Re: [Kamailio-Users] limits per subscriber and per gateway

On 20.08.2009 14:34 Uhr, Alex Balashov wrote:

Daniel-Constantin Mierla wrote:

...

This solution can be updated nicely in case of a farm of sip server, by replacing the htable with memcached module (new in devel).

Well, memcached may be a much faster and distributed key-value cache store, but I think using it for IPC (as opposed to, say, a cache!) is fundamentally committing the same fallacy as the "Database as an IPC mechanism" anti-pattern. Except that the outcome is not quite as harmful from a quantitative perspective.

right, sometimes we have to live with what we have :-). anyhow, shared memory is clearly defined as IPC mechanism

Cheers, Daniel