In "greenfield" builds, in which an organization deploys unified communications in a pristine environment, successful rollouts -- at least from the technical perspective -- are fairly straightforward. In most cases, however, UC is deployed on networks that already are supporting the business. In these cases, complex UC platforms are as dependent on fitting in smoothly with what already is in place as on the quality of the equipment the organization is buying.
"A lot of people in the field forget or don't know to benchmark existing voice and video services."
Just as it is impossible to build a solid house on a poor foundation, a good UC environment can't be set on a crumbling and ill-conceived infrastructure. And, just as the dangers of a creaky foundation grow as the structure gets higher, the problems introduced by a poor underlying network will manifest themselves to a greater degree as the UC environment grows deeper, more sophisticated and is used for more applications by a larger percentage of employees.
"The demand that VoIP puts on the network infrastructure is unlike anything that preceded it," says Jim Melvin, the president and CEO of Apparent Networks. "VoIP does not just require network connectivity, but predictability and consistent performance. The network really needs to be working right."
The first step, quite simply, is to see what is being dealt with. Experts say that a thorough pre-assessment is vital, and it should be conducted at a number of different levels. This should start, according to Dr. Michael Hollier, the CTO of performance management vendor Psytechnics, with a look at the level of service that end users are enjoying on the existing infrastructure. "A lot of people in the field forget or don't know to benchmark existing voice and video services," he says. "One of the little gems we learned is to do a quality benchmark before you press on."
The big leap that organizations are taking today is the transition from legacy infrastructure-known as time-division multiplexed (TDM) -- to Internet protocol (IP). This is the change from old phone networks to VoIP. Voice, of course, is the most important single application. Since IP networks combine it with other services and applications, such as instant messaging and streaming, the tests must assess how the network will perform once all the elements are supported by the same infrastructure. This is especially vital because there are a number of ways in which the applications can be mixed together.
More than Bandwidth Issues
This is true of all IP network builds. The final piece, where UC differentiates itself from the general IP networks, is the assessment of the way in which all of these services perform when they are linked together in a UC platform.
Bad performance at each of these overlapping levels will affect the level at which the entire system performs. However, the common thought among non-engineers is that an organization's UC implementation will be fine if there is enough bandwidth. This is enough of an oversimplification to be inaccurate.
Bandwidth is, of course, a vital element, but not the only one. The overused but very helpful highway analogy fits perfectly. A four-lane highway can carry a lot of traffic. But no matter how much overall capacity the highway has, its efficiency is affected if there are obstructions, detours and slow moving vehicles mixed haphazardly with those moving faster. To the person waiting for a bus, the bottom line is the same whether the problem is that the road is too narrow or it is strewn with obstacles: They are being picked up late.
In the world of telecommunications, the names for these obstructions are jitter, latency and delay. Not only do packets need to get through the system in enough number, but in the right order and, even within that the order, in a highly regular manner.
Next page: Less Room for Error