[c-nsp] Anycast services

[Joe Provo]

On Thu, May 18, 2006 at 04:15:58PM -0400, Tim Durack wrote:
> So everybody keeps saying. I know TCP is stateful, but for a web
> proxy, seems like it would work well enough. Might have to refresh to
> fix broken state, but users are used to that kind of behaviour on the
> Internet anyway.

Ah. The 'lowered expectations' school of thought. You might wish to 
just deploy it if you aren't concerned.

> Maybe dns with low ttl is a simpler solution, but it doesn't solve the
> problem of failing over when a proxy is dead.

Your DNS server changes the answer based on knowledge of the service's
availability.  Not rocket science.

Cheers!

joe

--
Joe Provo                                            Voice  617.670.2904
Senior Manager, Internet Planning & Design           Fax    617.670.2920
Network Engineering, RCN                             <joe.provo at rcn.com>