Re: H2: Should there be a limit to tolerance ?

--------
In message <CABkgnnVuWHQkpcu3yxAxisxziGqMmJ_vKVStUn-1UdbBKak0HQ@mail.gmail.com>, Martin Thomson writes:

>> What if it keeps hammering you with *only* frames which get ignored ?
>>
>> What if it does so at very high rate, because it is buggy or hostile ?
>>
>> What if the buggy implementation was in several million Internet-Of-Shit
>> things that got poured into concrete years ago ?
>>
>> There's got to be _some_ limit to patience ?
>
>See http://httpwg.org/specs/rfc7540.html#dos

Yes, I've seen that.

What I'm asking is if we should try to avoid have N different
critieria for N different implementations or if it makes sense
to try to coordinate some sort of same-ish criteria.

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.

Received on Friday, 17 February 2017 23:01:35 UTC