June 1, 2007

Telephony: The Top View Emphasizes Bandwidth

By Justin J. Junkus

Although people in cable tend to change companies, the fact that most of us keep working for some part of the industry gives us a pretty good sense of how things work and where the industry is going. Some of us who enter management also have the benefit of a larger view. This month, I'm sharing a conversation I had with Jim Chiddix that provides a perspective on bandwidth management, business services and telephony, as seen by an individual who is both an industry veteran and a member of the boards of Symmetricom and Vyyo.

JJ: Jim, a lot of what you've done in the past has been involved with cable's video side of the business, and now you've made a commitment to Symmetricom, which has been a strong player in telephone network synchronization. Does this say something about your views on convergence in the communications industry, or is it more a look at new markets for Symmetricom?

JC: A little bit of both, actually. Apart from its markets, Symmetricom interests me from a technical standpoint. I'm an old ham radio operator, and Symmetricom's products are a very precise implementation of the oscillator technology I used to build for ham radio transmitters. But beyond that, I do see a convergence, particularly in the area of network timing. Cable is offering more data services that bring them some challenges in terms of packet delivery and timing. Similarly, telcos are offering video and that has its own challenges in timing and transport.

JJ: If I look at some of your latest business activities, bandwidth seems to be one of your pet interests. Having spent a lot of time talking to cable about telephony, I'm curious how you see bandwidth bringing new telephony and telephony-related opportunities to cable. What's on the horizon?

JC: It's no surprise that cable sees a value to commercial services. Small to medium business in particular have been underserved. They can get T-1, but only at substantial cost, and some have been using DOCSIS to meet their bandwidth needs. There's a lot of money being spent on these services by small to medium businesses. I'm struck that cable's HFC plant usually passes within a block or even yards of these businesses, and that presents many opportunities. It could be voice over Internet protocol (VoIP) in terms of features, connectivity between branch offices, T-1 trunks for their private branch exchanges (PBXs), or delivery of T-1 to cell sites.

But regardless of what it is, the service is going to have one key characteristic, and that is the need for more bandwidth. Vyyo plays in one part of that game, but delivering technology to upgrade one leg of an operator's plant at a time. Symmetricom plays in another part by enabling a way to bond channels together for a much fatter, higher bandwidth channel. So it all ties together from my standpoint.

JJ: How are DOCSIS 3.0 and modular cable modem termination systems (M-CMTSs) going to affect the way CMTSs are engineered for telephony services? If it's about leveraging equipment and sharing applications use, are we still going to have enough CMTS capacity to share that piece of equipment among telephony, video and data?

JC: Electronics cost is about volume, and it makes sense to share equipment. As DOCSIS 3.0 gets introduced and more usage occurs, it's still a matter of network engineering, an old problem that both dialup and digital networks have had to deal with for years. But (implementing new architectures) will involve continual pressure to have enough bandwidth - using bonded DOCSIS channels, for example. The good news is that it all goes with revenue.

JJ: What is the future of timing, particularly for time division multiplexing (TDM) services?

JC: Video will be a driver for upgrades to existing TDM equipment. As telcos launch video over IP, there are timing specs that will require upgrades of their TDM timing shelves. Beyond that, there is a need to manage IP timing as packets move through a network. These timing and quality of service (QoS) issues are brought about by very different technology than cable has been using. Network management is going to require new tools that examine dropped packets, jitter and delay across an entire network.

JJ: Any other words about cable's opportunities and future?

JC: The original plumbing remains very useful. Fiber plays a much more important part, but the fundamentals haven't changed. There's been a tremendous amount of capital and labor in putting in this plant, and finding ways to continue to use coax makes practical sense.

Timing

Jim is obviously going to be heavily involved with bringing new synchronization and timing to cable. And speaking of timing, my good friend and co-editor Ron Hranac brought to my attention a timing discrepancy in my April column, which I'd like to correct. I mentioned that 5 nanoseconds (ns) is the time it takes for light to travel 50 feet, but Ron points out that the speed of light is 299,792,458 meters per second, or 983,571,056.43 feet per second. Therefore, light travels 1 meter in 3.34E-09 second, or 1 foot in 1.02E-09 second - roughly 1 foot per nanosecond.

So, in 5 ns, light travels 5 feet. That is, of course, for light in a vacuum, or at the high altitudes of Denver - I must have been thinking of the denser air in the Midwest! Thanks, Ron, you are absolutely correct.

Test tip

And finally, here's a test tip for the month. All this talk about bandwidth reminded me that VoIP quality can be adversely affected by bandwidth crunch anywhere in the call path. The Mean Opinion Score (MOS) is a parameter that measures the quality of a call, as heard by the subscriber. In the past, the only way to get a MOS was to assemble a group of "listeners" who rated the call on a scale from 5 (ideal) to 1. Now test equipment manufacturers have figured out how to build an accurate MOS from a set of other parameters, and MOS is included in VoIP test equipment. Even if your network has been engineered properly, the subscriber only cares about end-to-end call quality, and your company is the one that will get the trouble call. MOSs can be obtained from points within the call path, as well as at the call end points, and provide a way to identify exactly where voice quality is degraded. Get to know what equipment your company has to measure MOS and the location of network test points.

Justin J. Junkus is president of KnowledgeLink and telephony editor for Communications Technology. Reach him at jjunkus@knowledgelinkinc.com.