Re: two ideas...

>    I have two ideas which I would like to get some feedback on. One is
>to use compressed HTML/Java instead of regular HTML. The second is to
>implement intelligent pre-fetching for the web.
>
>    I would be interested in hearing some feedback from people on these ideas.
>
>P.S. For HTML, if your feedback consists of pointing out that the HTML
>content is a very small amount of the web traffic... please DON'T mail
>that to me unless you have some numbers... I have already covered this
>point in the paper and numbers on the traffic content would be more
>useful than relative terms... I'm looking for sources to provide me with
>data which can be useful for this analysis.
>
>Sneaker

We run a public caching proxy for the academic community in the UK.
From November 10th to November 19th our servers handled 9,800,000
requests delivering 126 gigabytes of data. Of this HTML accounted for
25% by access count and 12% by number of bytes.

We have very poor trans-atlantic bandwidth and so I would suggest that
even a 5% saving overall would be worth some effort.

Neil Smith
HENSA Unix

Received on Thursday, 30 November 1995 11:47:53 UTC