Going Back to Square One (VoIP Deployment)

Some issues seem to drag on for forever, especially with intermittent issues. The 5-percent static issue can linger for months while you provide call examples, your carrier works through them, and it asks you to retest.
After you finish every testing scenario you can and still have no definitive answer, start all over again from the most basic level. Any problem that can possibly befall your VoIP has a logical explanation. Looking too intently at the details may have caused you to miss the bigger picture, so review the results of all your testing based on the following perspectives:
Find a pattern to the problem. Look at time of day, day of week, geographic area, association of the problem with peak calling times, peak data transmission times, anything that links the occurrence of the issue to a larger environmental situation.
Review all configurations. The packet sampling size, codecs offered within the SIP server, hardware settings on VoIP phones, VLAN configurations on your LAN — everything.
Look at your network layout. If you haven’t drawn a diagram of your network, do so. It doesn’t have to be pretty. Simple boxes and lines work. You want to look at every piece of hardware and cable that connects it. Reviewing the network design may identify a bottleneck in the system that you were unaware of.
Identify all variables. Remember, not everything can cause a problem. If you have static, you’re looking for an electrical short functioning at the first layer of the OSI model. If it’s packet loss, you’re looking at the IP layer. After you identify the strata in which the problem is occurring, write down every potential piece of hardware, software, and cabling from your VoIP phone to the SBC of your carrier that are potential sources for the issue.
Isolate the variables. Construct tests to bypass the individual variables or use your VLM software to gather data on them so that you can either prove them to be clean or indicate whether they’re potentially the source of the issue.
Do the testing twice as methodically and twice as slowly the second time around. Something obviously slipped through the cracks, so write down all the testing that conduct, including the time, date, and result. Follow your testing where it leads. The truth is out there.