Whether the mega mergers roiling cable lead to intense new competition among top players or to new levels of cooperation, it seems safe to say that one result of the intensifying scramble to stay on top could be the eventual dominance of cable's infrastructure in local broadband communications transport.

This possibility rests on the fact that, no matter how AT&T and the new AOL Time Warner conglomerate work out their complex dealings, they've raised the stakes so high that they have no choice but to future-proof their networks as quickly as possible. And because they can do it in more or less uniform fashion across a newly standardized multiservice platform, they are assured of having the mass market scale essential to making their presence felt on retail shelves, where consumer purchases of modems, set-tops and other devices will decide who wins and loses in the broadband network services marketplace.

Ironically, the competitive threat AT&T and AOL Time Warner pose to each other could become the biggest goad toward maximizing their cable network potential against other platforms, because these are the entities who seem most motivated to serve the residential market with a full range of advanced services. That the threat of head-to-head competition between the two leading cable companies is real can be seen in AOL's aggressive broadband expansion strategy, which, as Robert Pittman, co-COO designate of AOL Time Warner, puts it, is to reach "every user in the country via cable, DSL, wireless and satellite."

That applies to AOL TV as well as the broadband-enhanced PC-based service, and it applies to Time Warner cable markets as well as other cable markets, where the AOL goal is to serve customers no matter what access means they use, says AOL spokeswoman Wendy Goldberg. With cable access now closed, AOL is focusing on the DSL and satellite options, with the satellite connection already on and DSL in limited commercial launches to existing AOL customers in several unnamed Bell Atlantic and SBC Communications markets.

"We're working closely with our partners to make the upgrade to DSL service as easy and convenient as possible for our customers," Goldberg says. This includes determining whether customers prefer the "splitterless" approach where they use the same in-home wiring for DSL and regular phone service, or the traditional approach that requires a technician to install a second wire in the home.

As for AOL TV, the company demonstrated its new service at the recent Consumer Electronics Show in Las Vegas and is planning to begin introducing the service commercially in the spring, Goldberg says, declining to say where the service will be introduced initially. AOL has contracted with Philips Electronics to supply set-tops for the DSL environment in support of AOL and with Liberate Technologies Inc. to provide the navigating software.

AOL has also worked out deals with DirecTV for supply of AOL TV via DBS and with TiVo Inc. to incorporate TiVo's personalized digital TV recording technology into the set-top.

Meanwhile, the need to reach a broader audience with broadband content has prompted Excite@Home, in which AT&T is the largest stakeholder, to launch a broadband portal on Excite, which will afford users of DSL and other broadband platforms access to much of the broadband-enhanced content developed for @Home, with the added benefit of streamlined delivery via the national broadband backbone @Home originally created for cable distribution. According to informed sources, Excite@Home was close to announcing the new portal at press time, with the likelihood that it would also name certain ISPs as partners in networking arrangements that would ensure the highest quality access to the portal through the Excite@Home network interface points.

AT&T has also made clear that it will deliver local phone and high-speed data services over fixed wireless facilities in markets where it doesn't have a local phone affiliation with cable operators. AT&T spokesman Mark Siegel says the company might also use DSL "on a selected basis" for high-speed data only, noting that the company's commitment to its wireless strategy remains the main focus as an alternative to cable distribution.

But Siegel stresses that the company is determined to maximize its voice services reach over cable through the types of partnerships it has tentatively arranged with Time Warner. "The AOL Time Warner deal only serves to affirm our view that cable is the right broadband platform," he says.

With completion of its acquisition of cable company MediaOne Group, AT&T Broadband and Internet Services will have a cable footprint covering about 25 percent of U.S. households. And, if it is able to come to terms with Time Warner on details of the voice services agreement that was reached with that multiple system operator nearly a year ago, its voice-over-cable footprint would extend to some 40 million households.

But that's a big if, as officials at Time Warner make clear. "The first order of business is to unravel the ownership question, and from that everything else will flow, including the voice deal," says a Time Warner official, speaking on background.

The ownership question concerns the fact that MediaOne now holds a 25.5 percent stake in Time Warner Entertainment, which includes the cable properties, programming networks and other assets. The two sides must come to terms on how AT&T's share of that ownership is to be structured, as well as what to do about the 32 percent stake MediaOne has in the Road Runner high-speed cable data service, which could pose regulatory problems, given AT&T's 58 percent stake in Excite@Home.

These issues were to have been worked out in tandem with the closing of the MediaOne/AT&T deal over the next few months, possibly leading to a rearrangement in the proportions so that AT&T would end up with a reduced share of Time Warner content properties in exchange for a greater share of the cable unit and completion of the voice deal. Now, with AOL as the new owner of Time Warner, such decisions will have to await conclusion of the AOL Time Warner deal, which will take the rest of the year, if not longer.

Further complicating matters is the fact that Time Warner is now preparing to enter the voice business on its own steam via the IP (Internet Protocol) platform that has been developed for the cable industry under the PacketCable banner. During the second quarter, the MSO plans to begin testing a second-line service over its Portland, Maine system, priced at $4.95 per month, with the intention of moving to widescale rollouts if the market trial pans out.

Similarly, Comcast Corp., another potential AT&T voice partner in cable, is moving ahead with its own voice service plans, using the new PacketCable platform, starting with a test involving equipment supplied by Lucent Technologies Inc. over Comcast's Union, N.J. facilities. Comcast last year joined forces with MSOs Cox Communications and MediaOne, before AT&T's bid for MediaOne, to form a united negotiating front with a combined footprint which they hoped would be sufficient to secure terms on voice services that would be comparable to those reached by the much larger Time Warner Cable, but now Comcast isn't counting on AT&T as it prepares for the leap into telephony.

In contrast to the growing support among cable companies for the new IP voice approach, AT&T is putting greater emphasis on the non-IP packet technology supplied by Arris Interactive, which it has been using since it began testing voice services in Fremont, Calif. last year and, even earlier, in Hartford, Conn. under the previous ownership of Tele-Communications Inc. Originally viewed as a stopgap while the firm awaited maturation of the PacketCable platform, the proprietary voice systems supplied by Arris and others have not only performed extremely well; their costs have dropped much faster than anticipated.

None of this is to say, of course, that AT&T can't move forward with telephony deals with Time Warner, Comcast or other cable companies. But the farther potential partners go without AT&T's participation, the harder it will be to put deals together, especially if the other companies are successful in their go-it-alone modes.

Meanwhile, AT&T BIS appears to be clicking on all cylinders with regard to all it has laid out as its internal cable agenda, including its efforts to improve on the underlying cable architecture as well as its move into other advanced service categories, including interactive TV, high-speed data and Web-to-the-TV applications. In fact, if work on the new architecture continues to progress along the latest lines outlined by senior engineers, AT&T could well end up revolutionizing the industry's approach to delivering services, with cost reductions affecting everything from capital spending on plant expansion to the means by which data signals are accessed by end users.

The prospects for dramatic improvements along these lines are reflected in plans for a third phase of the company's "LightWire" version of the traditional cable hybrid fiber/coax network. The company has been testing the first phase of this design over 600 miles of commercially operating plant in its Salt Lake City system with strong results, prompting it to begin using the design for upgrade projects that were already slated for this year under the old HFC approach.

As currently deployed, LightWire entails extension of fiber to "mini-fiber nodes" at interface points on the existing coaxial cable plant that have the effect of eliminating all in-line amplifiers between the node and end users. According to Oleh Sniezko, AT&T BIS vice president of engineering, this use of fiber reduces the number of people contending for bandwidth over any coaxial serving area and cuts the ongoing maintenance and performance hassles associated with use of traditional coaxial amplifiers.

Figure 1: AT&TÕs LightWire-II: Making it more passive. Source: Oleh Sniezko and Xiaolin Lu, AT&T BIS, 2000 SCTE Conf. on Emerging Technologies Proc. Manual, p. 505

But the long-term benefits of LightWire could be much greater, as reflected in the company's thinking about Phases II and III of the design migration, Sniezko says. "We want to do distributed processing," he says, which means that rather than keeping all the cable modem termination system functions that control distribution of data services in the headend, as is now the case, much of the CMTS functionality would be positioned at remote fiber node locations in the field.

I'd guess that in five years, distributed CMTSs will be used in 80 percent of our systems," Sniezko says.

The idea is to limit the tasks performed at the primary hubs to bundling and routing while moving the MAC (media access control) functions to the mini nodes in the LightWire architecture. These points of final fiber termination serving about 70 homes each will be able to enforce access policies locally, provide dynamic network security protection on a distributed basis and resolve upstream contention within the serving area independent of the rest of the network, Sniezko explains.

The key to this step is the ability to distribute everything from and to the headend over the fiber portion of the plant in the telecommunications industry's time division multiplexed format, rather than using the more processing-intensive and costly amplitude modulation process that characterizes today's cable distribution technology. By delivering interactive, dedicated signals, as opposed to the point-to-multipoint broadcast cable channels, in the digital baseband mode, AT&T will be able to use off-the-shelf components common to LANs to perform MAC functions at the mini nodes, Sniezko says.

Already, he adds, Broadcom Corp. and other chip suppliers are offering 100 megabit-per-second and gigabit Ethernet integrated circuits that would make it possible to interact with end users on the coax as if they were service nodes on a LAN. "We want to push the HFC (hybrid fiber/coax) network to look exactly like a PON (passive optical network) while still using the coax," Sniezko says. "If the network capacity is enough between the primary and secondary hubs, the narrowcast part of the (traditional cable) subcarrier multiplexing system can disappear," he notes.

The key to making all this possible is the emergence of a new generation of optical devices that can be tightly integrated onto electronic circuits. "The enabling devices exist today, but that must be packaged to our specifications," Sniezko says.

No other cable company has moved this far into the future, notwithstanding the likelihood that the more cable employs IP technology to support voice, data and interactive TV entertainment, the more bandwidth it will need to handle all the dedicated traffic. The advanced phases of LightWire not only address the need to efficiently manage such traffic across a streamlined optical infrastructure, but also the need to simplify provisioning of an increasingly complex array of services to end users using straight-forward digital telecommunications protocols as opposed to the more complex protocols of the current cable data and packet voice standards.

AT&T has also taken the lead in formulating an architecture for delivering advanced television services, which is beginning to appeal to Time Warner in light of the latter's recognition that tomorrow's service paradigm might require more functionality in the end-to-end client/server model than it once assumed.

Time Warner has been promoting its Pegasus digital set-top architecture, which it has been aggressively deploying throughout the past year, as a foundation for OpenCable, whereas AT&T has promoted a set of APIs for client (set-top) software and "middleware" linking clients and server platforms in the headend that assume a higher level of functionality in the set-top than Pegasus assumes. Now, however, Time Warner recognizes the system it chooses must integrate "many of the things that Jim (Wood, of AT&T BIS) has been talking about," says John Callahan, vice president of interactive services at Time Warner Cable (see below).

In a move that promises to bring the two MSOs closer on the overall OpenCable environment, Time Warner, which has been debating the question of "to middleware or not to middleware," is leaning toward a middleware approach where "some kind of virtual machine environment is the most likely candidate," Callahan says. Such an environment invites the kind of flexibility and processing power in the set-top box that AT&T envisions.

As described by Jim Wood, VP for advanced technology at AT&T BIS, this architecture includes a wide range of APIs that are designed to connect the set-top terminal operating system with server-based services via "middleware" that expands the range of services that can be delivered to end users.

Key components of this architecture include:

the broadcast service layer, which supports the ability of applications to "get hold" of TV channels or other applications in the box without relying on a click of the remote, whether those applications are in the electronic programming guide (EPG), the vertical blanking interval or some "out-of-band" location;

the HTML engine, an "absolutely critical" component that provides for navigation across different applications (services) and which, in AT&T's case, involves use of a specialized version of Microsoft's Internet Explorer navigator but which can also involve implementation of navigators in the Java Virtual Machine environment;

the foundation layer, the set of APIs for all services that provides access to underlying support mechanisms in the set-top, such as the graphics/video composition engine, the service policy aggregator and provider and the event/usage logging process;

the common content interface, which allows multiple Web sites to have access to the different resources in the set-top;

the EPG support module, which allows new EPGs to be loaded into the set-top;

and the Navshell Application, which is the master application running on the set-top to manage such functions as the provision of user graphical interfaces for various service categories, control over the combined graphical user experience as generated through the composition engine of the set-top, and the display of enhanced TV content.

All of this dovetails with AT&T's deal with Microsoft Corp. to introduce a set-top platform commercially this year that employs the Windows CE Plus operating system the software company has designed for interactive entertainment applications. If AT&T succeeds at drawing other cable MSOs to this platform, it will have more leverage in dealing with content providers as it attempts to generate advanced services that are distinct from those of anyone else, including AOL Time Warner.

Comcast, which, thanks to a $1-billion investment from Microsoft, is already in this camp, seems well positioned to emerge as a close ally of AT&T BIS, especially because it is partnered with the latter in ownership of Excite@Home and has expressed interest in doing a telephony deal. Sources report AT&T is also actively pursuing expansion of its 33 percent ownership stake in Cablevision Systems Corp., which, so far, has resisted doing a voice services deal with AT&T.

Questioned whether, in light of the reach of AOL as a provider of Time Warner content over DSL, cable and satellite, AT&T's partnering strategy might also involve efforts to build a cable coalition on content, Siegel replies: "We're looking at a variety of options, but I don't want to get into that."

Should these and other deals come to fruition, AT&T would be in a position to promote its new LightWire architecture on a much vaster scale, giving it a long-term migration strategy toward lower costs of delivering all types of broadband services in what would, ultimately, be a realization of the long-held passive optical network (PON) vision of what was once AT&T Bell Laboratories.

Roping in enough cable allies on the voice, data, interactive TV and network architecture components of its strategy to make its cable play really pay off may look like a long shot at this point; but if it succeeds, the company will have pulled off the evolution to the virtually unlimited bandwidth potential of fiber-based broadband technology at a fraction of the costs its competitors would incur to match it.

Whether AOL Time Warner somehow ends up in such a coalition or pushes forward with a competing coalition of in-house and third-party affiliations remains to be seen. But, whichever way it goes, it will be as motivated as AT&T to push the evolution of its various service platforms forward, not to mention the overall capacity and functionality of its cable networks. If all this means the cable-based fiber PON is closer at hand than once seemed possible, it's hard to imagine how carriers operating on other platforms will manage to match such network capabilities in anything approaching the same timeframe.