Re: Bandwidth

James Green (jmkgre@essex.ac.uk)
Tue, 25 Nov 1997 17:03:11 +0000 (GMT)


From: James Green <jmkgre@essex.ac.uk>
To: www-html@w3.org
Message-Id: <SIMEON.9711251711.B@s1675.essex.ac.uk>
Date: Tue, 25 Nov 1997 17:03:11 +0000 (GMT)
Subject: Re: Bandwidth


On Tue, 25 Nov 1997 13:36:53 +0000 James R Grinter <jrg@demon.net> 
wrote:

> the reason that there are currently still traffic problems, despite
> usage of caches, is because the cache usage isn't high enough. Too
> many people still try and go direct. Because it isn't high enough,
> many people won't invest in increasing their caching capacity (chicken
> and egg problem).

True enough currently, but only over the past two years or so have 
people come to accept the Internet as a household name. Give it another 
ten years, and a great many more people 'hooked up' to it, and see if 
the problem hasn't escalated. I know from experience from both home and 
University (both Britain), that sometimes sites over in the States are 
uncontactable due to high usage, so what happens when the Internet's 
usage increases, perhaps expotentially, over the coming years? If 
you're so concerned about getting people using caches, why not really 
push them, rather than mentioning them in that magazine the American's 
in the audience won't know about (Despatches)? Giving away IE 4 free 
on a free magazine supplement was a good idea with caching already 
set-up - those people who don't know about it won't know how to switch 
it off.

However, this relates only to Demon customers, so how about some 
readily-available public servers? Ones which are advertised and maybe 
even used by ISPs themselves? Then you start getting in to my thinking 
of country-specific versions.

> the trick, the thing that needs solving, is getting the client to go
> to the closest point where they can get that identical content from.

And also, to ensure that 'local' copies of the file are always up to 
date.

Perhaps if my idea *was* implemented, network traffic would be low 
enough to justify robots searching web sites themselves, updating their 
caches with new bits all the time. Now there's a thought - the user 
won't actually get to the site themselves at all, a robot gets it for a 
cache for others to use instead. Sounds a bit like a caching mechanism 
again!


Regards,

James Green

Term e-mail: jmkgre@essex.ac.uk   |   Home e-mail: jg@cyberstorm.demon.co.uk
Homepage: http://www.cyberstorm.demon.co.uk