Dompeting or complementary technologies?

Telephone companies and cable TV companies alike are searching for the best technology for a new media distribution network from the central office or headend to the user. How will cable TV companies offer telephone service? And how will telcos enter the multimedia market? A number of issues must be addressed before either will be successful.

The ideal new media distribution network must have three characteristics: • It must support the transmission of broadband signals suitable for video and high-quality graphics. • It must allow customers to choose their programs independently, regardless of time of day or what others are choosing. • It must be bidirectional so that users can send program or database selection commands to the central site.

Both wireline and wireless transmission technologies meet these criteria. To create an instant network, a few telephone companies have purchased existing cable TV companies. U S West, for example, bought Continental Cablevision. Other telephone companies have turned to wireless cable service-such as Pacific Bell's acquisition of a multichannel multipoint distribution service system in Southern California.

Most operators, however, are looking for ways to broaden the range of services with their existing wired networks, whether they are copper, coax, fiber optic or hybrid fiber/coax (HFC).

Telco Options The major problem with copper telephone cables is that they were not originally designed for broadband signals. But they can be upgraded to support broadband transmission.

Asymmetrical digital subscriber line is one such technology. ADSL appeals to traditional telcos for a number of reasons: It uses MPEG-type video compression; it may be cheaper than building HFC networks, especially for small numbers of customers; it operates over most existing copper cables (although range is sometimes a problem); and it provides regular public network service plus four downstream 1.5 Mb/s video channels and one two-way 64 kb/s control channel.

Today, ADSL delivers 6 Mb/s downstream and a 64 kb/s control channel upstream. Although it may sound ideal, the electronics aren't cheap. Press reports of BT's ADSL trials in the United Kingdom, for example, claim initial per-subscriber costs in excess of $30,000.

Another option for telcos is to supplement or even replace copper cable with coaxial cable, fiber optic cable or HFC. Pacific Bell has begun construction of an HFC network in California and plans to gradually replace all of its copper. Other telcos have announced similar plans.

The biggest issue with this scenario is providing electrical power to the electronic equipment required at the optical/electrical interface. Reliability also is an issue. Recent studies by GTE predict downtime for the current HFC network design to be about five times greater than traditional (non-ADSL) telephone networks.

Fiber to the curb (FTTC) is a third option, representing deeper penetration of fiber in the distribution network toward the customer. In this option, only the drop wires remain copper or coaxial cable. FTTC networks, with a star configuration of drops, offer better transmission and lower noise than HFC, but they also have more O/E points and thus more power problems.

The choice between enhancing copper cables with ADSL or replacing the network with a cable that is better able to handle broadband signals is not an easy one (see sidebar).

A key question to ask in making this technology choice is: How many customers will subscribe to the new services made possible by a broadband network? If only a few customers will buy a new service, an incremental approach like ADSL may make the most sense.

On the other hand, if a carrier is confident that most of its customers will want a particular new service, then a network replacement plan is a better choice. Predicting customer behavior, however, is risky business when introducing previously untried services.

Upgrading Cable Networks Many cable TV companies are already upgrading their coaxial cable networks to modern HFC designs. These companies find that their greatest need is adding upstream control channels and user selectivity or switching capabilities.

The first issue is being partly resolved through an industry agreement that designates spectrum for upstream channels (Figure 1).

In the United Kingdom, for example, most cable operators have installed switches at their headends and-by cooperating with carriers such as Mercury Communications-are offering public network services to their customers. Meanwhile, Motorola and other suppliers have introduced modems for Internet access over cable.

But does it work? The full service network trial run by Time Warner in Florida demonstrated that it would-but not without enormous amounts of engineering and construction to deal with noise, especially on the upstream channels.

Providing the electrical power for remote electronics is no less of a problem for cable TV companies than it is for telcos. The decision on whether to use cable television networks for voice, data, and other two-way services is more of an economic choice than a technical one.

As expected, each camp has its strengths and weaknesses (Figure 2). The best design depends upon what a carrier is starting with, the range of services to be offered, the expected demand and the regulatory environment.

Regardless of the technology selected, a major problem for both camps is adapting billing and customer care systems for new services. Telcos are not accustomed to customer service issues regarding content, and cable TV companies are not set up to respond to detailed questions about telephone bills. And because neither has direct links to financial institutions, considerably more work is needed to find the perfect answer.

Steve Saltwick is Director of Telecommunications Sales for Tandem Computers Inc., Cupertino, Calif.