If only unified communications (UC) could be as simple as the ‘two cans and a string’ method of yesteryear. But it’s not, and as more businesses embark on unified communications projects they are failing to account for unplanned spikes in traffic and, as a result, are experiencing network downtime.
A lot about unified communications is murky, from the terminology (as Josh Finke illustrates in this post), to the deployment (are you just integrating voice and text, or is it more complex?). Whatever the case may be it is important to understand that UC is a growing and complex organism, but by base-lining your network you are one step ahead in understanding and eliminating problems that may surface with increased UC traffic.
There are three data elements you should measure in your UC implementation:
Overall utilization: Per segment and per server, so you can see both overall effects and the effects on specific applications.
Top talkers: On a routine basis, who/what is most active? Hopefully these are not individuals, but servers. Remember to look at this overtime – during the day it might be the application servers, but at night it might be your backup servers.
Top protocols and applications: Again, look at this overtime as it is highly time dependent.
These measurements should become more granular based on the complexity of your UC implementation, and if you have a VoIP system in place or other areas relating to UC, these are always good to measure as well.
Remember that all of these are dependent on time, so it is important to measure daily, weekly, monthly and even seasonally – does your network experience a spike of activity in the Spring around Wimbeldon or in the Fall during hurricane season? Also keep in mind that UC is ever changing, so always plan for your current implementation as well as for future forecasted requirements.
Check out what WildPackets has to offer in network base-lining.