THE DEFENSE DEPARTMENT'S ACQUISITION process isn't keeping pace with the environment in which our forward-deployed forces must do battle.

Making appropriate technology available to those forces challenges the department for two distinctly different reasons: one, the availability of the technology itself, and two, the process of acquiring it and getting it to the field.

Compared with just a decade ago, the technology commercially available is vastly superior. The continual progression of Moore's Law-that computing power doubles roughly every 12 to 18 months-means entrepreneurs now can offer varied solutions for gathering information, distributing it around the world and processing it to make sense of the inputs. Even so, the right combination of tools to fit the job of our deployed forces is often relatively impossible to find commercially.

DOD's acquisition system meanwhile has evolved as a most bureaucratic process: It's cumbersome and slow, and any organization that might be remotely affected by decisions gets the chance to express its point of view. But instead of a decision process that lets everybody have their say before someone responsible decides what to do, Defense substitutes a set of meetings that are nothing more than forums to ensure completeness of the inputs and to hope for agreement by consensus. Layers of such meetings culminate in a Defense Acquisition Board where consensus once again supplants crisp decision-making. Finally, during the acquisition process, everybody is still empowered because nobody told anybody else that their ideas were not important enough to make the cut.

Amazingly, this process works-as long as the department can control the pace of technology advance, such as with nuclear weapons, stealth applications and other uniquely military technologies. It absolutely does not work when technologies are commercially available. Adaptation is usually necessary, but meeting the requirements of a bunch of teams seeking consensus is impossible.

The Global Information Grid-Bandwidth Expansion program is an example of a successful new acquisition path. By using commercial standards to let multiple bidders compete for each subsystem and choosing the correct partitioning of subsystems, the Defense Information Systems Agency has deployed a worldwide, high-bandwidth, fiber-optic network at almost 100 sites within three years-and under budget.

One reason it worked: no large groups trying to come to joint agreement. The program office, the Joint Chiefs of Staff and the Office of the Secretary of Defense discussed the locations and the levels of bandwidth at each location, and then the assistant secretary of Defense for networks and information integration (which was me at the time) made decisions if there was a disagreement.

This program should be seen as a model for other DOD programs. And, DISA now knows how to run such programs and could do so again.

But I must add a footnote about how process can still strangle a good program. To cover GIG-BE's operation and maintenance costs, I favored yearly appropriations that would encourage experimentation into how this new network-centric capability could be used by making the system seem free to its users. OSD, however, has opted to charge organizations by usage, which can also work.

But there's a troubling rub. Because some users voiced concerns that their circuit-based applications might falter when cut over to the IP network, the GIG-BE program permits the use of special software and hardware to emulate circuits. For IP traffic, a user can just plug in to the edge of the network.

In choosing the reimbursement rate for users of GIG-BE, OSD set the rates for users demanding circuits at a lower cost than the fee for IP users. So DOD is subsidizing users who do not move to GIG's net-centric standards and penalizing those who use the new technology. I am amazed that the power of the system's secondary users is so great that they can command discounts out of the hide of primary users. It's a major error that needs an immediate fix.

John P. Stenbit was Defense Department CIO from August 2001 to March 2004, when he retired after a 30-year IT career in government and industry. He consults and serves on corporate and other advisory boards. E-mail at johnstenbit@cox.net.

purchase reprint

link to this page