[hydra-dev] Use Case: Distributed Queue System

[Leif Madsen]

Looking at emails, I guess we're writing about use cases for the system :)

One of the things I'm seeing a lot with queues is the ability to distribute them across multiple servers. What Asterisk currently can't do is know the position of callers no matter which server they have entered.

So lets say we have 3 Asterisk servers, and callers are distributed (somehow) evenly throughout the cluster. What I've been seeing is a want for the need to have a single queue loaded across all 3 servers, and to distribute callers as if they were on a single instance of Asterisk. So if caller 1 enters Asterisk A, and caller 2 enters Asterisk C, then caller 1 is distributed to an agent and caller 2 is distributed to an agent, while caller 3 (on Asterisk B) sits in the queue waiting for the next available agent.

I wasn't at the meeting, so I'm not sure if this delves into an application realm that was discussed as not being applicable to Hydra. But I think if you generalized it down to, "Information should be available across several physical systems acting as a single logical system". I think some of the logical volume managers work like this, in that you can just add another hard drive (server) to the mix, and increase the amount of memory (calls) in the logical volume.

Leif.

-- 
Leif Madsen
Digium, Inc. | VoIP Engineer and Consultant
445 Jan Davis Drive NW - Huntsville, AL 35806 - USA
Check us out at: www.digium.com & www.asterisk.org