October - 2001 - issue > Cover Feature

In early 2000, at the peak of the CLEC boom, the challengers to incumbent Regional Bell Operating Companies (RBOCs) represented almost 20 percent of the market capitalization of all telecom services companies. In the past year, that share has fallen by 90 percent to represent only two percent of the total. The incumbents — SBC, Qwest (US West), Verizon and Bell South — once again rule the roost. In fact, together with the big IXCs, the incumbents represent more than 85 percent of all capital spending in the industry. On the other side, in the last six months alone challengers such as Teligent, Winstar, Covad, Northpoint, Rythms, and 360 networks have all filed for Chapter 11 (see figure 1).
And, of course, equipment and component vendors have felt the pain of their customers.

So, does this mean that the Telecom Act of 1996 was unsuccessful? It failed in some ways. But it has also changed the face of telecom forever. In terms of forcing a government-mandated competitive environment, the Act did fail. But, in that process the demand and end-user expectations have significantly changed. The implications are that those competitive service providers with sustainable business models will have a chance at success. Similarly, in the equipment space the rate of innovation has dramatically changed. From three- to five-year innovation cycles driven by telephony requirements, the cycle is now six to 12 months.

All Is Not Lost

Despite the mayhem in the service provider market, we are in a supply-constrained, not a demand-constrained, world. McKinsey and JP Morgan forecast strong growth in consumer broadband, driven by DSL and cable. Currently, 10 million U.S. households have broadband access. By 2005, this number is expected to grow to more than 40 million, plus four million small businesses.

Layer on top of that the increasing demand for IP-based value-added services such as virtual private networks and you have a long-term opportunity to build service-enabled networks.

End-User Pain

For end-users, the ability to get broadband services on demand is still not there. Often it takes six to eight weeks and several technician visits to set up connectivity. In the business world, if an enterprise wants T-1 or DS-3 connectivity, it has to wait three months. The demand is robust, but no one supply demand in a manner that is sustainable, scalable and profitable.

Investment Flows in the Future

Over the last 10 years, roughly $100 billion has gone into building the next-generation long-haul data backbone networks that span the country. Level 3, one of the eight major backbone providers, boasts 12 conduits, each with 96 fiber pairs, each fiber carrying 40 DWDM channels of OC-192 (i.e., 10 Gbps) each. The result? 440 terabits per second of capacity in one network alone. This equates to enough capacity to carry all of the nation’s telecom traffic! That capacity increases by 16 times if technology advances on the horizon, such as 160 DWDM channels at OC-768 (i.e., 40 Gbps) speeds are used. Hence, regardless of the controversy regarding “dark capacity” vs. “lit capacity,” the reality is that for now there is significant capacity in the long-haul portion of the network and opportunities in this segment are limited.

However the edge market, commonly called the “on-ramp,” composed of the first mile and the metro market, has not seen the kind of investment that the core has seen, and remains a bottleneck.

This market has a number of attractive characteristics (see figure 2). From a service provider standpoint, it is a point in the network where traffic from different customers is aggregated and services delivered for real revenue. From the equipment manufacturer’s perspective, it represents higher intelligence and higher margin products. Various technology options are being deployed in the last mile (DSL, cable, passive optical networks, Ethernet, etc.) and the metro market (next generation SONET, DWDM, IP routing, and aggregation and subscriber management). All of these elements must have service visibility as well as the ability to intelligently react to the type of service or function required.

So what are the implications for the evolving network architecture and business models?

Provisioning and Operations

In this area the equipment vendor community has failed the service providers. There have been many promises about equipment that is easy to inter-work, provision and operate. While there has been progress, significant work lies ahead. The speed with which services can be created and delivered will depend on progress in this area. A service provider business model cannot be based on expensive truck-rolls for initial provisioning and activation. Costs will drive the decision-making.

Service-Enabled Network Edge

If mission-critical services have to be delivered in a consistent and reliable fashion so that service providers can provide and live up to service level agreements (SLAs), the network must be able to increase the reliability and consistency with which it carries traffic. Quality of service and class of service mechanisms (DiffeServ and MPLS) that distinguish between traffic and “guarantee” delivery will be critical to delivering these services. These attributes must be set at the network edge and carried across the backbone. This will also enable latency-sensitive voice and video traffic to be carried reliably, leading to true convergence.

The idea here is to consistently and reliably deliver services that customers need, and to not only drive revenue today, but also to add capabilities to the network and deliver new services that can drive high-margin revenues.

Common Control Plane

As the core becomes all-optical and different network layers such as transport (DWDM, SONET, Ethernet), switching (ATM, frame relay, optical cross-connects) and routing (IP, MPLS) all need to work together to deliver services, the development of a common control mechanism becomes critical. A common control plane also means a common operational model and easier and faster provisioning.

The result is a unified network that combines all of the elements of intelligent transport with the elements of service aggregation and delivery. This network will allow many advanced services to be supported at the lowest possible cost thanks to a single network layer that simplifies provisioning, network management, billing and control. This evolution will take at least a decade, but major elements will be in place in three to five years.

Regulation and Government Action
Federal regulation could significantly accelerate spending levels. Several bills are working through Congress, as well as rulings by the FCC, that could materially impact a recovery.

HR 1542 (Tauzin Dingell) has been moving through the House of Representatives for the last four months. This bill is strongly supported by several of the major RBOCs, including SBC and Verizon. In its current form, the bill would require the RBOCs to roll out high-speed access to 20 percent of their COs within in one year, 40 percent within two years and 100 percent within five years. It would also aim to protect the RBOC investment by waiving the unbundling requirements of the 1996 Act for new access networks. The net result could be a revitalization of the RBOC monopolistic advantage.

Former FCC Chairman Bill Kennard, recently noted that “the incumbents are trying to seduce members of Congress and financially-strapped equipment manufacturers into supporting legislation that would effectively monopolize deployment. They are saying, ‘just give us a monopoly and we will make sure that everyone is served.’ I call this the seduction of monopoly. It would reverse years of pro-competitive development.”

Needless to say, we view this bill skeptically. Even though it might accelerate spending in the near term, the effects of reduced competition could have severe repercussions over the long term by retarding competition. In contrast to H.R. 1542, Senator Fritz Hollings recently introduced S.1364. This bill is focused on penalizing those service providers who do not comply with the unbundling requirements imposed by the 1996 Act. This bill creates an environment favorable to sustained competition and therefore sustained spending. Not surprisingly, there is minimal RBOC support for this bill. Pending legislation over the next six to 18 months could materially impact the telecommunications spending environment. If the right combination of bills pass, we could see an acceleration of RBOC spending as early as the second half of 2002. Conversely, if the wrong legislation passes, spending could be deferred well into 2004-2005.

The Road Ahead

The shakeout in the telecom sector is a healthy event that will cause weak service providers and suppliers with unsustainable business models to fall by the wayside. The market is returning to more classical growth, spending and valuation metrics. The underlying demand drivers are strong. Required changes to the infrastructure are real and will therefore drive growth in selected areas, especially at the edge of the network. Entrepreneurs with ambitions of hitting it big still have sufficient opportunities. The weather is clearing, but it is still “slippery when wet” — tread carefully!

Anand Gowda is a principal at Carlyle Venture Partners. Shailesh Shukla is vice president of strategy at Redback Networks. si