W3C home > Mailing lists > Public > ietf-http-wg@w3.org > April to June 2012

Re: HTTP -> Messages -> Transport factoring

From: Carsten Bormann <cabo@tzi.org>
Date: Thu, 5 Apr 2012 23:06:09 +0200
Cc: "ietf-http-wg@w3.org Group" <ietf-http-wg@w3.org>
Message-Id: <67C5C37B-D1E4-4BCD-BAF3-926B119FFE35@tzi.org>
To: "Poul-Henning Kamp" <phk@phk.freebsd.dk>
> The HTTP/UDP use-case is different, in that we are specifically
> talking about non-transactional small object retrival, but the
> 64k limit is of course the same thing.

How does the intermediary know that the object will be small (< 64 KiB minus epsilon)?
(Maybe there is some configuration that makes this prediction possible.)
What happens if the prediction was wrong?
(http://tools.ietf.org/html/draft-ietf-core-block is one way of handling this case, but maybe not the one you have in mind.)

> Once of the attractions of HTTP/UDP is that you can multicast and
> whichever server happens to have that icon will send it to you.

If all those servers can handle the load of discarding lots of requests...
(And assuming the set of servers actually responding is small enough not to cause implosion issues.)
Is there anything written up on this server-side "HTTP over UDP" that I can consume before bothering you further?

Gre, Carsten
Received on Thursday, 5 April 2012 21:06:43 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:51:59 GMT