On the other hand, China is a developing military power, with real ambitions complicated by economic change. Having embraced a kind of capitalism, the Chinese Communist Party can no longer claim that it is the proper guide to the Chinese future; its ideology is largely defunct. However, it badly wants to retain power. It now seems to see nationalism, i.e., the recovery of past Chinese glory, as its main argument for legitimacy. That argument probably spells trouble in the South China Sea. Given Chinese nationalist ambitions, the penetration of the Chinese military into what in any other country might be the civilian economy can have real consequences. It is now probably impossible to say whether any Chinese high-technology enterprise is or is not subject to direct military control.
In 1996, the Chinese approached Hughes, the main U.S. maker of communications satellites, to help them develop a mobile telephone system. At about the same time, their military communications satellites were running out of operating time, and an attempt to launch a German satellite failed. The Chinese military began to buy satellite telephone time on commercial satellites. Hughes found itself hiring the son of a senior Chinese general involved in the Chinese space program. To make matters even more interesting, the general began to advocate improved intelligence collection, which by implication meant collection by listening satellites.
President Clinton approved the satellite export, waiving a ban on military and satellite exports imposed after the 1989 Tiananmen Square massacre. His action is hardly unprecedented, but it is now being raised in the context of heavy (and illegal) Chinese contributions to the President's 1996 election campaign. Should he have known better? Did he willingly hand over critical technology? How important was it that at about the same time he streamlined the process of gaining export approval, placing it under the Commerce Department, which is oriented toward exports? Or is this a case of truly commercial technology advancing to the point where it looks better than much existing military equipment?
The commercial satellites the Chinese were already using, like other commercial satellites, offered highly reliable long-range service, but only on a limited point-to-point basis, with a limited number of circuits. Transmitters and receivers needed fairly large dishes—with limited mobility, particularly for ground forces. By way of contrast, a mobile telephone system employs units small enough to be hand-held. To be useful, the system must be able to switch among thousands of conversations. The satellite communication system can be used tactically, but a mobile telephone system would seem to offer far more flexibility than older and more conventional types.
Mobile telephones produce very weak signals. To receive them for retransmission, a satellite needs a high-gain antenna. The Hughes project has recently attracted additional attention because the antenna it uses is apparently quite similar to those used by the U.S. government to intercept foreign communications. There can be no question but that any such antenna-receiver combination can pick up not only Chinese mobile telephones, but a great deal more—such as low-power military communication. Moreover, it is difficult to imagine how the Hughes satellite (or any comparable device) can somehow be limited to the telephones it is intended to support. That depends largely on the ground electronics with which the satellite communicates.
To make matters still more interesting, the Chinese claim that their mobile phone network may grow quickly to about 200,000 subscribers. So small a number suggests that the subscribers will be Chinese government users, perhaps mainly military. A system with this sort of capacity would certainly suffice to connect the Chinese general staff with its subordinate commands. Compared to conventional radio, mobile phones would be much more difficult for a hostile power to intercept. Interception could be made even more difficult by attaching the telephones to simple directional antennas. Thus the mobile phone system certainly has potential military value.
The worldwide demand for mobile systems is very strong, however. It is difficult to imagine how military subscribers can be excluded or, for that matter, how coded transmissions can be controlled. In the Chinese case, the issue is that the Chinese government may use its ownership of the mobile system to eavesdrop on its own citizens. That seems inescapable. The more interesting question is whether a mobile system, effective well beyond Chinese borders, gives the Chinese military a major advantage in projecting power.
Behind the interest in RMA lies an assumption that the United States can maintain a monopoly on advanced forms of command and control. To some extent that is probably true; we spend much more than anyone else on intelligence collection. The Hughes deal, however, would seem to imply that we cannot control communication itself. Moreover, U.S. attempts to limit the power of commercial encryption systems have generally failed. Thus, even if we can intercept the weak signals which foreign users of mobile phone systems use, they may well evade our attempts to decode them. Yet advocates of encryption can point out reasonably that commercial users demand security—from, among other things, their competitors.
We obviously depend heavily on signal interception and location. The impact of wide-spread mobile telephone systems would seem to be to make interception far more difficult. The new generation of encryption systems adds to the difficulties. It would be foolish to imagine that U.S. capability is almost gone. It would be equally foolish, however, to base our strategy on the assumption that we can continue to track forces as we have in the past. After all, how can we tell whether a particular mobile telephone is in the hands of a general or in the hands of a salesman traveling through, say, China?
The Revolution in Military Affairs matters enormously to us. It is a way of maintaining U.S. global power despite drastically reduced resources, both in money and in human effort. The idea behind the RMA (or network-centric warfare, or Joint Vision 2010, which all are much the same thing) is that U.S. technological leverage in information collection and dissemination should translate into military superiority. Yet the Hughes deal is a demonstration that the most advanced U.S. technology is often available for sale, because it is civilian rather than military technology. That does not mean that the United States cannot prevail. However, it does mean that we are unlikely to enjoy for long the almost effortless superiority that advocates of information warfare and the RMA seem to expect.
The Hughes deal, then, is symbolic of the shift we see everywhere around us: the military procurement system cannot keep up with the avalanche of new civilian devices. The military is now too small a market segment to dominate any sort of electronic development. Companies literally cannot afford to keep making military-rated components (in some cases, they claim that their standard civilian lines are up to military standards).
The outcome of this sea change in the electronic industry has forced the military to use commercial off-the-shelf components. Such components are certainly less expensive than their older military equivalents, but we are beginning to see that they impose their own very significant costs.
For some time it has been a given in the computer industry that the speed of cutting-edge computing chips doubles every 18 months. Unfortunately, the average time to move a military project from design to production is about a decade—and the time span is increasing. How, then, do we design new weapons? The approach in the New Nuclear Attack Submarine, probably the best we can do, is to provide open-architecture electronic enclosures and the most capacious possible data bus. If the submarine is to last 18 years, the rule of thumb dictates that the bus should be at least 4,000 times more capable than would seem necessary now.
The really difficult problem is software. It probably takes about as long to write and, more important, to test the submarine's software as to design and build her hull. Experience suggests that software performance can be intimately connected to obscure details of the computer chips on which the software runs. What happens when the chips for which the software was designed are longer made? In the past, the assumption has generally been that software could be made upwardly compatible. True, the original chip might go out of production, but the maker would have a strong incentive to develop more powerful ones that could run the same software.
However, each chip technology has its limits. What the market seems to be saying is that chip makers are finding it more profitable to design entirely new chips for higher speeds, rather than to maintain family relationships so that software can continue to run. The market is still expanding so quickly that many buyers do not care whether their new machines can run software compatible with their older ones. Many businesses dispose of their computers wholesale every two years or so, simply because they cannot afford to be handicapped by slow chips.
The military procurement system, however, treats computers as major capital investments rather than as expendables. Its software must be much more carefully tested than civilian software, since the consequences of failure tend to be devastating. The implication is that we must often accept that systems will use the chips available (or, perhaps, in the late design stage) at the time when weapon systems are developed. We can get around withdrawals from production by stockpiling, but in that case we will find ourselves with weapons powered by chips only one-thousandth as powerful as those available on the open civilian market at the moment those weapons come into service. How do we design equipment so that the change in computer technology can somehow be isolated from other system elements?
Today, the Chinese show little ability to translate their wishes into military power. Most of their equipment is still antiquated, and they seem unable to integrate their lively new computer software industry into their state-owned military industry. It would be foolhardy to imagine that the problem will never be solved. When it is, Chinese designers undoubtedly will insert current computer chips into their new weaponry. It would be most unfortunate if we faced those weapons with equipment built around chips a thousand or so times slower, simply because ours were designed a decade or so earlier.