On Tue, 21 Apr 1998, Michael Ferioli wrote:
> At 09:12 AM 4/21/98 +1200, Joe Abley wrote:
> >
> >I think the original poster was talking about distributing three caches
> >(or cache clusters) in the wide area. A cluster of unix boxes running
> >squid in each city (with round-robin DNS set up to spread load across each
> >machine in the cluster) should work pretty well.
>
> Exactly. But do you really think DNS round robin is a good solution?
> Why would an Ankara user use the Istanbul Cache? That wastes
> bandwidth which kind of defeats the whole idea of proxying.
Yep, I completely missed the point of what you were trying to do :)
You could use it to tie a cache farm within one city to a single name for
the customer, though. If you set all your clients up to fetch an autoproxy
script in order to configure switch settings, you might be able to serve
up a different collection of settings based on the IP address of the
caller. That might do the trick quite nicely.
Alternatively you could maybe use something like Cisco's LocalDirector or
Alteon's layer-4 switching to make a "virtual" cache address redirect to
one or more real caches.
> >I didn't notice the "NT" word in the original post (but then again, I
> >didn't read it very well :)
>
> Absolutely no NT. I don't even know how to spell NT.
All kneel and praise Bill.
-- Joe Abley <jabley@clear.co.nz> Tel +64 9 912-4065, Fax +64 9 912-5008 Network Architect, CLEAR Net http://www.clear.net.nz/Received on Tue Apr 21 1998 - 05:30:23 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:39:47 MST