Pac Bell commits to hybrid networks
PAUL PALUMBO
The ongoing debate in broadband network technology circles about which fiber-rich network to deploy currently pits advocates of switched digital video fiber-to-the-curb against proponents of hybrid fiber/coaxial cable networks. For its part, Pacific Bell is committed to hybrid fiber/coaxial cable and has gone to great lengths to support its decision.
Although many subtle distinctions can affect the capabilities of both network architectures, the chief dividing line between the two is that switched digital video takes fiber to the curb or deeper, into neighborhoods. On the other hand, the hybrid fiber/coaxial-cable architecture is a less media-intensive design that runs fiber to a neighborhood node. Pacific Bell`s internal analysis of hybrid fiber/ coaxial cable versus switched digital video network design costs suggests that the latter is as much as 30% more expensive.
According to Pacific Bell`s chief network architect, Keith Cambron, the telephone company prefers hybrid fiber/coaxial cable because there are currently no commercial deployments underway involving switched digital video (nor are there likely to be for several years), and forecasted user demands simply do not warrant the return-path bandwidth.
The substance of the debate, according to Cambron, boils down to whether you believe there will be "large amounts of data coming from people`s homes, as opposed to large amounts of data going to people`s homes. Assuming you believe broadband data to be `symmetrical` in nature (namely, data packets moving equally to and from the home), then switched digital video makes more sense."
Commenting on the architecture debate, Tony Robertson, a tele communications analyst with Robertson & Stephens Co. in San Francisco, says the decision to deploy switched digital video or hybrid fiber/coaxial cable involves a "judgment as to what proportion of users will be satisfied with a large number of traditional cable channels, versus some kind of video-on-demand. The answer might be different for diverse areas of the country, or even [within] communities."
Cambron admits that companies opting for switched digital video will eventually get there by building hybrid fiber/coaxial-cable networks first and then "overlaying" switched digital video with a fiber-to-the-curb design.
Other Pacific Bell strategists expect that early on, most revenue-generating traffic will be analog (one-way video, two-way voice).
Moreover, Pacific Bell planners argue that a pure, switched digital video solution requires set-top boxes on every device in the home. It`s unlikely, they contend, that net subscriber capture would warrant the cost at this juncture. Pacific Bell`s business plan envisions that telephony and broadcast TV are going to be its two primary revenue sources well into the next century.
Cambron says there will be digital revenues (by means of cable modems, for example); however, Pacific Bell believes those new market positions are going to take a long time to develop. The main issue for the company is to build its network so it can accommodate a variety of traffic. By the time its broadband network is deployed, Cambron asserts, fully 70% of the traffic will be in video bits (analog and digital broadcasts, as well as interactive [game] business segments). Cambron believes that whatever network architecture is going into the ground or strung today is working on a 5- to 15-year life cycle.
From an investment standpoint, Pacific Bell notes that hybrid fiber/coaxial cable has already passed several key milestones that make it the most appealing choice today:
Feasibility: Hybrid fiber/ coaxial-cable networks have proven they can achieve the performance necessary to deliver broadband applications.
Economics: The cost of installing hybrid fiber/ coaxial-cable plant (about $750 to $800 per home passed for Pacific Bell) is acceptable, and the company believes that it can make money using the architecture.
Cost reduction: Hybrid fiber/coaxial-cable architecture is entering the adolescent phase of its deployment evolution, where costs are beginning to come down.
Cambron says that two areas are being attacked for cost reductions: optics (including lasers, receivers, optical transceivers and amplifiers) and radio-frequency modulation techniques. For example, Pacific Bell`s network will initially use 64-quadrature amplitude modulation for AM-modulated video downstream because it is commercially available.
But Cambron expects 256-QAM systems to begin appearing by next year, and that migration will increase throughput from 6 bits of digital information for every hert¥of analog bandwidth to 8 bits (+33%) per hertz.
"For the return paths, the multitone techniques appear to have promise," says Cambron. What Pacific Bell is looking for in the return path is a common bandwidth management approach, as opposed to time-division multiplexing technologies it uses today. That approach will allow the company to "make much better use of forward and return bandwidth by sharing synchronous mode traffic with asynchronous mode traffic and thereby expand the use of the return path significantly," according to Cambron.
Using AT&T as its systems integrator, Pacific Bell has designed its network so that each 500-home node is split into four quadrants, and it dedicates full reverse bandwidth (5 to 40 MHz) for each 100 to 120 home node. Three fiber links deliver up to 750 MH¥downstream to each 500-home optical network unit. q
Paul Palumbo writes from Seaside, CA.