[aklug] Re: GCI, ond others, bandwith restrictions

On Wed, 20 Oct 2010, Jeremy Austin wrote:

> I see about a tenth of Arthur's traffic through squid, but the
> percentages are about the same. I've seen savings drop from over 20%
> to around 15% due to the rise of streaming media.
>
> Streaming media (by which I mostly mean YouTube) easily count for 40%
> of HTTP traffic. I'm experimenting with YouTube caching at the moment;
> so far I'm seeing about 30% savings, which should get us back closer
> to 20% savings overall.

I've been tempted to try some of that. We are dropping all GET queries, but
I've thought about caching queries to MS & Apple update servers, etc. I'd
appreciate hearing about any success and/or problems you run into doing
that.

> What we really need is an extension of HTTP that allows a
> Bittorrent-like collaborative transfer, with the addition of
> nearest-neighbors only, and reporting of stats back to the original
> HTTP server for hit counting. I know there are sites that do things
> like this with Java or Flash, but until it's in apache and Firefox, it
> will go nowhere.

For those providers that have peering agreements I think it would be useful
to be able to set up HTCP peers between our clusters. Whether ACS or GCI is
doing any caching at all, though, is a good question.

         --Arthur Corliss
           Live Free or Die
---------
To unsubscribe, send email to <aklug-request@aklug.org>
with 'unsubscribe' in the message body.
Received on Wed Oct 20 21:10:11 2010