lt;/td&gt;&lt;td&gt;ML-DSA in X.509 certificates and CRLs&lt;/td&gt;&lt;td&gt;Proposed Standard (Oct 2025)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;RFC 9909&lt;/td&gt;&lt;td&gt;SLH-DSA in X.509&lt;/td&gt;&lt;td&gt;Proposed Standard (Dec 2025)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;RFC 10073&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;Composite ML-DSA (18 combinations of ML-DSA-44/65/87 with RSA, ECDSA, Ed25519/Ed448)&lt;/td&gt;&lt;td&gt;&lt;strong&gt;Proposed Standard (Nov 2026)&lt;/strong&gt;, formerly &lt;code&gt;draft-ietf-lamps-pq-composite-sigs&lt;/code&gt;&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;h3 id=&quot;52-pquip-post-quantum-use-in-protocols&quot;&gt;5.2 pquip: Post-Quantum Use In Protocols&lt;/h3&gt;
&lt;p&gt;PQUIP does not define protocols. It produces cross-cutting guidance, published as Informational RFCs.&lt;/p&gt;








































&lt;table&gt;&lt;thead&gt;&lt;tr&gt;&lt;th&gt;Document&lt;/th&gt;&lt;th&gt;Subject&lt;/th&gt;&lt;th&gt;Status&lt;/th&gt;&lt;/tr&gt;&lt;/thead&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;RFC 9794&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;Terminology for PQ/T hybrid schemes&lt;/td&gt;&lt;td&gt;Informational (Jun 2025)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;RFC 9958&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;&lt;em&gt;Post-Quantum Cryptography for Engineers&lt;/em&gt;, formerly &lt;code&gt;draft-ietf-pquip-pqc-engineers&lt;/code&gt;&lt;/td&gt;&lt;td&gt;Informational (Jun 2026)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;strong&gt;RFC 9955&lt;/strong&gt;&lt;/td&gt;&lt;td&gt;Hybrid signature spectrums&lt;/td&gt;&lt;td&gt;Informational (Jul 2026)&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;code&gt;draft-ietf-pquip-hbs-state&lt;/code&gt;&lt;/td&gt;&lt;td&gt;Hash-based signatures: state and backup management (LMS/XMSS)&lt;/td&gt;&lt;td&gt;Active with the IESG&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;code&gt;draft-ietf-pquip-pqc-hsm-constrained&lt;/code&gt;&lt;/td&gt;&lt;td&gt;Adapting constrained devices (HSMs, embedded) for PQC&lt;/td&gt;&lt;td&gt;Active with the IESG&lt;/td&gt;&lt;/tr&gt;&lt;tr&gt;&lt;td&gt;&lt;code&gt;draft-prabel-pquip-pqc-overview-00&lt;/code&gt;&lt;/td&gt;&lt;td&gt;Overview of PQ algorithms (parameters, sizes, assumptions)&lt;/td&gt;&lt;td&gt;Individual draft, &lt;strong&gt;expired&lt;/strong&gt; Sep 2026; no formal standing&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;
&lt;p&gt;If you only read one document, read &lt;strong&gt;RFC 9958&lt;/strong&gt;. It explains the threat model, KEMs versus Diffie-Hellman, signature trade-offs, hybrids, key and signature sizes, and migration in engineering terms.&lt;/p&gt;
&lt;hr/&gt;
&lt;h2 id=&quot;6-sources&quot;&gt;6. Sources&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;IETF: RFCs and drafts&lt;/strong&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;a href=&quot;https://www.rfc-editor.org/info/rfc2026&quot;&gt;RFC 2026&lt;/a&gt;, &lt;a href=&quot;https://www.rfc-editor.org/info/rfc6410&quot;&gt;RFC 6410&lt;/a&gt;, &lt;a href=&quot;https://www.rfc-editor.org/info/rfc7127&quot;&gt;RFC 7127&lt;/a&gt;: IETF standards process&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.rfc-editor.org/info/rfc7696&quot;&gt;RFC 7696: Guidelines for Cryptographic Algorithm Agility (BCP 201)&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.rfc-editor.org/rfc/rfc9794.html&quot;&gt;RFC 9794: Terminology for PQ/T Hybrid Schemes&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/rfc9958/&quot;&gt;RFC 9958: Post-Quantum Cryptography for Engineers&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/rfc9955/&quot;&gt;RFC 9955: Hybrid Signature Spectrums&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.rfc-editor.org/info/rfc9881/&quot;&gt;RFC 9881: ML-DSA in X.509&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.rfc-editor.org/rfc/rfc9909.html&quot;&gt;RFC 9909: SLH-DSA in X.509&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://datatracker.ietf.org/doc/draft-ietf-lamps-pq-composite-sigs/&quot;&gt;RFC 10073: Composite ML-DSA&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://rfc-editor.org/info/rfc6379&quot;&gt;RFC 6379&lt;/a&gt;, &lt;a href=&quot;https://www.rfc-editor.org/info/rfc6380&quot;&gt;RFC 6380&lt;/a&gt; (Suite B for IPsec, Historic), &lt;a href=&quot;https://datatracker.ietf.org/doc/rfc8423/&quot;&gt;RFC 8423&lt;/a&gt;, &lt;a href=&quot;https://www.rfc-editor.org/info/rfc9206/&quot;&gt;RFC 9206&lt;/a&gt; (CNSA for IPsec)&lt;/li&gt;
&lt;li&gt;pquip, active with the IESG: &lt;a href=&quot;https://datatracker.ietf.org/doc/draft-ietf-pquip-hbs-state/&quot;&gt;draft-ietf-pquip-hbs-state&lt;/a&gt;, &lt;a href=&quot;https://datatracker.ietf.org/doc/draft-ietf-pquip-pqc-hsm-constrained/&quot;&gt;draft-ietf-pquip-pqc-hsm-constrained&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Individual: &lt;a href=&quot;https://datatracker.ietf.org/doc/html/draft-guthrie-cnsa2-ipsec-profile/&quot;&gt;draft-guthrie-cnsa2-ipsec-profile&lt;/a&gt;, &lt;a href=&quot;https://datatracker.ietf.org/doc/draft-prabel-pquip-pqc-overview/&quot;&gt;draft-prabel-pquip-pqc-overview&lt;/a&gt; (expired)&lt;/li&gt;
&lt;li&gt;Working group pages: &lt;a href=&quot;https://datatracker.ietf.org/wg/pquip&quot;&gt;pquip&lt;/a&gt; · &lt;a href=&quot;https://datatracker.ietf.org/wg/lamps/documents/&quot;&gt;lamps&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;&lt;strong&gt;NIST&lt;/strong&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;a href=&quot;https://csrc.nist.gov/projects/post-quantum-cryptography&quot;&gt;NIST Post-Quantum Cryptography project&lt;/a&gt; and &lt;a href=&quot;https://csrc.nist.gov/Projects/post-quantum-cryptography/publications&quot;&gt;publications&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://csrc.nist.gov/pubs/sp/800/227/final&quot;&gt;NIST SP 800-227: Recommendations for Key-Encapsulation Mechanisms&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://csrc.nist.gov/pubs/ir/8547/final&quot;&gt;NIST IR 8547 (final)&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;&lt;strong&gt;NSA&lt;/strong&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;a href=&quot;https://media.defense.gov/2022/Sep/07/2003071836/-1/-1/0/CSI_CNSA_2.0_FAQ_.PDF&quot;&gt;CNSA 2.0 and Quantum Computing FAQ&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;&lt;strong&gt;European and Australian authorities&lt;/strong&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;a href=&quot;https://digital-strategy.ec.europa.eu/en/policies/post-quantum-cryptography&quot;&gt;European Commission: Post-Quantum Cryptography policy page&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://digital-strategy.ec.europa.eu/en/library/coordinated-implementation-roadmap-transition-post-quantum-cryptography&quot;&gt;NIS Cooperation Group: Coordinated Implementation Roadmap (June 2025)&lt;/a&gt; and &lt;a href=&quot;https://digital-strategy.ec.europa.eu/en/library/eu-roadmap-post-quantum-cryptography-survey-feedback&quot;&gt;survey feedback / FAQ (Sep 2026)&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://etsi.org/deliver/etsi_ts/103700_103799/103744/01.02.02_60/ts_103744v010202p.pdf&quot;&gt;ETSI TS 103 744 V1.2.2: Quantum-safe Hybrid Key Establishment&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://ccb.belgium.be/news/preparing-quantum-era-strategic-imperative-belgian-organisations&quot;&gt;CCB: Preparing for the Quantum Era&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://bipt.be/operators/publication/communication-of-13-january-2026-on-minimum-requirements-and-roadmap-for-the-post-quantum-transition-in-the-telecommunications-sector&quot;&gt;BIPT: Communication of 13 January 2026 on the post-quantum transition in telecommunications&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://cyber.gouv.fr/enjeux-technologiques/cryptographie-post-quantique/&quot;&gt;ANSSI: Post-quantum cryptography&lt;/a&gt; and &lt;a href=&quot;https://messervices.cyber.gouv.fr/guides/Transition-post-quantique-protocole-IPsec&quot;&gt;Transition post-quantique d’IPsec (Feb 2026)&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://ir.cwi.nl/pub/34604/34604D.pdf&quot;&gt;AIVD, CWI, TNO: The PQC Migration Handbook, 2nd edition&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://ncsc.gov.uk/guidance/pqc-migration-timelines&quot;&gt;UK NCSC: Timelines for migration to post-quantum cryptography&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.cyber.gov.au/business-government/secure-design/planning-post-quantum-cryptography&quot;&gt;ASD: Planning for post-quantum cryptography&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;&lt;strong&gt;Research and news on quantum computing&lt;/strong&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;a href=&quot;https://arxiv.org/abs/2505.15917&quot;&gt;C. Gidney, &lt;em&gt;How to factor 2048 bit RSA integers with less than a million noisy qubits&lt;/em&gt; (2025)&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://tech.yahoo.com/science/articles/quantum-computers-just-10-000-152714950.html&quot;&gt;Live Science / Yahoo: Quantum computers might need just 10,000 qubits to break encryption (Mar 2026)&lt;/a&gt;, on the Caltech/Oratomic preprint arXiv:2603.28627&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://interestingengineering.com/innovation/quantum-computing-fewer-qubits-breakthrough&quot;&gt;Interesting Engineering: Caltech and Oratomic error-correction architecture (Mar 2026)&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.discovermagazine.com/quantum-computing-is-beginning-to-take-shape-here-are-three-recent-breakthroughs-48938&quot;&gt;Discover Magazine: three recent quantum computing breakthroughs (Apr 2026)&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.livescience.com/technology/quantum/ibm-quantum-processor-achieves-highest-fidelity-calculations-for-the-longest-period-of-time-on-record&quot;&gt;Live Science: IBM logical-qubit fidelity record (Apr 2026)&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://www.theblock.co/news/business/2026-04-24-researcher-breaks-15-bit-elliptic-curve-key-wins-1-bitcoin-bounty-project-eleven-398792&quot;&gt;The Block: 15-bit elliptic-curve key broken, Project Eleven Q-Day Prize (Apr 2026)&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href=&quot;https://thequantuminsider.com/2026/04/21/cryptologist-finds-aes-128-likely-safe-from-quantum-attack/&quot;&gt;The Quantum Insider: AES-128 likely safe from quantum attack (Apr 2026)&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;aside class=&quot;status-note&quot;&gt;&lt;p&gt;&lt;strong&gt;Status note.&lt;/strong&gt; Every RFC, Internet-Draft and NIST publication in this article was checked against the IETF Datatracker, the RFC Editor, IANA and NIST in &lt;strong&gt;November 2026&lt;/strong&gt;. Several of the Internet-Drafts mentioned are in their final stage before becoming an RFC (in the RFC Editor queue or with the IESG), so their numbers, versions and details can change quickly. Always check the Datatracker for the latest version before you rely on one.&lt;/p&gt;&lt;/aside&gt;</content:encoded><category>PQC</category><category>Cryptography</category></item></channel></rss>