- From: Willy Tarreau <w@1wt.eu>
- Date: Thu, 24 Jan 2013 08:08:18 +0100
- To: Patrick McManus <mcmanus@ducksong.com>
- Cc: "William Chan (?????????)" <willchan@chromium.org>, James M Snell <jasnell@gmail.com>, Nico Williams <nico@cryptonector.com>, Roberto Peon <grmocg@gmail.com>, "ietf-http-wg@w3.org" <ietf-http-wg@w3.org>
Hi Pat, On Wed, Jan 23, 2013 at 07:55:42AM -0500, Patrick McManus wrote: > On Wed, Jan 23, 2013 at 2:33 AM, Willy Tarreau <w@1wt.eu> wrote: > > > As I said, I think that if the state itself is never larger than a request > > and substitutes for the request, it's not that big of a deal. > > > > > I'm more comfortable saying that the cost of the state should be > proportional to any latency-scaled win it can give... honestly, the trend > in ram prices is a lot friendlier to making progress than the trend in > bw-to-latency ratio. You don't count the trend in user count which is faster than the trend in RAM price unfortunately. I've had several times people ask me what to change in their kernel to go beyond 1 million concurrent connections in haproxy. A few years ago, they were only talking about several tens of thousands. With everything device connected to everything all the time, I don't expect this trend to revert any time soon :-/ Willy
Received on Thursday, 24 January 2013 07:08:54 UTC