Currently, I've got sipx installed as a small cluster of servers, two redundant servers and one as a voicemail only server on an IBM BladeCenter. The two redundant servers are on 2 x 2.8Ghz, 64Bit, 4GB RAM blades and the voicemail server is a guest on an ESX server running the same hardware and three other low use guests.
A common suggestion seems to be 'run multi-core CPU's'. I could migrate these over to my 2 x dual core-2.4Ghz, 64Bit, 4GB blades but before I bother, I wonder about the following. My plan is to have 2 or 3 separate locations to run the cluster at some point so wonder what might be easier. A blade server is a big box while similar power in individual boxes takes up less space, less heat, cooling, power, etc. Easy to put into various locations. So for example, I've got multi-CPU servers as well, some of which are dual XEON 2.8Ghz servers and even 8-way 3Ghz servers. The thing about these servers is that some of them can handle 32 and 64GB of memory. So, not multi-core processors, single core, multiple CPU's but tons of memory. I'm trying to get an idea of which might be the better way to go, for future flexibility and performance. For example, on a box which has 32GB of memory, perhaps I could use 4GB for the main app, then create a huge cache for other things? I know that this question is based on many variables but it's something I'd like to get right before going into production. Sure would welcome some input based on what you've been doing. Thanks. _______________________________________________ sipx-users mailing list sipx-users@list.sipfoundry.org List Archive: http://list.sipfoundry.org/archive/sipx-users Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-users sipXecs IP PBX -- http://www.sipfoundry.org/
