Finding the Way: Taking Web Infrastructure to the Network's Edge

January 2010
Topics: Command and Control, Netcentricity, Systems Modernization, Technological Innovations
When the DoD needed to evolve the net-centric capabilities of the Global Information Grid, MITRE engineers Steve Farish and Stan Ames found a solution in a commercial company's innovative approach to Web infrastructure.
man on teleconference

Since coming into existence in the late 1990s, the Global Information Grid (GIG) has continued to evolve as a network that can support the Department of Defense in its goal to achieve network-centric operations—or net-centricity. GIG Content Delivery Services, for example, enable the delivery of content and applications through standards-based Web technology. File sharing, video conferencing, and other enterprise services now routinely arrive at the edge of the network in sub-second speeds.

It wasn't always this way, however. For example, video conferences in Kabul, Afghanistan, didn't happen overnight. The origins date back to the summer of 2002, when mission applications were becoming more widespread and email was no longer the primary use of the Internet. It was clear by then that business system modernization meant moving to the Web.

MITRE's Stan Ames and Steve Farish were both working on a business system modernization activity for the Air Force's Global Combat Support System (GCSS). To see if their Web-based program would work, Ames and Farish were testing software applications—but they soon ran into trouble. "As we started to roll out apps, we discovered that we didn't have a network that could stand the traffic," says Farish, now associate executive director of MITRE's Warfighter Integration Division.

Not only had they discovered a problem that couldn't be solved within budget, it was even bigger than they could get their arms around. "There's no bigger orphan in the world than infrastructure," Farish says. "I dont care if it's IT or bridges. There's essentially no champion for it."

Considering a Commercial Solution

As they considered various solutions, Ames and Farish didn't rule out commercial technology. "If we can identify companies that have a unique idea that we can apply to our Air Force systems, then we all win," says Ames, a senior principal engineer in MITRE's Enterprise Systems Integration department.

As they considered possible solutions, the two engineers decided to look into Akamai Technologies, a company in Cambridge, Mass., that had begun to acquire a reputation for its content-delivery network. The company already had an impressive list of high-profile customers, including the White House, the FBI, and others. Ames and Farish arranged a visit to their headquarters.

"Lots of companies are doing now what Akamai was doing then, but at the time, they were the primary company pushing that concept," Ames says. It was up to him and Farish to figure out how best to incorporate the technology within the Global Combat Support System program.

"Not Invented Here"

Early on, Farish recognized that security was going to be a challenge. They were, after all, proposing to outsource protected bandwidth. "Stan has a great deal of expertise and some level of gravitas on security," Farish says. "Thanks to him, we won the security battle."

The real challenge, however, would be even more daunting—the "not-invented-here" mindset. "We had a harder time convincing the DoD that we should just buy it instead of building it ourselves," he adds.

There wasn't any doubt that the technology was going to be used. The real question for Ames and Farish was: When would people realize that this was the best approach? In order to succeed, the two engineers recognized they would need to strategically engage and influence the stakeholders and eventually find a champion.

"A lot of people don't recognize a good idea the first time they see it," Farish says. "There's no malfeasance there; it's just that they're locked into their current way of doing things. It's human nature."

Road Trips Pay Off

Akamai is located less than an hour's drive from the Electronic Systems Center at Hanscom Air Force Base, and a succession of general officers up to the two- and three-star level soon found themselves taking the trip into Cambridge. On each visit, more influential stakeholders became convinced.

"It helps if you get to the senior leadership early enough that they can see the value," Farish says. Among those who saw the wisdom of this commercial capability were the commander of the Electronic Systems Center at Hanscom, the vice commander, and the Air Force chief information officer. "We were lucky because we did have some senior people in the Air Force who understood how important it was going to be," says Farish.

One of the biggest proponents was Joe Besselman, the Air Force GCSS program manager, who conducted several successful pilot programs. Following these pilots, the Defense Information Systems Agency adopted the concept and put it on their CipherNet. At that point it became a supported capability within the Air Force community.

Off the Charts

Altogether, it took three or four years for the pieces to finally align. "It took awhile, because we planned to use this capability for systems that were also in the process of being created," Farish says. Ames agrees. "The concepts of the technology were only just beginning," he says.

Many people were ultimately involved in the implementation, but it was Ames and Farish who initially identified Akamai as the ideal solution. While both engineers downplay their role, most agree that their innovative and nontraditional approach to industry outreach, as well as their skillful strategy for influencing stakeholders, had impact well beyond their own work program. Once their Air Force program had adopted Akamai's services, this solution went on to have a significant role across the DoD. PowerPoint charts at the Pentagon now attest to the massive use of the commercial solution first proposed by the MITRE team.

"Applying the best technology to a problem is just part of what MITRE does," says Ames. "It was really, in the end, a no-brainer situation—it cost a lot less to hire another company to do something than to try and do it ourselves." But Joe Besselman begs to differ, believing that Ames is far too modest in remembering his and Farish's role. He summed up their work as "really an off-the-charts contribution."

—by David Savold


Publication Search