|
[intro] [export control policy] [current situation] [looking ahead] [conclusion] |
CURRENT RISKS |
Case studies: Russia, China, and Supercomputers |
|
China Concerns In 1993, US approval to sell an $8 million Cray supercomputer to China for weater analysis gained publicity. Yet after supercomputer export control was relaxed in 1995 for computers that can calculate between 2,000 and 7,000 MTOPS, China went on a shopping spree, buying 46 supercomputers over the next two years. Because discretion is left to the American company as to whether the products will be used for military purposes, the government does not know where the computers are being used. The buyers must be civilian for the sale to not need federal approval, but it is possible for the computers to be diverted after the sale. American intelligence officials fear that they could be used by the Chinese to design and test new nuclear warheads. In 1997, Washington officials did find out that one Sun Microsystems supercomputer that had been sold to China had been diverted from civilian to military use. The U.S. was able to pressure China to return the computer to Sun. There was no evidence of similar diversions happening to other computers. In 1998, the Department of Commerce required that American companies exporting computers performing 2,000 to 7,000 MTOPS to certain countries, including China, notify the Department, which can raise an objection to the sale. (described here [link]). These concerns are part of the debate about US technology export controls. Do the supercomputer sales constitute a possible threat to national security? The Russian and IBM Incident In accordance with US export controls over supercomputers (discussed here [link]), sales of computers to foreign military installations must have federal approval. In July, 1998, and IBM subsidiary pleasded guilty to 17 felony violations of these export control laws. IBM East Europe/Asia Ltd. participated in selling 17 supercomputers without federal approval to Arzamas-16, a nuclear weapons laboratory in Russia. The deal took place in late 1996 and early 1997, and used middle-man companies to handle shipping from the German IBM plant through Amsterdam and back across Europe into Russia. American IBM executives had no idea the deal took place. Russia bought the computers in a misunderstanding that extended from the nuclear test-ban treaty they signed in early 1996. Russian officials said they thought the computers would be approved after the deal as alternatives to actual nuclear testing. Requiring federal approval allows the US government to know where dual-use technologies are being sold to. After finally getting Russia to sign a test-ban treaty, is the US being backhanded in being so restrictive about giving Russia the means to test weapons on a computer? The US requested the return of the supercomputers, but Russia refused. Does this indicate that the US would not have approved the sale had the IBM subsidiary requested it? What kind of effect does this policy have on American economics versus American security? The debate on supercomputer and other high-tech export control continues.
|
|
[intro] [export control policy] [current situation] [looking ahead] [conclusion] |