[quicwg/base-drafts] Stateless Reset becomes a larger risk of amplification with longer CIDs (#2770)

[ianswett <[e-mail]>]

The two principles previously espoused were:
1) MUST ensure every stateless reset is smaller than the packet received to prevent looping: https://tools.ietf.org/html/draft-ietf-quic-transport-20#section-10.4.3
2) Stateless resets MUST be 69 bytes with #2749 in order to make them non-identifiable.

This means clients that want stateless resets in response to keepalives(ie: PINGs) must send 70 byte packets or risk getting nothing in response, therefore waiting for an RTO.

This seems like an unnecessary risk/complexity, when we could decrease the minimum size of the stateless reset being processed by a client.  The risk/cost to the client seems very low of processing stateless resets after a packet has already failed decryption.

Additionally, it's difficult to imagine a client ever use a 16 byte connection ID, and even more implausible the client will use a 48 byte connection ID, given most clients are expected to have one connection per port and even those with multiple likely won't have so many they need >8 byte CIDs.

-- 
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/2770