Re: [quicwg/base-drafts] If you want a Stateless Reset you need to send a much larger packet than before (#2770)

[ianswett <[e-mail]>]

I remember a few goals for stateless rejects(there are probably more):
 1) Cannot be spoofed
 2) Ideally an active attacker cannot cause a server to generate one for an active connection
 3) Indistinguishable from other traffic
 4) No amplification attacks
 5) No endless loops
 6) Can be sent in response to any packet that can't be associated with an active connection.

We compromised 6 in order to achieve 3 more often.  I think we can tweak the existing design to not compromise 6 and still provide 3.

We can put a bit more burden on the client and add text to indicate that if the client wants a stateless reject that's indistinguishable from other traffic, it needs to send a packet that's at least 1 byte larger than the minimum plausible incoming packet.  In most circumstances, I expect this to naturally occur because the server connection ID will be longer than the client's.

-- 
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#issuecomment-499894384