- From: Nicolas Mailhot <nicolas.mailhot@laposte.net>
- Date: Tue, 10 Apr 2012 08:56:26 +0200
- To: "Per Buer" <perbu@varnish-software.com>
- Cc: "Nicolas Mailhot" <nicolas.mailhot@laposte.net>, "Poul-Henning Kamp" <phk@phk.freebsd.dk>, "Adrien W. de Croy" <adrien@qbik.com>, "ietf-http-wg@w3.org" <ietf-http-wg@w3.org>
Le Lun 9 avril 2012 18:30, Per Buer a écrit : > On Sun, Apr 8, 2012 at 11:12 PM, Nicolas Mailhot > <nicolas.mailhot@laposte.net> wrote: >> >> Ah, but on a big proxy farm setup, the load balancer may orient the second >> request to an new proxy (since the first one is busy). And the origin web >> site >> may also be part of a cdn, and the second request may not end on the same >> delivery server as the first one > > So, if the proxy farm fails to hash incoming requests on source IP or > target URL then this might happen. That breaks load balancing as soon as your network is big enough, with different parts that get active at different points of the day. > But either of these methods will > easily help avoid the problem. No they won't. To scale network equipments need to be as stupid as possible, and as much smarts as possible kept in the endpoints. You're breaking this principle there. And anyway even if your solution was possible, you still get unhappy users that serial refresh because they're not seeing initial progress in their web clients -- Nicolas Mailhot
Received on Tuesday, 10 April 2012 06:56:58 UTC