Hi Yoav,
just for clarification: That was me. When they started talking about about PQC a decade ago, they mentioned algorithms like McEliece with public keys around 1MB. Not that this is a deal-killer. If necessary, we would come up with an IKE extension to have “jumbo-sized payloads” that had 24-bit lengths. IKE over TCP (RFC 8229) can handle this easily. Actually, not so easily. Despite the fact that IKE Message header has a length field of 32 bits, the message prefix in RFC 8229 is only 16 bits in size, so no single IKE message transferred over TCP can be greater than 64K. IKE fragmentation could help in this case, but RFC 8229 says that “implementation SHOULD NOT send fragments when going over a TCP”. However the good news is that “it MUST support receiving fragments.” So we just need to violate that SHOULD. Regards, Valery. But anyway, since the current state of the art seems to not need such an extension, I guess there’s no point it bringing this up now. Yoav On 25 Mar 2018, at 20:36, Scott Fluhrer (sfluhrer) <sfluh...@cisco.com> wrote: During the WG meeting in London, somebody (I forget who) asked me whether KE payloads larger than 64k. I thought I ought to clarify matters (as they are more complex than the brief answer I gave indicated). Of the proposed postquantum key exchange (and public key encryption algorithms, which can be used to transport keys) submitted to NIST, the majority of them have key shares (or public keys/ciphertexts) smaller than 64k; there are a handful that are larger. Now, it is possible (albeit unlikely) that all the algorithms with key shares < 64k will be broken; unless this happens, it would be reasonable (IMHO) that we mandate that any algorithm he allow have a KE payload that fits within 64k. Now, in the event that we feel the need to support larger key shares, there are possible ways to support that; I don’t feel that we need to talk about those options now. _______________________________________________ IPsec mailing list <mailto:IPsec@ietf.org> IPsec@ietf.org <https://www.ietf.org/mailman/listinfo/ipsec> https://www.ietf.org/mailman/listinfo/ipsec
_______________________________________________ IPsec mailing list IPsec@ietf.org https://www.ietf.org/mailman/listinfo/ipsec