Re: [quicwg/base-drafts] handling of coalesced packets with decryption errors creates DoS opportunity (#2308)

[Marten Seemann <[e-mail]>]

> That could be true, though I am still not sure if we need to talk specifically about the attack.

I was not suggesting to add language about this attack. My suggestion was to stop processing a datagram once unprotecting one of the QUIC packets fails, although you have the keys available for that encryption level.

> For example, an attacker can also send a packet that successfully decrypts, that contains many tiny CRYPTO frames (4-bytes each). I'd assume that processing of that would be more costly than AEAD-decrypting every 35 bytes.

I don't understand this. Why is processing CRYPTO frames expensive? In my implementation, assembling stream (and crypto stream) data is not an expensive operation.

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/quicwg/base-drafts/issues/2308#issuecomment-451828441