[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-core-devel
Subject:    Re: kio_http, persistent connections
From:       Rik Hemsley <rik () kde ! org>
Date:       2001-06-04 20:08:00
[Download RAW message or body]

#if Lars Knoll
> > that would explain why KIO is still a few magnitudes slower in retrieving
> > webpages than i.e. IE/Netscape.
> 
> If pipelining is not implemented, that would explain a lot. 

Yes, it looks like persistent connections are implemented, just not
pipelining. I'm not sure though, it's quite complicated to test :)
It seems that way, because why else would kio be creating new HTTP
connections when there are already 4 open ?

> The http server send us a timeout anyway. We can just keep it open up to 
> MAX(server_timeout, our_max_timeout).

Or MIN :)

> > > I think that the best solution would simply be to ensure that no
> > > more than 4 connections are made to one host.
> >
> > why ?
> 
> Because of the http specs? Because opening 20 connections to one host 
> thrashes the server if every client does it.

Load on the server is one problem, but the problem that concerns me most
is that if there are many idle connections, the server reaches its
simultaneous connection limit, when it really shouldn't. This blocks
new connections until the old ones timeout.

Yes, I should set the timeout lower, but if konqy comes in and takes
up 6 connections out of a maximum of 32 (just from looking at one page)
and then doesn't close them, that leaves a lot less for any other
clients for the next (timeout) seconds.

I say 6 connections, because that happens consistently when kio_http
requests one of my web pages, finds that 4 of its connections are busy
(getting tiny graphics) and then spawns a couple more connections. Ouch.

Rik

[prev in list] [next in list] [prev in thread] [next in thread]