those of industry. Because the DoD is today a relatively small customer for information technology, industrial R&D programs are considerably larger than those affordable by the Air Force. Clearly, the commercial sector provides immense incentive for going after scientific and technological advances in this field. However, some technologies that are unique to the military or that are not yet commercially viable require military investment:
. . . For example, many sensor applications are unique to government requirements and hence are funded solely by the government. Similarly, there are additional technologies that are essential for government missions but which may have or develop commercial application as well; however, the cost of their development is usually so high that industry cannot make a business case for maturing them commercially. Examples include the Global Positioning System, or development of new propulsion concepts.2
Assessing the appropriate R&D investment by the Air Force in light of the ongoing revolution in information technology in the commercial sector is challenging. In this section the committee examines some of the specific sectors of information technology (IT) to draw distinctions for the Air Force.
This section starts with computing devices. Transistors, switches, and integrated circuits are covered; a separate section on space electronics is included because of the unique environmental requirements of space and its importance to the Air Force. Also covered are storage and display technologies. Computing architectures explores alternative paradigms for computing. Under communications, a number of areas are explored: optical materials and devices, radio frequency (RF) materials and devices; and RF and optical MEMS. Finally, information and signal processing and data fusion requirements are discussed.
Advances in information technology are a result of the ever-shrinking transistor, applied to almost every aspect of gathering and treating information. Continuing increases in information technology capabilities are dependent on continuing advances in the fabrication of ever more powerful computational hardware. Moore’s law, the exponential increase in integrated circuit functionality, continues today. Although there are potential limitations on the horizon, the semiconductor industry roadmap, ITRS,3 which is based on the scaling from larger devices that has served us so well for the last 40 years, foresees a continuation of the current rate of Moore’s law to the present roadmap horizon of 2016, corresponding to a 32-fold improvement in device density. Barriers to reaching and surpassing this density include the following, among others: lithography, gate oxide current leakage, interconnect requirements, and thermal issues. These challenges have spurred many research efforts, both to address the