Re: squid features

From: Henrik Nordstrom <hno@dont-contact.us>
Date: Thu, 24 Jun 1999 00:33:31 +0200

Lawrence R. Mulder wrote:
 
> I was cruising the squid web site, but didn't see the info I was looking for
> but was wondering if you could answer a couple of questions about where the
> squid project might be going.
>
> If it doesn't already (haven't read that it can) does squid plan on
> supporting HTTP 1.1?

Eventually yes, but not in the immediate future. A lot of changes is
needed to support HTTP/1.1 (it is a fairly complex protocol to
proxy/cache, compared to HTTP/1.0 by which Squid was designed).

> Is their planned support for Streaming media caching not just proxying or
> tunneling?

Squid is only a HTTP proxy. No plans to include other protocols. Our
statement is that other protocols is better processed by separate proxy
softwares.

As for caching of streaming on top of HTTP. Squid may cache this,
provided that the data complies with HTTP cachability rules.

> Finally, would it be worth it at this point in squid's developement to write
> a proprietary database for cache objects and write them to a raw disk
> instead of an OS filesystem formated disk?

Work is being made in this area to provide a general framework to
develop and test such object stores. We will probably see three or four
different object stores for Squid within the next 6 months or so.

> Basically I would just like to see squid be able to scale up to what some of
> the big boys like Inktomi Traffic Server. See
> http://www.inktomi.com/products/traffic/

Squids primary goal is not to compete at the ultra high scales. Squid is
primary a research tool to try out the feasibility of web caching
techniques.

That said, Squid will in the future scale a lot better than it does
today, and hopefully provide a more rubust service.

--
Henrik Nordstrom
Spare time Squid hacker
Received on Wed Jun 23 1999 - 16:36:34 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:46:58 MST