W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > January to April 1997

Re: Pipelining and compression effect on HTTP/1.1 proxies

From: Benjamin Franz <snowhare@netimages.com>
Date: Tue, 22 Apr 1997 17:18:25 -0700 (PDT)
To: http-wg@cuckoo.hpl.hp.com
Message-Id: <Pine.LNX.3.95.970422170128.6540A-100000@ns.viet.net>
X-Mailing-List: <http-wg@cuckoo.hpl.hp.com> archive/latest/3134
On Tue, 22 Apr 1997, Bob Monsour wrote:

> Most modem dial-up links employ PPP to initiate their connection and PPP
> compression, as implemented in Win95 and NT, negates most if not all of the
> effects of modem (v.42bis) compression.

Don't underestimate the modem compression. When moving highly compressible
(textual) data it can easily double to triple throughput in my experience.

> A bigger-picture issue for compressing HTML is the potential to reduce the
> number of IP packets traveling over those 'hops'. This is an effect that
> neither PPP nor modem compression can offer. The fewer packets handled at
> each hop, the more capacity available for other traffic at those hops and
> thus, more overall network improvement. That is, the raw 'local' efficiency
> of the compression is not the only benefit, even when that efficiency is
> relatively low.

Even without the modem compression you are only talking on the order of a
7.5% savings - or about 3/4 of a second difference on a ten second load -
for a typical web page with a mix of about 30K of text and graphics
combined. This seems to me to be chasing the point of diminishing returns.

Benjamin Franz
Received on Tuesday, 22 April 1997 17:19:55 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:40:19 UTC