Information warfare is a relatively new rubric, which is receiving increasing attention within the United States from both the government and the general population. Recent studies and Congressional hearings have discussed the vulnerability of the U.S. civil infrastructure to information sabotage, perpetrated by both state and non-state actors. Most recently, President Clinton established the President's Commission on Critical Infrastructure Protection to identify vulnerabilities in the nation's overall infrastructure and to recommend policy actions to reduce them. One of the areas that the Commission will investigate is the nation's information infrastructure. For instance, the armed services foresee new uses for digital systems to enhance military capabilities, but they also recognize the growing U.S. vulnerability that might be exploited with the techniques of information warfare.

The existence of softer and perhaps more critical homeland targets is creating interest in information warfare at a strategic level. That interest has two very different themes: new weapons the United States might use against an adversary and, in the hands of others, new threats to U.S. civil information-system-dependent infrastructure. The latter, the defensive concern, is currently receiving the larger measure of public attention.

This paper reports the results of a study of the special export control regime for high-performance computers. The history and purpose of this export control regime are reviewed, and a framework for analysis is established, which can be used to test the basic premises on which the control regime rests and to suggest viable control thresholds. The fact that the export of certain computer systems cannot be effectively controlled is established, and the limits of controllability are defined. U.S. government applications for high-performance computers are reviewed with respect to the requirement for and criticality of such computing for national security. Finally, judgments are made as to the levels of control that are possible, and the desirability and feasibility of maintaining Such controls. Near- and intermediate-term problems that may erode the liability of the basic premises underlying high-performance computer export controls are identified.