Systems based on the simplicity of Ethernet and the power of optics will dominate the broadband access mix.
HERB MARTIN, Salira Optical Network Systems
Who can forget the Apple Newton? Apple probably can. The super-hyped handheld gizmo that burst onto the scene some years back-the premier entry into the then-new category of personal digital assistants (PDAs)-was going to keep people organized, punctual, and operating at peak performance. For a short while, it was the new, new thing. Whip one out at a meeting and you were guaranteed immediate attention. Great notion. Wobbly technology. Outrageous price tag. Newton stumbled, then fell-a victim of economic and technological gravity.
But Newton's conceptual descendants proliferate today, of course, in the form of Palm Pilots and the like. It's a device category we've come to take for granted-simple, reliable technology in the right package at the right price.Today's access networks share more with the nascent struggles of the PDA than a lot of people in the communications industry would like to admit. Who doesn't want faster access to the Internet? But even now, access technologies, especially those with costly and complex ATM underpinnings, look a lot more like the Apple Newton than the Palm Pilot (see Photo).
Just as there is proliferation in the types of PDAs now available, so too will there be proliferation in varying types of broadband access. As is shown in the Table, the coming broadband access mix will contain a veritable stew of coexisting and competing access methods. The common thread among the successful technologies in this mix will be the ability to deliver access that's simple, powerful, and affordable. Systems based on the simplicity of Ethernet and the power of optics will dominate the broadband access mix.
While the raw horsepower exists in many of these solutions, failure to harness it in a mass-consumption fashion leaves unchanged the fundamental disconnection that still exists between the end user and affordable broadband services. To the end user, less is more: less cost associated with high-speed Internet access to voice, video, and data services; and less hassle in modifying services (self-provisioning) when need be. And more is better: more ability to allocate bandwidth dynamically through an e-commerce model, more service providers to choose from, and more richness of multiservices.
Today, access technologies, especially those with costly and complex ATM underpinnings, may have more in common with the Apple Newton than the Palm Pilot. The spectacular failure of the Apple Newton was caused by the product's high price tag, steep user learning curve, and lack of interoperability. In contrast, sales of the Palm Pilot (shown here) and its descendants have exploded precisely because they are affordable, simple to use, and able to interact with other personal digital assistants and computers.
For carriers, a dramatic leap in cost-effectiveness enabled by a solution based on Ethernet is just the beginning of the checklist. Emerging solutions must deliver a compendium of benefits that enable widespread broadband access. Simple remote provisioning cuts expenses and unlocks lucrative revenue opportunities hitherto unavailable. As new subscribers come online, capital expenditures associated with build-outs can keep pace. As end users demand more bandwidth-which is suddenly more affordable-the network infrastructure can grow accordingly. Value-added services protect against price erosion and guard against competitors' inroads. Finally, better deployment of space in the central office and even more efficient use of electrical utilities enhance overall return on investment.
The transition of the access market calls for an affordable solution that's Ethernet-simple and fiber-fast. To deliver the promise of fiber-fast transport to the mass market, point-to-multipoint or passive optical networking is the architecture, not the finished structure. It's the basis upon which the superior solution is designed. Does point-to-multipoint yield our optimum solution? Partially, yes. Carriers must be able to distribute broadband services to unprecedented numbers of users, not just a fortunate few with access to DSL or commercial sites budgeted for T1 (1.544 Mbits/sec) and T3 (44.736 Mbits/sec). But point-to-multipoint is nothing more or less than topology. It's a way to leverage an efficient transport topology.
No one would dispute the superiority of fiber optics over copper as a transport medium. Since 1996, 50 million miles of fiber-optic cable have been laid-nearly 40% of it last year. The real issue is cost-effectiveness-economics preclude the old point-to-point, ATM-driven solutions. More users, using more bandwidth fanned out from fewer central locations, forces costs down not just in terms of build-outs, but also in equipment and maintenance, among other factors.
What about the traffic? What form does it take? Just as "think globally, act locally" became the mantra of political and environmental activism, affordable access may follow a similar directive: wide-area thinking and recognition of the local requirements of LAN traffic-95% of which is related to Ethernet-based applications. Thus, we are witnessing the advent of Ethernet passive-optical-network (EPON) systems in which optics can be precisely delivered to end users, and the end-user traffic is aggregated and sent back to the network without using complicated protocols. Design globally, implement locally.
Finally, we circle back to affordability. EPON enables simple, flexible provisioning of voice, video, and data traffic. But the ultimate package, the one that will resonate in the mass market, must come within economic reach of that market. Part of the technological challenge is that the Ethernet/IP ingredient cannot be regarded as an afterthought, add-on, or even an add-in to the point-to-multipoint architecture that accommodates LAN applications associated with the mass market. So far, that is exactly how it's been approached. The inadequacy of the results to date speaks for itself. The optimal solution calls for a system designed from the ground up as an Ethernet solution. It's not the result of an equation-no matter how elegant-that merely factors in the IP protocol. Hardware and software that promise compelling benefits to carriers and end users must be based upon a view of EPON as the architectural basis of the solution and not the solution itself.
To deliver on the lofty promise of accessible access, end-user requirements for simplicity and affordability must be factored in from the beginning. Carrier requirements for flexible provisioning of profitable, value-added services must be incorporated into a design model that views the network holistically. Who stands the best chance of success in this new age? Bet on the vendors that deliver carrier-class, Ethernet-based products that move the world beyond inaccessible access.
Herb Martin is president and CEO of Salira Optical Network Systems Inc. (San Jose, CA).