One of the advantages of deploying a feature rich call center software suite in the Cloud is that there are feature rich APIs and tools available for managing both software and cloud infrastructure. The APIs of more advanced Cloud providers are of interest where they allow a call center to bring pre-configured instances up and down on demand. This can result in a significant cost savings when a major factor in the cost of the service is CPU usage.
The ability to load balance servers to expand service capabilities in the contact center is a key attribute in easily managing a call center software deployment. In the case of an Asterisk based ACD, for instance, calls and agents must be spread out to avoid creating an unnecessary bottleneck. This has been discussed here very recently, so it’s not necessary to belabour this point. Being able to balance load over a set of servers may allow you to specify the set of servers being used. If this can be managed in real time, or on a scheduled basis, then it should be possible to move load to or from servers as capacity demands.
This point is especially relevant for call centers. While web and application traffic may stay at consistent levels in our 24/7 world with all corners of the globe wired in to the Internet, most call centers still have busy and quiet times. If we know we are likely to only need a third of our capacity for a reasonable length of time (ie midnight to 6am UTC), bringing down a half of our servers should certainly be feasible. As long as the software allows the system to mark certain servers as unavailable for new calls or agent connections, or mark them as available again, we can be open to bringing servers down.
If the call center software can be configured via API to manage load and server deployment, then the same method can be used to signal the Cloud provider that instances can be brought down and up again. With a reporting API available, it would even be possible to devise a system to bring capacity up and down automatically. Cases would have to be devised to cope with any delays that are possible in such a procedure, but could certainly be worked around.