While military applications will increasingly rely on COTS, there remain unique requirements for which technological advances in basic computing hardware and software will be required. These fall into the realm of so–called "grand challenge," which will require trillions of floating point operations per second (teraflops). Several approaches are being considered, each of which, if ultimately realized, is likely to offer certain inherent advantages for different applications. HPC and scalable parallel systems are of particular importance. Optical processing techniques combine elements of both and are being pursued as a means of increasing inherent parallelism and computational throughput. Software advances are seen as a way to allow aggregation of very large numbers of computing elements. Both of these approaches lend themselves to solutions to complex deterministic problems (i.e., problems for which a sequence of calculations to reach a specific solution can be defined). By contrast, neural networks provide a better way of attacking less determinate problems.

Table E–10 highlights significant capabilities and trends in key areas of computing and software. HPC is an area of international R&D. In addition to France, which is recognized as a world leader in photonics, Japan and Russia have had strong programs in optical computing. Germany has a growing interest and has strong capabilities in production of photodynamically active bacteriorhodopsin films that may be an enabling technology for future optical/molecular computers. Israel has a small but sound and growing EO infrastructure as well. The growth of the Internet and multimedia are producing growing demand for development and global implementation of very high–speed digital networks. Development of these is an international activity, with cooperation among major telecommunications firms. One example is the Japanese Real–World Computing (RWC) program, which includes a number of other countries as participants.

The United States has dominated and is projected to continue to drive the state–of–the art in HPC; Japan has strong capabilities. However, Japan has dominated in areas of "traditional" supercomputing high–cost mainframes and vector processors

The United States has pioneered a variety of technologies for scalable distributed processing based on U.S. microprocessor designs whose computational power continues to double approximately every 18 months. These configurations now dominate the market. Availability of affordable HPC capability has also led to a growing level of international interest and work in intelligent systems and human–composite interfaces.

Massively parallel processing (MPP) and neural network programming could be applied to numerous applications covered by ASTMP milestones and objectives. M&S are examples of applications requiring the computing speed and power offered by MPP techniques, while neural network programming may be more useful in the development of decision aids. Only a few countries have the supporting infrastructure necessary for major R&D in these technologies. World leaders include the United States, Japan, Germany (with whom an active Data Exchange Agreement (DEA) exists), and to a lesser extent the United Kingdom and France.

MPP and neural network programming are important aspects of the Army’s electronic battlefield (EBF) concept. MPP will contribute significantly to simulation and virtual reality (VR) components of the EBF.

Military requirements for processing real–time signals and imagery data severely challenge existing computing capabilities. Optical processing offers potential advantages for these applications and is an important area of technology development where other nations have world–leading capabilities.

ASTMP goals include demonstration of a two–dimensional (2D) optical processor capable of running real–time automatic target recognition (ATR) and signal processing algorithms on data from imaging sensors such as the synthetic aperture radar (SAR) or EO systems. Near–term ASTMP milestones include the development of optical interconnections for computers; photonic and electronic devices integrated on the same chip; image–forming light modulators, and an order of magnitude improvement in spatial light modulation dynamic range.

Optical processing techniques are well suited for analysis of data generated by these high–volume throughput applications. The development of photonic devices necessary for optical computing are of significant interest to the U.S. Army and have numerous military applications. World leaders in photonics/EO include the United States and Japan, followed by France, the U.K., and Germany.

The network throughput demands of international telecommunications firms are primary drivers of the state of the art in most networking areas, including fiber optic communications and optical switching (including wave division multiplexing techniques). All of the major telecommunications–producing nations—the United States, United Kingdom, Japan, France, Germany, and Canada, followed closely by China and Israel, have good capabilities in fiber optic networks. The implementation of the 5–10 gigabits per second (Gbps) fiber optic cable that will link Europe and intermediate points in Africa and Asia with Japan will almost certainly speed proliferation of this technology. While Japan and selected regions of Europe may lead in deployment of high speed fiber–optic cables, implementation in other areas is limited primarily by economic considerations rather than technology. In the critical area of switching the United States, Canada, and the United Kingdom have the strongest technological positions, followed very closely by Germany and France.

International software developments are enabled by widespread availability of very powerful microprocessor–based symmetrical multiprocessing systems. A number of countries, including Israel, India, and Russia, are actively engaged in commercial cooperative software developments.

In software one key to achieving our goals for M&S is the implementation of advanced algorithms, specifically for MPP. Currently only a few countries possess the supporting infrastructure necessary for major R&D in this area. World leaders include the United States, Japan, Germany (with whom there is an active agreement), the United Kingdom, and France.

AI (or machine intelligence or intelligent systems) is an area of worldwide research interest. One area that is particularly promising for international collaboration is artificial neural networks (ANNs); for example, the optical ANNs being pursued by Japan as part of the RWC initiative. Another area is the application of AI to so–called intelligent agents for collecting information and managing operations in a distributed battlefield command, control, communication computers, intelligence, and information system. For example, Australia has a particularly strong presence and activity on the Internet World Wide Web. Much of the work is theoretical in nature, and many of the problems are tractable with modest computing power, widely available in the commercial market. This active and effective research in AI can be found in most developed or developing countries. Much of this work is being driven by the Internet or by requirements for managing and administering extremely large, complex telecommunications systems. In addition to work in the United States, which is the world leader in this area, Japan’s RWC initiative has a strong component of AI. Strong capabilities in intelligent agents also reside in the U.K. and Germany, followed closely by France. AI capabilities are found in many other countries.

One of the effects of increased computer hardware performance and communications bandwidth has been to spur rapid interest and growth in VR. While the U.S. holds or shares a lead in most areas of HCI research, the U.K. (which has an existing cooperative effort in helmet–mounted displays (HMDs) with the Air Force and NASA Ames), France (visually–coupled displays and digital scene generation), Canada (head–mounted stereo displays and large data set visualization), Germany (applications to robotics and teleoperations), and Japan (visually–coupled systems) have world–leading development efforts. Other countries have niches of capability, two notable examples being the strong capability in haptic devices at the University of Pisa in Italy, and Israel’s work in heads–up displays.