On Fri, Jun 01, 2001 at 03:31:34AM -0500, Joe Cooper wrote:
> Nathan Hand wrote:
> > On Thu, May 31, 2001 at 10:38:27AM +0100, Pat Newby wrote:
>
> >>On a similiar topic, he says that the same box will cache streaming
> >>media, which squid is unable to do at the moment. (And that it will
> >>do a lot of other things as well of course).
> >
> > He's right. There are several limitations in Squid. But NetCache
> > has its own set of limitations too. The hard bit is figuring out
> > which limitations are relevant to your site.
> >
> > NetCache is better than Squid for performance/$. NetCache gets a
> > higher peak throughput too. NetCache is also proprietary and has
> > a hefty price-tag. Both are stable and feature-rich.
>
> Actually, NetCache is not better than Squid on price/performance, though
> it can do a higher peak throughput, as you point out. See the last
> cacheoff results and compare the NetApp, Squid, and Swell numbers:
>
> From this it's pretty clear that Squid on FreeBSD is the equal of the
> low-end NetCache. I'll leave it to the reader to note how the
> Squid-based Swell entry compares (I don't want to try to use the Squid
> list for advertising).
>
> NetCache is not a very high performance caching system. In fact, it
> shows it's hereditary link to Squid (both originated from Harvest in the
> distant past) by performing very similarly on similar hardware. They've
> done a few performance enhancements that make it scale better than Squid
> to big hardware...but not a whole lot better.
I'm happy to be wrong :-)
And while Squid won't scale to big hardware you can just stick several
Squid boxes behind a L4 switch. So the low peak-throughput of Squid is
not really an issue.
> Interestingly, I've found that many of the 'low-end' entries that
> provided better price performance than the Squid entries at the cacheoff
> don't actually exist for purchase anymore (if they ever did). I did a
> little price research a couple of weeks ago to make sure we are still
> very competitively priced...I needn't have worried. I couldn't seem to
> locate most of the under $5k entries on the vendors websites, or if I
> could, they were significantly more expensive than the price reported at
> the cacheoff...Seems a bit odd, and misleading.
Unethical, I'd say.
-- The more I know about the WIN32 API the more I dislike it. It is complex and for the most part poorly designed, inconsistent, and poorly documented. - David KornReceived on Fri Jun 01 2001 - 05:55:14 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:00:27 MST