For the Department of Defense, enabling computers to recognize objects within an image using sophisticated algorithms has long been a top priority.
Known as Image Understanding, these algorithms can create a description of the world to help build battlefield awareness and improve target recognition for military personnel. However, when it came to sharing discoveries and data relating to Image Understanding, the DOD simply wasnt getting the picture.
At the crux of the problem was the DODs multifaceted research community. Composed of a wide array of both internal and external players, this community includes U.S. Air Force, U.S. Army and U.S. Navy personnel; Sandia National Labs; Lockheed Martin; Raytheon; Booz Allen Hamilton; Cal Tech; and New York University.
For years, these factions have been conducting research, tests and evaluations relating to a variety of disciplines, including Image Understanding technology and sensors-driven electronic warfare. But despite these sophisticated, high-tech pursuits, the DODs research community was years away from being able to share information.
In response, under the auspices of the Office of the Secretary of Defense, the DOD launched the VDL (Virtual Distributed Laboratory). Created in 1996, the VDL is a Web-based repository that lets researchers talk about, share, store, search and retrieve research information—particularly data focused on Image Understanding.
Although run at Wright-Patterson Air Force Base in Ohio, the VDL is hosted and managed by General Dynamics, a Falls Church, Va., defense contractor, and is sponsored by the office of the Deputy Under Secretary of Defense for Science and Technology.
While the VDLs mandate was to encourage collaboration among disparate research groups, it didnt take long for obstacles to surface. For starters, the VDLs user base represents a diverse group of entities including DOD factions, academic institutions and corporate facilities.
All of these entities depend on a variety of platforms ranging from Windows to Linux, as well as a combination of open-source, government-developed and commercial applications. As a result, the VDL was hard-pressed to accommodate such a mixed bag of vendors, platforms and software packages.
For those who wished to pool their information with an outside entity, the only alternative was often regular mail—a lengthy process that included delivering research work to internal government departments, which would then forward the work on to external contractors.
"Unfortunately, in a lot of cases, [research groups] were FedExing CDs and tapes around. It was a very inefficient mechanism," said Eric Baenen, a program manager for the Advanced Information Systems strategic business unit of General Dynamics.
E-mail proved to be no more productive. Given the confidential nature of the DOD research communitys work, any online exchange of information demanded adherence to strict encryption and privacy policies.
Unfortunately, said Baenen, "the technology for sending encrypted e-mail back and forth wasnt quite as user-friendly as it needed to be, while sending unencrypted e-mails was not an acceptable solution." Whats more, many of the communitys external organizations had strict limitations on the size of e-mail attachments their servers could process, thereby preventing the receipt of large files.
Those entities that opted to avoid collaboration altogether also created their fair share of inefficiencies.
"The problem was, youd have lots of different people throughout the DOD collecting similar kinds of data," said Baenen. The result was not only a collection of redundant research findings but a squandering of government money: "If people would have talked and pooled their research, we could have had much more efficient use of our funding," Baenen said.
Because of General Dynamics role hosting and managing the VDL project, the company was also selected to serve as the DODs software integrator.
Finding an appropriate software solution, however, involved satisfying a long list of criteria. Among the DODs chief priorities was identifying software with strong authentication features, as well as point-to-point encryption capabilities, to meet DOD security requirements. Furthermore, because some external parties, such as academic institutions, are restricted from accessing certain projects and portions of research, the software would also have to feature fine-grain permission control and password protection technology.
But topping the DODs wish list was a Web-based document and file management system that would support open-source protocols.