Think about what happens when a skyscraper is about to be built. The construction crew obviously wouldn't begin by making the observation deck. Building a new contact center is a lot like the construction of a new building. It all starts with a plan, along with a solid foundation.
Plan Basic Infrastructure Capacity
Preparing for a new implementation begins with knowing the intended capacity of your system. Once you understand what the capacity is supposed to be, you can start to make a plan where all the components of each system will work the way they're supposed to. Start with your infrastructure and verify that the capacity is there in terms of the number of concurrent interactions you're able to handle. The infrastructure should be able to handle the velocity of both incoming and outgoing interactions.
I always recommend that people first perform the testing necessary to confirm that the basic infrastructure has the expected capacity in terms of the number of simultaneous phone calls or browser interactions. Don't shoot for the moon on the first attempt at testing. Instead, give yourself enough time to go through a very controlled process. Remember, your goal is to make sure that the whole complex works as intended.
There are occasional situations where there's a sudden deluge of a large number of telephone calls to the contact center. In a TDM environment, it wasn't much of a problem (except it took the system a little while to get everything taken care of). With today's VoIP environments along with SIP connectivity to the public network, the same situation could result in a catastrophic system failure. If the SBCs and the network environment haven't been provisioned to handle the anticipated velocity, it's the equivalent to constructing a building with a weak foundation. Read our complete guide to SIP Trunk fundamentals.
Add IVR and CTI Applications
Once you have a solid foundation of basic infrastructure capacity, you can start to build the other components in your contact center. Add the IVR and CTI applications and run them at 60 to 80 percent of your anticipated full load in a production environment. Make sure they still function they way they're expected to.
It's easy to run a single call or browser session into a scenario and see that it'll work from start to finish. You really need to make sure that all the integrations within the system, data processes, and component capacities that are put into place can actually handle the velocity of interactions on top of that infrastructure that you've already proven out.
Perform a Soak Test
It's very important to make sure that systems can continue to run for extended periods of time (not just five or six minutes). In our experience, if a system is going to have issues, they tend to show up within 20 to 30 minutes of running under full load. That's how much time it usually takes for memory leaks to become evident. Soak tests should use externally provided traffic to continuously exercise the system at a production level.
With Prognosis installed on the inside, you can get a bird's-eye view of how the system performs while ensuring that all the different elements in your environment are working within their designed capacity. You'll see if the CPU gets pegged at a particular server/service and have the capability to ensure that the various network segments have adequate capacity to keep up.
Careful instrumentation of the system using Prognosis and the native OA&M tools that come with your system components will help you determine whether or not your system is behaving the way it's supposed to. You'll see if it can tolerate 30%, 60%, or 100% full load.
Stay tuned for our next post, when we'll talk about how to ensure that your system is resilient in the face of component failure. For more information about this topic, join our webinar on March 24.
