by Mitch Shapiro

"Imagine a world," says Mike Bookey, "where private monopolies owned the public road system and set the access rules. Federal Express might be allowed to use the local roads to deliver its packages, but not Airborne, UPS and other competitors."

Most of us would find such a world inefficient, unfair and downright offensive. As his listener's sense of outrage begins to stir, Bookey continues. "Why should the building and operating of the digital road system," he asks, "be any different than the public road system?" "Hmm…well, uh…let me think about that a minute," replies the listener. Bookey's achieved his first goal. He's got your attention, and, more importantly, he's got you beginning to think outside the telecom-industry box.

Meet Mike Bookey, and the concept of Community LANs--gigabit-speed optical LANs owned and controlled by local communities. In contrast to the vertically-integrated gatekeeper business models preferred today by cable operators and ILECs, C-LANs are wholesale "IP-utilities" that provide local high-capacity managed IP platforms designed to support a wide range of competitive retail service providers.

Bookey has spent most of his 30-years in the datacom industry using networks to transform economic and social institutions. From the early 70s to the mid-80s he was a pioneer in the development of electronic banking systems. In the late 80s through the mid-90s he was a leader in the movement to wire school computers to each other and to the emerging Internet.

Today Bookey is among those who see a growing need for networking initiatives at the local community level to parallel the rise of the global Internet. Optical C-LANs, he argues, can provide a strong foundation for creating: 1) abundant, ubiquitous and standards-based network capacity; 2) healthy competition and reduced barriers to entry; 3) rapid development and provisioning of new broadband services; 4) substantial savings in transportation costs and; 5) enhanced economic growth and quality of life. Like many in the industry, he questions the ability of existing local broadband platforms (DSL and DOCSIS cable modems) and the associated industry structure (vertically integrated duopoly) to efficiently achieve these goals.

Bookey's concept of C-LANs is similar to the notion of Public LANs (P-LANs) outlined in a White Paper published last year by Nortel. The paper described P-LANs as "city or regional networks modeled after the Internet, but designed like corporate Intranets. They combine the openness of the Internet with the performance and economics of local area networks." P-LANs' geographic focus, said Nortel, "allows for improved quality/performance and positions them as innovators for advanced Internet applications."

The Nortel paper concluded that the "unique architectural design" of P-LANs, "based on providing a simple, yet comprehensive address plan and large amounts of end-to-end bandwidth…offers a new model that can be leveraged by users and providers alike."

To help make that new model a reality, Bookey has founded two companies dedicated to designing, building and managing optical C-LANs. The first, ViaLight, already has an optical C-LAN operational at Issaquah Highlands, a 3,350-home residential development 15 miles east of Seattle that is the future home of Microsoft’s second major business campus. The goal of his most recent startup, FTTX Systems, is to develop C-LANs in new housing developments and through alliances with public utilities.

In Fat Pipe's March issue, an article entitled "Death Trap" by editor Gary Kim began with a statement that reflected the thoughts and feelings of many in the industry: "We hate to say it," wrote Gary, "but the Telecommunications Act is dead." The article closed with a warning to "get ready for monopoly, at least in the consumer space."

In its review of the regulatory and competitive landscape, the article cited the Open Market Plan model adopted by Rochester Telephone as one approach with potential for salvaging competition in the local loop. The ILEC (now part of Frontier Communications) had split itself into a common carrier wholesale infrastructure company and a retail sales organization. As Gary noted, "the Rochester model provides a business model clearly differentiated from what we seem to be facing today."

The same is true of C-LANs. And, importantly, they don't require ILECs to voluntarily split themselves into separate wholesale and retail entities, nor legislation that mandates such break-ups--neither of which seems very likely.

In light of recent industry developments, some have argued that the local broadband market is regressing to a private monopoly/duopoly mode of operation and that this trend calls for regulation to compensate for the resulting distortions in market function.

A better approach, says Bookey, is for direct public control of the network. Local digital networks, he says, need to be controlled by the community for the same reason local roads are: they are too important to the community to have decision-making in the hands of remote executives and shareholders.

Bookey's arguments rely on the premise that ubiquitous optical last-mile networks, like roads, are natural monopolies (or, at most, duopolies). The recent cutoff of capital and subsequent collapse of competitive CLEC and DLEC sectors suggests that capital markets increasingly view local broadband networks as natural monopolies or duopolies. And incumbents often appear to confirm this notion in their presentations to investors about the market impacts and financial viability of upstart competitors.

As is the case with roads, says Bookey, it makes most sense economically and logistically for a community to build just one high-capacity open-access network. He challenges critics of public ownership to present convincing evidence that private monopolies provide better overall value than publicly-run monopolies.

Bookey sees other parallels between local roads and C-LANs. Both require a comprehensive addressing plan (street addresses; IP addresses), active routing technology (smart traffic lights; high-speed routers), connectivity rules (rules of the road; network protocols), security and rules enforcement (laws & police; network security and privacy protection). And, importantly, both are open to the general public on a fair and equal access basis and achieve beneficial "network effects" by interconnecting every location within a community.

And just as a combination of public funding and usage fees (e.g., gasoline taxes and tolls) have financed road construction and maintenance, Bookey believes a similar financial model makes sense for C-LANs.

In our public road system, private industry makes its money by building roads, manufacturing concrete, asphalt, traffic lights, etc. and using roads to deliver goods and services. But the public retains ownership and makes key policy decisions (e.g., when and what roads to build), sets the rules, insures open access, etc. Bookey argues that a similar set of relationships makes sense for C-LANs.

Among those likely to benefit from the proliferation of C-LANs, says Bookey, are CLECS and IP service providers that would have access to a high-capacity optical network on a fair and equitable basis.

An assortment of vendors would also benefit. These would include suppliers of switches, routers, servers, optical cable and components, service providers, information appliance manufacturers, home automation vendors and software developers.

Last but not least, are the benefits to the local community, starting with lower prices and improved services generated by healthier competition among retail service providers. Another potential benefit, says Bookey, is significant savings in road construction, mass transit projects and commuting costs if even a modest percentage of car trips could be transferred to the digital road system.

Bookey also foresees major benefits to a community if every one of its government, utility, public service, education and health care facilities had a symmetrical gigabit-per-second link to each other and to every home and business in that community.

Critics of public ownership argue that, in addition to being poorly equipped to manage a telecom business, "public" entities would have unfair competitive advantages relative to private-sector incumbents, including access to lower-cost capital, close ties to local governments and, in some cases, tax-related advantages.

Bookey counters that incumbents could benefit by exploiting the vast capacity of optical C-LANs. As incumbents, he says, they would be well positioned to retain existing customers, especially if they use C-LANs to deliver new services not readily supported on their existing networks. C-LANs would also allow them to compete for customers outside their network footprint, focus investments on high margin services and free themselves from costly upgrades and maintenance of legacy copper and coax networks.

And, since the transition from telephone and cable last-mile infrastructures to C-LANs would be gradual, says Bookey, incumbents would have ample time to transition their business services and investments. And, to the extent they use C-LANs to deliver services, incumbents would benefit from the latter's access to low-cost capital.

As critics point out, the history of public networks has had its share of failures. They cite reasons ranging from too much politics and bureaucracy to insufficient technical, managerial and marketing expertise.

Why, they ask, should be we expect C-LANs to perform any better than the institutional networks built by cable operators since the early franchising days? Bookey contends that community-wide IP-based optical C-LANs operating in an increasingly IP-based world would be a far cry from these early I-Nets. These pre-Internet local networks typically grafted proprietary datacom technologies to low-performance cable networks designed to deliver analog broadcast video. In many cases cable operators did not view I-Nets as an investment opportunity but as a cost they must incur to win a local franchise. This attitude was not conducive to strong long-term planning, management and maintenance of these networks. And, importantly, I-Nets were not well integrated with the residential cable network.

Some public network supporters, citing the industry's current rash of business failures, challenge their critics to prove that the private sector has had a significantly better record of success.

Bookey argues that the real issue is not whether early failures occur, but whether their lessons are learned, whether a business model is fundamentally sound and--if given a chance to take root in reality--it can achieve desired results faster, better and cheaper than the alternatives.

Though Bookey believes C-LANs have a place in metro markets, he sees the nation's smaller towns--especially those served by publicly owned utilities--as the natural place to start. Often suffering economically and lacking broadband service from incumbents, many of these communities feel a growing urgency about their need for broadband. At the same time, some have local utilities able to contribute key resources needed to support a C-LAN--including access to low cost, patient capital.

Some optical vendors are also targeting smaller communities. Bernard Daines, CEO of World Wide Packets, for example, sees public power companies as "first movers" in the deployment of his company's gigabit Ethernet technology. Among the key reasons he cites are their longer financial time horizons, community-service orientation and experience with utility-related telecom.

The Grant County Public Utility District (GCPUD) in Washington state is one of WWPs first customers and one of the first public utilities to deploy an optical C-LAN. Under a Washington state law passed in March 2000 GCPUD is limited to providing wholesale service, a restriction that does not conflict with the C-LAN business model (see sidebar for a review of market and legal trends in other states).

According to Jonathan Moore, senior telecommunications engineer at Grant County Public Utility District (GCPUD), the utility is spending $3,500 per customer location today as it begins to deploy its optical gigabit network, dubbed "Zipp." He expects this cost to fall to $3,000 by yearend and $2,500-$2,600 by early 2002. Roughly 85% of the network will consist of aerial plant.

Though these prices would not be well received by private investors evaluating a largely residential broadband network--especially in today's tight capital markets--Moore says GCPUD, "as a government agency" is "less time constrained." He says the utility’s business plan includes a 15 year capital recovery period, with network electronics amortized over five years and the system expected to go cash positive in 6-7 years. This assumes that take rates rise from an initial 5% to 40%, which Moore suggests could prove conservative.

GCPUD's approach to the C-LAN wholesale model is an interesting one. It charges $4.00 per terabit of data flowing into the network with a minimum monthly access charge of $40, which covers up to 10 terabits of traffic (GCPUD does not charge customers for downstream traffic they receive from the network, though retail service providers may). There is also a one-time installation fee of $300. This fee structure applies to both consumers and businesses, including ISPs and other service providers.

Moore says initial ISP retail prices started in the $35-$40 range, but have since come down to $9-$25 as competition has taken hold. Northwest Telephone, based in Wenatchee, WA, has begun offering voice-over-IP service for $15 a month. Northwest Broadband is charging $40 a month for a package of roughly 70 video channels. Moore predicts this price will decline as additional video providers enter the market, a process already in the works.

GCPUD is hoping that the combination of its low wholesale fee structure for retail service providers coupled with competitive price cutting will allow county residents to receive a much higher level of service using Zipp for roughly the same price (including its minimum $40 per month access fee) they might otherwise pay for voice, video and data services using alternatives. It is also using the network for internal utility functions, including electronic meter reading and, later on, more sophisticated customer-controlled energy management functions.

Zipp's link to the outside world is NoaNet (Northwest Open Access Network), which operates a fiber backbone with 24 points of presence in the region. NoaNet is a non-profit corporation formed last year by 18 PUDs in the region, including GCPUD.

To make fiber transport more affordable in the state's low density areas, NoaNet charges distance-insensitive "postage stamp" rates. This improves the economics of linking remote communities to centralized "master headend" facilities by allowing the latter's equipment to be more economically shared among a larger but geographically dispersed user base. Other public utilities linked to NoaNet are watching Grant County closely and considering whether and how to pursue a similar model in their own communities.

Though their needs and resources are different than small towns, Bookey believes a strong argument can also be made for building C-LANs in the nation's major markets.

One potential spawning ground for the C-LAN model is low-income urban areas that have been a low priority for incumbent's broadband investments. The city of Chicago, for example, is in the process of soliciting proposals to build and operate CivicNet, an optical network reaching government offices, non-profit agencies--and eventually private businesses and homes-- in underserved parts of the city. With its $32 million in annual telecom expenditures, the city will be an anchor tenant on the network. It is also contributing resources to the project, including rights of way, conduit space and fiber.

City officials say the primary driver of the project is to promote economic development in the targeted areas. Rather than mandate a particular model for ownership, control and financing of the network, they are encouraging the development of new models by companies and consortia submitting proposals. The C-LAN model would seem to have some relevance here.

According to Bookey, C-LANs could help reduce the traffic problems that plague most major metro markets. "How many cars in rush hour traffic," he asks, "are actually delivering information…transporting employees' bodies to work so their employers can gain high-bandwidth access to their minds?"

Citing Seattle as an example, Bookey says the public is currently spending billions of dollars on ineffective solutions to transportation problems that could be better addressed by investments in C-LANs.

According to King County Councilman Rob McKenna, the Puget Sound Regional Council is considering spending $18 billion over 30 years to construct a light rail system that is projected to shift an additional 1.1% of commuter trips from cars to mass transit. Bookey believes a C-LAN could take a significantly bigger bite out of the city's traffic problems at a much lower cost. And, whereas the rail system would provide a mere 125 miles of connectivity, a C-LAN could interconnect every property in the Seattle area.

The table below outlines the impact a C-LAN might have if funded with some of the $18 billion Seattle is considering spending on its light rail system. Based on numbers suggested by Moore and Bookey, the table considers two different C-LAN capital cost scenarios: $3,500 and $2,300 per connection. The former might apply if Gigabit Ethernet electronics were used today, while the latter might apply to 100 Mbps Ethernet today or Gigabit Ethernet a year or two from now.

Based on these cost assumptions, a C-LAN serving Seattle could reach all of the metro area's 1.1 million homes and nearly 80,000 businesses for just 16-23% of the cost to construct the planned rail system.

The table also considers two different scenarios in terms of the shift from automobile to C-LAN as a mode of commuting. In the more conservative scenario the average household shifts just one monthly commute to the C-LAN. The second scenario assumes one weekly C-LAN commute per household. In both cases we assume an average of 1.5 commuters per household. Based on these assumptions the conservative scenario would yield a 3% incremental shift away from highway commuting, nearly three times as much as the proposed rail system. The more aggressive scenario would shift 13% of total commutes to the C-LAN.

Other assumptions used in the table include an average round trip commute of 20 miles, fuel efficiency and gasoline costs of 18 miles and $1.65 per gallon, respectively, and total vehicle operating costs of 31 cents per mile.

The C-LAN's capital costs are 50 times annual commute savings in the scenario with $3,500 per-connection costs and one monthly C-LAN commute per household. This drops to less than 12 times annual savings when we assume an average of one C-LAN commute per week and 7.6 times annual savings when we also cut C-LAN capital costs to $2,300 per connection.

On a per-household basis, the average monthly commute savings came to $6.20 in the scenarios based on one C-LAN commute per month and $26.85 in those based on one weekly C-LAN commute.

This analysis--simplistic as it may be--suggests that, for less cost, C-LANs might generate greater reductions in commuter traffic than the rail systems now being considered as major public spending priorities in cities around the country. It also suggests the possibility that telecommuting-related savings alone could justify public C-LAN investments. To the extent this is true, it means that the other benefits of a ubiquitous high-capacity, multi-purpose C-LAN could be realized at minimal public expense.