Interesting comments! Thanks, Thom.

My replies inlines.

Guilin

发件人:Thom Wiggers <[email protected]<mailto:[email protected]>>
收件人:Wang Guilin 
<[email protected]<mailto:[email protected]>>
抄 送:[email protected] <[email protected]<mailto:[email protected]>>;Valery Smyslov 
<[email protected]<mailto:[email protected]>>;Wang Guilin 
<[email protected]<mailto:[email protected]>>
时 间:2026-04-08 19:50:31
主 题:Re: [IPsec] [IPsec]: New Version Notification for 
draft-wang-ipsecme-kem-auth-ikev2-03.txt

Hi Guilin and Valery,


Op 7 apr 2026, om 11:12 heeft Wang Guilin 
<[email protected]<mailto:[email protected]>> 
het volgende geschreven:

In addition, beside using ML-KEM to replace ML-DSA for authentication in IKEv2, 
we also noticed that some other KEMs could be good candidates as well. For 
example, Classic McEliece has public key sizes from 260KB to 1.36 MB, which is 
huge compared to ML-KEM. However, the ciphertext sizes are just 96-208 bytes, 
very short. Therefore, in the case two entities need to authentication with 
each other frequently,  Classic McEliece could  be a good choice to save 
communication overhead, by assuming that each side can store public key or 
certificate of the other side. If a few MB storage is not an issue, using 
Classic McEliece as KEM based authentication may be even practical for IoT 
devices with constrained capability, but only communicating with fixed parties.

Table 6 in [1] gives the exact sizes of Classic McEliece variants.

> I think that the use of Classic McEliece to avoid transmission suits IKEv2 
> well, since it supports some “Certificate” formats such as the hash-and-url 
> scheme that very naturally work for this out-of-band distribution of public 
> keys.

[Reply from Guilin] Yes, this sounds naturally working. We can intergrate this 
approach expilictly in the draft, I think.

> Of course this mechanism will also be useful for UOV-style schemes that have 
> very large public keys but small signatures.

[Reply from Guilin] Hope more PQ KEMs with shorter ciphertext can be 
standardized in the near future.


Cheers,
Thom


_______________________________________________
IPsec mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to