The idea of smart pull is that in a net-centric environment the user can be asynchronous in both time and space, choosing when to pull down the information he needs from anywhere. When describing the advantages of the net-centric environment over today’s primarily broadcast environment, first on the list is this exact capability. It’s a major improvement over a system where information generators decide when to publish over a broadcasting system. Time is the difference as, to a lesser degree, is the specialization of use assumed in how information is sent.

In the bad old days before the broadcast system that has been deployed over the past three decades, an information generator determined not only when to send information but also to whom, because the systems in those days were basically telephones. The dissemination of information was therefore synchronous in both time and space. The receiver had to be where he was expected to be, and he had to be there when the information was sent to answer the phone.

The current broadcast approach lets a user be asynchronous in space but keeps him synchronous in time: He can be anywhere and receive the information from satellites beaming it across the globe, but he has to be listening when it is transmitted or he will miss it.

For the next-generation net-centric system to be effective, several conditions must be satisfied, starting with the availability of a common communications hub of virtually unlimited bandwidth connecting major information sources to users. At the Defense Department, the Global Information Grid-Bandwidth Expansion program will be that hub.

Plus, we require data sources on the network that are readily discoverable and open to all authorized users. This in turn requires applications that manipulate the data to derive answers. The apps must also be discoverable and open to all authorized users. Both the data and apps exist in various forms today, but DOD needs to adapt them so they are discoverable using metadata tags. Defense must create these tags and establish standard interfaces to allow general use across the department.

Neither of these tasks is extremely difficult, but DOD must perform them for virtually all its legacy information systems. It’s a problem of scale. Even if it completes the task successfully, DOD must also add something new to the network: value-added services. These business process controls will marry the apps to the data in a sequence so a user can set up an automated rule for retrieving information easily and in a format that’s compatible with his or her situation.

Tying this back to the reader’s comment, for example, would mean that a user could define a geographical box and then discover all of the data that he could find that fits in the box and then assemble the data himself, or he could call on a value-added service like Google Earth to do it. Development of such services will not be easy to stimulate because they are not normal business, per se. But there are promising signs of a natural proclivity for vendors to create them for some data types.

Achieving the benefits of the smart-pull approach will require that DOD deploy more enterprisewide services and use some creative management to encourage the hard work of tagging data and applications for discovery and—most of all—for creating the value-added services.

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.

purchase reprint

link to this page