On Mon, Apr 14, 2008, Charles Glass wrote: > Testing with Squid (2 caches) and mask assignment I was able to get 32 > mask values per cache. The weight attribute did not appear to have any > effect on distribution of the mask values as it does with hash > assignment.
Then use squid. :) > I would like to use mask assignment for obvious reasons but having > only 3 mask values per cache is not feasible. > > Thanks for the info on this it is greatly appreciated. You're just not going to get sensible performance scaling on these devices without the whole path being in hardware. (ObNote; I haven't touched or adminned WCCPv2 on a 7600 w/ ES modules; I've just touched it on the 6500 and I've heard lots of stories from 7600 + squid users. :) Adrian _______________________________________________ cisco-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
