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: Luigi Rizzo <luigi@labinfo.iet.unipi.it>
Date: Wed, 23 Apr 1997 19:15:44 +0200 (MET DST)
Message-Id: <199704231715.TAA23431@labinfo.iet.unipi.it>
To: Koen Holtman <koen@win.tue.nl>
Cc: frystyk@w3.org, snowhare@netimages.com, http-wg@cuckoo.hpl.hp.com
Well it looks like we are discussing a bunch of different problems.

I was looking at storage management at the proxy (and thought the
subject was clearly indicating this, although now I see it might
be a bit ambiguous). Probably, wrong mailing list...

Some of the other postings and pointers were referring to minimizing
traffic over the link (or perhaps transfer time, which is yet another
issue ?).

Apart from using caches, you can reduce the traffic by using:

    - a more compact source encoding;
    - link layer compression;
    - shorter/compressed protocol headers;

to me the former two approaches seem to be slightly out of scope:
although in the short term it might be worthwile to try to compress
redundant encodings such as HTML or JAVASCRIPT files, as others
have commented, this is just a temporary solution which will probably
become less and less encessary as providers will realize that they have
to cut data at the source.

I'd also note that many of the notes on the web pages focus on TCP
features such as Nagle, slow start, delayed acks and try to defeat them
in various ways. Again, once a problem is known, just solve it the
right way (e.g. by using appropriate socket options).

	Cheers
	Luigi
-----------------------------+--------------------------------------
Luigi Rizzo                  |  Dip. di Ingegneria dell'Informazione
email: luigi@iet.unipi.it    |  Universita' di Pisa
tel: +39-50-568533           |  via Diotisalvi 2, 56126 PISA (Italy)
fax: +39-50-568522           |  http://www.iet.unipi.it/~luigi/
_____________________________|______________________________________
Received on Wednesday, 23 April 1997 11:03:22 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:32:35 EDT