On Tue, Jun 01, 1999 at 10:42:25PM -0600, Dax Kelson wrote:
>
> I was working on a commercial caching box today, the only feature it had
> that Squid didn't was the following.
>
> When a user requests a html document, the commercial box scans that html
> document for additional objects (mostly images) and starts retrieving
> those immediately. Shortly after, when the client requests those objects,
> the cache box either has them already, or is already getting them.
>
> Could this be implemented in Squid?
>
This could have nasty side effects. In Australia we're mostly charged by
the megabyte or by % utilisation, a user visiting a site with say 1000
linked MP3, movies or FTP site (each around 5-10Meg) would generate
around 500-5000Meg of traffic just from visiting a single page!
The commercial solution must have some kind of sanity checking to prevent
this (perhaps HEADing the objects and checking their size or limiting the
spidering effect to only a few links at time or in total).
tom@interact.net.au
Received on Wed Jun 02 1999 - 03:27:32 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:46:42 MST