February 9, 2004

by Jenny Carless, News@Cisco

A nation's capital is a good place to learn about the latest in legislative affairs. But some might not think to look there for the latest in networking.

In the case of Washington, D.C., that would be a mistake: the Office of the Chief Technology Officer (OCTO) of the Government of the District of Columbia is implementing a forward-looking technology initiative that municipalities around the world can learn from. OCTO has selected optical, IP and Ethernet networking equipment from Cisco Systems to build a high-speed fiberoptic network that will provide city-wide voice and data communications services.

By owning its own metropolitan network, OCTO is getting more power and reliability - for a much lower price - than it could otherwise.

Still in the development phases on an OC-48 SONET backbone, OCTO expects to unveil its first DC-Net service within the first half of 2004. DC-Net will eventually link more than 300 sites, allowing OCTO to consolidate data and voice communications traffic from all of the city's agencies and provide services such as real-time interactive video for public schools and an updated emergency communications system for the police and fire departments.

News@Cisco asked Peter Roy, deputy chief technology officer for OCTO, to talk about the DC-Net service and what other municipalities can learn from his organization's experiences.

What business and/or technical factors led OCTO to build DC-Net?

Peter Roy: Initially, we thought we could save a few dollars by putting in a few point-to-point data circuits, instead of paying the ILEC for leased lines. But after we negotiated our franchise agreement and completed some engineering, we realized that a modest incremental investment would allow us to connect virtually all of our locations on our own private network.

We also realized that moving from Centrex voice to PBXs on a private network would save a lot more money than the data circuits ever would. We have about 40,000 Centrex stations, and this large number simplifies our business case.

Later we realized that we could use the reliability features of a self-healing SONET network to provide the high availability we need for backhauling our public safety wireless voice and data traffic and for moving E911 traffic between our own selective routers and the District's public safety answering points.

As time goes on, we've found more and more applications that make owning our own metropolitan network make even more technical and financial sense.

Basically, we will have much more power and reliability for a significantly lower price. On a straight payback basis, the project will pay for itself in five to seven years, depending on how many government agencies use the system.

What were the challenges of taking on an IT project of this scope?

Peter Roy: Budgets in state and local government are tight. It's hard to convince the decision makers that this sort of project is worth investing in, when there are so many programs competing for declining dollars. When funds are scarce, it's tempting for government decision makers to pay more in the long term for lower-performance communications systems, rather than divert the limited funds to invest in a high-performance system with an excellent payback. But in the District, we're fortunate to have leadership that understood the need to invest in a high-performance, high-return private network, particularly in support of our critical public safety needs.

The start of the project was complicated by the fact that engineering started in 2000, at the peak of the tech boom, and talent was both hard to find and expensive.

There is no clear pathway or established state of the art for contracting for services on a large private government network. We had several false starts, because the market was not always responsive to some of our initial attempts to package work in ways that we thought would be advantageous. It's important to take the time to research different alternatives for the contracts needed to manage the network and to perform the market research necessary to see whether the marketplace will respond positively to your contracting plan.

What will your next steps be?

Peter Roy: In the next few months, we'll be switching over several thousand voice and data circuits in our largest buildings. The buildings with the highest numbers and concentrations of personnel produce the highest savings, so we start with them. By October 1, 2004, we'll have our first 80 buildings on the system.

What can other municipalities learn from OCTO's experiences with DC-Net?

Peter Roy: Develop a solid business case and make the financial aspect of the project sell it for you. Once people realize that you will get more (and better) stuff for less money, the ride is a lot easier. Recognize that this stuff is arcane and that most people will need help to understand what it's all about - and why they should become a cheerleader for the idea of a private government network when many people think that such work is inconsistent with government's typical core competencies.

Invest the time and effort to explain - and "sell," if necessary - what the project is about. Make people inside and outside your organization understand that you're not going into the telephone business - instead, you're building a high-performance infrastructure that will allow the government to compete better for services. For us, the work of switching over and operating the network will be performed by contractors who had to compete for the privilege. We have a few employees who will perform inherently governmental functions.

Also, do the research to show that this is not a harebrained idea that you're the first to think of. There are many successful private government networks in place and many more in the works.

We stayed away from becoming a carrier, such as a CLEC. This has been very beneficial. Remain a private government network provider as long as you can, because it keeps you from being regulated like a CLEC and avoids many of the troublesome legal issues that have plagued CLECs in the last few years.

Allow enough time, and don't promise a specific switchover date unless you have all your ducks in a row. Everything takes much longer than you think it will, especially the myriad steps leading up to switchover.

It's important to be aware of the right-of-way issues that will crop up in an urban location, also.

Realize that the best business model is usually to get contractors to compete for the right to operate your network. That is the path we have selected, and we are very satisfied with the results, especially from a price standpoint.

If you're going to run mission-critical systems on your network, make a proven track record an important criterion for the equipment you select.

Realize that running a multi-service network will make the organization much closer to a private business, requiring non-traditional government functions, such as a billing system, outreach and customer service, a help desk and quick-response teams.

The usual rules of managing a technology project apply: