Subscribe to the Free Print Edition now!
Defense Systems Wednesday, August 20, 2008

Current Issue eSeminars Jobs FAQ
1105 Media [purity]
quickfind
purchase
reprint
link to
this page
categories
C4ISR
Network-Centric Warfare
Training and Simulation
Security and Intelligence
online resources
White Papers
RSS Feed
Military Links
1105 Media, Inc.
» Government Computer News
» Government Leader
» Washington Technology
» FOSE

home > news > article

|  News  |

Delivering Information Superiority in the Global Information Grid: Why Smart Push Dominates Smart Pull

The Defense Department will spend billions on a new services-oriented architecture called the Global Information Grid and associated Network-Centric Enterprise Services program. The primary driver for this new initiative is a belief that the GIG and NCES will produce information superiority for military operators, a state where friendly forces have more valuable information than the opposition and use that information to produce better decisions that result in superior outcomes.

These ideas have been spelled out over the last decade in many educational and policy publications. Implicit in many of these papers is the tenet that information needs to be metadata tagged to give operators the ability to search and pull information when and where needed.

The so-called Smart Pull theory is that information superiority results from a system that enables every user to pull relevant information. This theory motivates DOD efforts to develop XML metadata schemas for existing data sources, to “post before processing,” making raw intelligence data immediately available to operators, and to emphasize information search and access.

Although Smart Pull has intuitive appeal, the intuition is flawed and can lead to serious architectural errors and lost opportunities. The simplest and most direct alternative is the Smart Push theory, which assumes that many operators can and should delegate the work of monitoring for newsworthy information to a system with enough smarts to determine when conditions of interest occur.

While this might seem like semantic quibbling, it’s much more than that. The two theories, in practice, differ in efficiency by more than five orders of magnitude. Where the pull theory might require an operator to process 10 million bits of relevant information, the Smart Push theory reduces that to fewer than 100 bits. Put another way, Smart Push is a hundred thousand times better than Smart Pull. Smart Push reduces the number of bits that need to be communicated, considered, and acted upon by 99.99 percent. This, of course, reserves scarce resources so operators can focus on important information and respond quickly. Specifically, by filtering out all the unimportant bits, we assure that the important bits get to the people who need them and that those people aren’t glutted with vast numbers of extraneous data.

The Naval Postgraduate School has been working with the World-Wide Consortium for the Grid (W2COG) to implement the technology and deliver the benefits that this superior theory entails. NPS and the W2COG call the concept VIRT, which stands for Valuable Information at the Right Time. Recalling Shannon’s original concept of a bit as the amount of information required to reduce a receiver’s uncertainty by 50 percent, we pursue information superiority by maximizing the flow of valuable information delivered to each operator. In many operations, the operators first consider alternative courses of action in light of expected conditions and then settle on an actual plan. The chosen plan is deemed best when the planners foresee that it achieves the best outcomes under the expected conditions. From that point forward, operators need to know when some new information indicates a significant change in expected conditions, so that the plan’s optimality becomes suspect. In short, an operator engaged in carrying out a mission considers newsworthy just those changes that can undercut the previous expectation of success.

VIRT focuses on assuring that each operator receives high-value bits by giving those bits priority and filtering out low-value bits. What bits do operators value most? When operators are conducting planned missions, the outcomes of those missions are jeopardized when unanticipated events occur that contradict the planners’ assumptions. For example, if we want to rescue hostages using helicopters that require refueling in the desert, a sudden, unexpected sandstorm makes the planned mission untenable. To achieve the goals of information superiority, we need our systems to understand that a sudden sandstorm is an event that adversely impacts any mission that assumes helicopters will be refueled in that vicinity. VIRT services deliver this information quickly by understanding how operator missions depend on assumptions and by continually seeking information that indicates adverse events might be occuring. Each operator values just those few bits indicating that an assumed condition has gone from “acceptable as expected and planned for” to “unacceptable, not expected, and undercutting the plan for the current mission.”

The VIRT working group in the W2COG includes government employees, academics, and a number of leading IT contractors including Oracle, Intel, Lockheed Martin, and Teknowledge. The group is creating an open architecture for VIRT as well as some off-the-shelf implementations.

The key components of VIRT include:

* Tools for operators to describe events that they need to be alerted about, which we call “conditions of interest” * Vocabularies and semantic schemas that relate available information sources to such conditions of interest * Geospatial and time-indexed databases for maintaining current and forecast values of the relevant variables * Continuous monitoring for changing values of conditions of interest, including projected future states * Alerting methods for communicating significant news quickly and concisely.

While there are many situations where pulling information can be useful, time-stressed operators can’t achieve information superiority that way. They need systems that can understand how their operations depend on expected conditions and can exploit plentiful computational power to cull valuable bits from vast amounts of relevant data. VIRT, enabling Smart Push, will provide the needed approach to do that.

DOD will need to shift the underlying theory driving GIG and NCES to deliver information superiority to operators at the edge.

Rick Hayes-Roth is a professor in the Information Sciences Department at the Naval Postgraduate School. He can be reached at hayes-roth@nps.edu.


purchase
reprint
link to
this page
advertisement
advertisement
advertisement
ADVERTISE CONTACT US CUSTOMER HELP EDITORIAL INFO SITE MAP