No-nonsense networking: connecting dozens of schools in a high-speed network while on a budget is no small feat. The key to success? Finding partners who know what they're doing.

When Philip Brody arrived at Clark County (Nev.) School District in September of 1998, he was charged with upgrading the district's computer network so schools could become competitive in the newly dawned Internet age.

The only problem, Brody noticed, is that there was no network. Nothing to speak of, anyway. Says Brody, "Some schools had dialup access. Some had ISDN. Some had T1s. And some had no connectivity. And there was no plan on how to deal with any of it."

Problems like Brody's plague school districts today, which are inevitably coping with the challenges of networking their schools together so they can purchase connectivity as a group, share resources and offer up-to-date services to their students. And of course it all has to be done for a reasonable cost.

Districts have come up with a variety of solutions to the networking problem, and today there are a few common methods of dealing with the challenge. Brody's district is a perfect case in point.

As the fifth-largest school district in America, Clark County comprises 317 schools (and counting), nearly 300,000 students and a land area the size of New Jersey. Some schools are separated by up to 90 miles. And growth has been explosive: Beginning in the early '00s, Clark County was opening one new school every month, a pace that has remained at this breakneck rate for more than five years.

Plugging the Holes

As his first order of business, Brody decided he needed to at least get the schools a base level of connectivity. He knew that whatever the district decided to do for the long term, the necessary upgrades would take years to complete. Better to give everyone a base level of access immediately than force the unconnected schools to wait for the ultimate system to be installed.

For the short term, Brody connected all the schools directly to the Internet via a frame relay system from Sprint. The advantage of this system was that the company could completely handle the installation and management without intervention from Clark County. While the price was more than Brody would like to have paid, and this system didn't allow for any central oversight over internetworking among the various schools in the district, it did at least give the schools a reliable and reasonably fast connection to the Internet for basic Web browsing and e-mail.

But the Clark County plan was much larger than this. Enter Carlos A. Garcia, who was superintendent of district until late last year. (He is now vice president of urban markets for McGraw-Hill Education.) Garcia took the reins at Clark County in 2000 and spent his early months visiting classrooms and assuaging the technology situation. He was shocked to find many schools without computers. In schools with computers, several were still clinging to mainframe technology that was 20 or 30 years out of date--and employees didn't want to let that go.

Planning, Planning, Planning

Together with Brody, Garcia embarked on what might be the most ambitious network upgrade in the history of public K-12 schools. Their overarching goals: consider the long term and find ways to leverage the technology to save money. For example, Garcia wanted a telephone in every classroom and Brody wanted to ensure the schools had enough bandwidth. "Bandwidth was king," he explains. "You could never have enough. Adding bandwidth is expensive and time consuming, and we wanted to get ahead of the curve so we didn't have to play the bandwidth game every year or two."

In order to ensure that bandwidth level was high enough, Brody planned for a network that would support the school district for up to 10 years.

In a network project this large, the first step is to craft an RFP. Brody and Garcia enlisted the help of PriceWaterhouse and engineering firm SAIC to help them determine their exact technological needs and to generate a plan. Garcia is candid about their ignorance of cutting-edge technology, saying, "We knew we didn't have the expertise to write the RFR We didn't even have the expertise to know what questions to ask."

According to Garcia, the biggest hurdle is making sure you partner with good technology providers. "A lot of people and vendors say they can do a lot of things, but we knew that some people would over-promise and realized that no matter who we hired, there were going to be a lot of bumps in the road." From his experience, Garcia learned that it's vital to be patient and take the time to vet the vendors and develop partnerships with people you want to work with. That way, they'll be upfront and honest about the obstacles, training and the amount of work the project will take.

During completion of the RFP and a lengthy bidding process that Brody remembers not-too-fondly, the district rolled out a pilot project based on the technology it had outlined in the RFP. That pilot project connected seven schools, loaded them up with instruction applications and allowed for central monitoring at the district offices. It also made Garcia's goal a reality: Voiceover IP technology allowed the schools to route telephone calls over the Internet, taking the phone company out of the picture and saving millions of dollars per year. In the small ring of pilot schools, Brody was able to work the kinks out of the system and try out new technologies, like streaming video and the district's Virtual High School, which now serves more than 6,000 students.

With planning done and a bid accepted, Clark County installed its new network. The network is based on gigabit Ethernet and divides the member schools into 12 subnetworks. Each subnetwork is a ring of a dozen or so schools, served by two connections to the district administration headquarters, which provides a centralized connection to the Internet. (In fact, it provides multiple connections and offers at least two interconnections to each subnetwork ring. This provides complete redundancy for the network. If one line is cut or broken, the backup line can instantly pick up the slack.)

The network took 18 months to complete, which is a blink of an eye for a project of this magnitude, and cost $17 million. Clark County employees did most of the work themselves, which involved connecting 100,000 computers to the vast district network. Brody's team manages the network, with SAIC providing supplemental assistance when things get difficult. Most impressive of all are Brody's estimates of the cost savings: The district paid about $177 per computer per year for maintenance before the upgrade. Now that cost is down to $27 per PC.

The Consortium Connection

Clark County was lucky in having resources and staff to build its new network without much outside help, but most districts don't have that kind of expertise. Enter consortiums, a relatively new approach to buying technology services in bulk that lets someone else shoulder the burden of the work.

David Mabe is the deputy executive director of Texas' Region VIII Education Service Center. He relates a tale that has become all too familiar to American schools. In 1995, Region VIII was at the pinnacle of K-12 technology. The 48 schools in this northeast Texas district formed the Northeast Texas Regional Education Telecommunications Network, putting a T1 line in every school, managed centrally by Mabe's staff. It was a monument to progress ... for a while. As technology expanded and bandwidth needs increased, the T1 became overburdened and the schools couldn't meet even basic demands.

By 2003 it was time for an upgrade, and Mabe scrapped the old network entirely. He decided to partner with a company called Trillion, which convinced him to completely change the district's infrastructure.

Forty-six of the 48 schools opted to form a consortium to work with Trillion on the upgrade. Trillion specializes in forming groups of schools within a district (or spanning multiple districts) to help push down technology costs.

In the case of Texas's Region VIII, Trillion installed two massive rings of connectivity that circle the northeast part of the state. The network has 90Mbps of bandwidth, and individual schools have a minimum of 28Mbps, even tiny Marietta, which has just 50 students. Mabe is enthusiastic, saying, "There's no way some of the rural schools that we serve could have ever afforded this level of broadband Internet access." Trillion also provides hosted e-mail, filtering and firewall services, and the company also files the schools' E-rate applications.

Along with the impressive bandwidth it provides, that kind of cost leveling is the main draw for people like Mabe, who don't want to see smaller schools left behind in the rush to better technology. With a consortium, Mabe was able to allocate costs for the new network proportionally among the schools, so that no one institution felt a massive surge in expenses. "Member schools saw a 7 to 10 percent increase in costs, but got about a 2,000 percent increase in bandwidth," he says. Those costs are fixed for five years, too, so there will be no rate increases until 2010.

Mabe says this solution is probably the only way this network ever could have been built. The reason is size: Region VII covers 6,500+ square miles and measures 170 miles across. As with Clark County, telecommunications services are spotty and broadband services are nonexistent in some of the more rural locations. (A combination of fiber and point-to-point wireless connections are used to connect the schools.)

Administrators in other regions report similar results, with even larger areas. Floyd E. Beard, executive director of the Colorado East Central Board of Cooperative Educational Services, manages a collection of 21 districts that span from Denver to the Kansas border. He started a consortium to connect all the districts into one meganetwork that is now under construction for launch in the fall. He says the cost, which is equivalent to "one teacher's salary per district," is the best part.

The state of Utah may represent the largest consortium of all: The state connects all 40 school districts and provides centralized services like security and content filtering. Most of Utah's work, however, is done independently and without third-party help.

Mabe sums up his feelings on the consortium system succinctly, saying, "I would highly recommend the consortium because when we all pool our resources and share in costs and administration, everyone's a winner."

Christopher Null is a freelance writer based in San Francisco.

Tips for Successful Networking

Planning and installing a network is an exhausting and complicated process. Remember these lessons when it's time to upgrade your district.

Hire people who know what they're doing. You absolutely need someone on staff that's an expert with technology, no matter how much you plan on outsourcing.

Consider a stopgap system, installing a large network takes time. ISPs and telcos can install a simple, rudimentary network to provide services while the real network is being built.

Get help understanding what you need. Just understanding what you want your network to do and what equipment that requires can be the hardest part of the installation. Spend twice as much time on generating the specifications for the network than you think you need.

A pilot project can be a big help. If you're installing services at a large number of sites (30 or more), consider a pilot network of just a few schools where you can test how well your network will operate.

Create a meticulously detailed network plan. You won't be able to personally inspect every piece of equipment every day, unless you like the idea of spending thousands of hours on the road.

Plan for the long term. The Internet is a rapidly changing beast. Don't plan for what you'll need next year. Plan for eight or 10 years out, especially regarding any new schools you'll be opening.

Get help implementing that plan. Doing the legwork of running cable, installing routers and connecting computers is tough, but it's a well-understood business. Literally thousands of service providers are available to help you with this work.

Don't forget about maintenance. Clark County invested in extended warranties up front to minimize the risk of big repair payouts down the line.

Don't forget training and education. Get technology staff interested in the upgrade process from the start through training. Continue training regularly, especially as new services are introduced.