VOIP technology is highly in demand with many right now because it provides a means of communicating at a relatively cheaper rate. But while price is a major consideration, performance is equally important. To have a VOIP test quality, a virtual network test bed may be used to make sure that VOIP ownership and deployment is truly worth it. By using a test bed, developers, QA Specialists, network managers and the other IT personnel can observe the behavior of network applications in a controlled environment where they can truly simulate the realistic demands and actual performance of the network. This simulation should include all the relevant aspects of the service such as latency, loss, jitter, packet loss, sequence errors, echo and the actual quality of the service, to name a few.

To get a better grasp, it is best to familiarize yourself with the terms to watch out for. Latency means the delay caused by the actual transmission. A long latency therefore is not ideal and unacceptable in voice-grade transmissions. To get some perspective, an acceptable delay would be approximately less than 150ms- anything more than that greatly lowers the quality of the service. Loss is another factor which is likewise unacceptable. While some amount is certainly unavoidable, more than 1% is considered too much and will degrade the conversation significantly. A hurdle to the voice quality is called jitter. This happens when the VOIP is run on a congested network. Too much jitter makes the conversation undecipherable and ultimately useless. Sequence errors occur when the packets arrive in the destination in the wrong order. Like jitter, too much sequence errors and the messages become impossible to understand as well. Finally there is Echo which occurs when the signal is sent back to the sender. Obviously, too much of this also lowers the quality of the conversation.

In setting up your virtual network test bed, there are several best practices standards often used. The first is to capture conditions on the network to define all the scenarios possible. To realistically reflect the demands of the environment, data must be gathered over a period of time and analyzed to determine the best, average and worst case scenarios. Results of tests like these relate directly to call quality. The second step is to use the virtual network to run VOIP test services in the lab under the real-world scenarios. When the results for every possible scenario have been taken, we can now test the VOIP by running actual voice traffic. Then, we proceed with the third step which is to analyze the call quality using technical metrics. A sample of a technical metric may be the mean option score (MOS) used to pinpoint the specific points where voice quality is poor. Once this is done, the fourth step is to validate the actual call quality by listening to actual calls. This is still needed since technical metrics can sometimes prove unreliable and because the actual users are the true test of the calls quality. After these steps have been run, it may be repeated from time to time as needed. It may also be of value to invite end users to see if they like the service. This step is necessary before rolling out the service just to see if it is really of good quality and worthwhile service to apply.