VoIP views

VoIP musings.

Regarding “Quality question remains for VoIP :” This past year I rolled out a new, 1,000-phone PBX. I considered voice over IP briefly but decided that it wasn’t ready for prime time, for the following reasons:

• Up time: Our phones need to be up 24-7; most networks are not. Don’t rely on a transport layer to support phones that is not as reliable as what you require.

• Converged infrastructure: This sounds great, but what if you are on the phone and a denial-of-service attack is launched against anything in the path of the call? Phone traffic has no business on the data network unless it is very low-priority traffic, or it is too expensive to run a second network to a location.

• Price: VoIP equipment still costs at least five times as much as analog equipment, and probably two to three times as much as digital equipment. If your telephone cable plant is in place, why not use it rather than pay someone to pull it out?

• Maturation of equipment: As far as I could find, there is no interoperability of different brands of systems. Sure, you can call a phone by a different vendor, but the bells and whistles that end users come to expect don’t generally work across brands. Once this problem is straightened out, VoIP has a much better shot at taking off.

Our break-even point for the new PBX is a little more than two years. After that, it’s free. The difficulty with VoIP might be taken care of before we have to replace our switch again; only time will tell.

After reading “Quality question remains for VoIP,” I feel it necessary to clarify the statement that “The result of latency is jitter, which can cause an IP voice conversation to break up.” The variation in latency on a data network is what results in jitter, not the actual delay itself. Jitter buffers on most VoIP products are dynamic, meaning they resize themselves according to conditions on the data network. If the delay is relatively constant, the buffers don’t need to work as hard, and voice quality doesn’t suffer as much. Under these circumstances, as long as the total delay doesn’t exceed a customer’s budget, jitter doesn’t contribute as much to VoIP quality as other network factors.

On the other hand, as most IP networks aren’t managed with comprehensive quality-of-service (QoS) policies, the variation in delay can be dramatic, and therefore damaging to VoIP streams. QoS policies aren’t enforced to just reduce delay, but to limit the variation in that delay. Delay isn’t the only cause of poor-quality VoIP conversations. There are other factors involved.

What I never see mentioned is the old adage about putting all of your eggs in one basket. In the 1980s, those of us on the voice side were trying to convince customers to run data through the PBX. But end users feared that if the PBX went down, they would lose all communication, both voice and data. Now it seems that no one is too worried about this. The PBX has proved to be a much more stable platform than the network. It might be less expensive to go with an IP solution, but is it worth the instability? The best solution, at this point, might be to use a combination of these two platforms.