Re: Large Frame Proposal

In message <CAOdDvNptJiAUcr1sShO6dbfKNya9xm8TTkqc7fk7-cYy=5d10w@mail.gmail.com>, Patrick McManus 
writes:

>obviously if jumbo frames are not used jumbo frames dont matter
>
>but when they are used, all of the receivers streams are put at risk. So
>the client is incented to use multiple connections so the risk is not
>linked. And that's a fail for h2.

First, this is rather trivially detectable by any moderately compehensive DoS
detection scheme.

Second, what you describe amounts to a client using HTTP/2 in HTTP/1 fashion.

That would be the almost trivial way to add HTTP/2 support to many
HTTP/1 implementation, without having to fundamentally change the
applications internal scheduling.

It may aid HTTP/2 adoption if we are not too harsh too soon on this "hack".

-- 
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 Wednesday, 9 July 2014 12:32:41 UTC