[overview and site map] |
CONCLUSION |
|
Supercomputer Policy Over the past years, the restrictions on supercomputer exportation have undergone many revisions and we believe is on its way to stabilizing itself into a thoughtful policy. Originally the policies were too restrictive and were in need of a "90s" update. In its zeal to promote free trade (and perhaps to honor campaign promises) the Clinton administration made sweeping changes to the supercomputer export policy by removing many restrictions. In retrospect, while the changes had the effect of promoting trade and bestowing economic benefits on American computer manufacturers, it is clear that many of the revisions went too far (see the Russia and China Incidents page). Since then, the policy has been amended and allows for the government to object to certain sales and to track statistics about buyers in suspected countries. These policies are very reasonable given that usually the sales are large (millions of dollars), and the countries that are being dealt with are possible threats to national security. Of course the policy will never be perfect, but it is definitely in the right ballpark. Some critics continue to argue that US policy is too relaxed. Any policy to allow the export of dual use items will face criticism. Supercomputers, although militarily useful, also have valuable non-military purposes such as weather modeling, safety and maintenance functions and can contribute to scientific discovery. It would be a shame if xenophobia would win out completely and deny these benefits to a large portion of the world's people. Furthermore, the rate of advancement in this area is truly astounding. To put some of this in perspective, consider that by 2004, the United States plans to have computers capable of performing more than one hundred trillion theoretical operations per second (100,000,000 MTOPS). The average Intel Pentium 200 MHz processor performs at a level of approximately 200 MTOPS. The computers that formed the basis of our disagreements between Russia and China had a performance that was measured in the range of 10,000 MTOPS; As of October 1998, the world's fastest computer was owned by the US Department of Energy (Livermore Labs), running almost 4 trillion operations per second, and in 1997, the Chinese unveiled a domestic supercomputer, the Yinhe-3, capable of performing 13,000 MTOPS. It is our contention that while the computers that we are exporting are capable of being used for destructive means, they will, one, be quickly outdated, and two, soon be domestically available in China itself. Exportation still situates Chinese equipment orders of magnitude less powerful than US government equipment, and as a side benefit, US companies will enjoy greater economic success. Encryption PolicyIt is our view that the current US encryption export policy has failed. As the ESI report concludes: Export controls on encryption technology should be dropped. The record shows that these controls have had no discernible impact on national security, but have demonstrably compromised America's economic security. Foreign encryption products are present in the free international market, their competitiveness is increasing at the expense of American companies, and their products are outside U.S. regulatory authority. In this light, export controls are indefensible. We tend to agree. Further, we believe that the governmental proposed key escrow and key recovery systems are also doomed to fail due to lack of foreign support. Given the choice of a US product that leaves open the potential for eavesdropping by the FBI or a secure foreign product, it is clear that most institutions, especially foreign corporations or governments, will tend to favor the secure product. The technological lead that US companies once enjoyed in this area is shrinking, and as a result, export controls should be eased immediately in order to prevent any further erosion. |
[overview and site map] |