RE: my broken TCP is faster on broken networks

Josh Cohen (joshco@microsoft.com)
Mon, 14 Sep 1998 02:15:29 -0700

Greg Maxwell said:
> Then why did we have a microsoftie on here saying IE complies and only
> does 2. Hehe I'm supprised that Windows' TCP/IP stack can handle 32
> connections at once. :)

Dancer,
Could you please reference the case and version of IE you are
talking about ? IE 4+, which supports http/1.1, limits connections
to 2 per server, as per the spec. If you could show otherwise,
I'd be interested in seeing what your seeing.
( Assuming there is no registry tweaking of those settings )

---
Josh Cohen <joshco@microsoft.com>
IE Lead Program Manager - Networking
 


> -----Original Message-----
> From: Dancer [mailto:dancer@brisnet.org.au]
> Sent: Sunday, September 13, 1998 3:21 PM
> To: David S. Miller
> Cc: linker@z.ml.org; mj@ucw.cz; acahalan@cs.uml.edu;
> alan@lxorguk.ukuu.org.uk; linux-kernel@vger.rutgers.edu
> Subject: Re: my broken TCP is faster on broken networks
> 
> 
> David S. Miller wrote:
> > 
> >    Date:        Fri, 11 Sep 1998 19:05:19 -0400 (EDT)
> >    From: Gregory Maxwell <linker@z.ml.org>
> > 
> >    >    The right answer is that HTTP should have supported 
> persistent
> >    > connections since version -1.-1 and all browsers 
> should use that.
> > 
> >    Yes, they should. But they dont seem to. I hit Apache 1.3 sites
> >    with Netscape 4.5b and see multiple TCP connects.
> > 
> > It is another vendors attempt to make their browser look 
> "faster" than
> > others.  Netscape always opens up 4 connections in parallel 
> by default
> > and uses persistant transactions over each one.  You can 
> configure it
> > to use only one open connection.
> 
> MSIE evilly uses up to 32 (by default). Anyhow, there's a 
> long-standing
> rationale behind the parallelizing of connections (which I hate, yes),
> side-by-side with the reasoning that was used to develop the pipelined
> connections. Check www.w3.org for the full story.
> 
> D
> 
> 
> -- 
> -----BEGIN GEEK CODE BLOCK-----
> Version: 3.1
> GAT d- s++: a C++++$ UL++++B+++S+++C++H++U++V+++$ P+++$ L+++ E-
> W+++(--)$ N++ w++$>--- t+ 5++ X+() R+ tv b++++ DI+++ e- h-@ 
> ------END GEEK CODE BLOCK------
> 
> -
> To unsubscribe from this list: send the line "unsubscribe 
> linux-kernel" in
> the body of a message to majordomo@vger.rutgers.edu
> Please read the FAQ at http://www.tux.org/lkml/faq.html
> 


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/faq.html