Pages: pp. 22
Most of us think of communication as a technology supporting computation, especially since networking has so dramatically changed how we do things over the course of this past decade. However, as we hope to convey in this issue, the modeling and design of tomorrow's communications systems require some of the most advanced scientific techniques available today. Communications technology, in all of its forms, provides fascinatingly rich computational problems, running the gamut from modeling physical properties of a medium and the signals propagating through it to simulation and analysis of global computer networks.
Donald McManus and yours truly, the guest editors, coauthor the first of the theme articles, "Horizons in Scientific and Distributed Computing." It's a bit of an outlier in that it surveys several software and networking technologies that are on the periphery of scientific computing today but hold promise for playing a greater role in the future. Topics covered include Java, MPI, OpenMP, components, agents, IPv6, and ATM. We briefly introduce these technologies and their current status, and we give some projections about future prospects.
"Fast Computational Techniques for Indoor Radio Channel Estimation," by Marc Kimpe, Harry Leib, Olivier Maquelin, and Ted H. Szymanski is next, and it develops parallel-ray-tracing techniques for radio coverage. The revolution in cellular telephony will shortly be felt in mobile computing, using both laptop computers and personal digital assistants. Offices, laboratories, and even cities will soon be festooned with wireless base stations for ubiquitous network access. This article introduces a propagation model and ray-tracing method that can quickly and accurately simulate channel properties of indoor radio base stations, as a function of their location.
The next article, "Modeling the Global Internet," by James H. Cowie, David Nicol and Andy T. Ogielski, presents techniques for modeling packet traffic in huge data networks. Analytic queuing and other "classical" approaches to network modeling are simply not scaling to the size and complexity needed for understanding how global networks behave. We have already seen "Internet storms" and other types of phenomena occur but have no real understanding of how and why they arise. This article delves into the technology needed to make progress there.
The third article, by Tim Olson, Dennis Healy, and Ulf Österberg entitled "Wavelets in Optical Communications," delves into the challenges of maximizing usable bandwidth in optical fiber communication channels. Those challenges include generating and detecting signals with enormous bandwidths using electronics-based hardware. The authors develop orthogonal filterbanks based on wavelet theory and show how this approach can improve the achievable bandwidth in any such channel.
The last article, "Satellite-Constellation Design" by Martin Lo, is an introduction to the problems associated with multiple satellite placement subject to the constraints of orbital mechanics and applications needs. Some industry projections indicate that the number of telecommunications and imaging satellites in orbit will double over the next five years. Two satellite-based telephone and data network projects alone, Teledesic (underwritten largely by Bill Gates and Craig McCaw) and Iridium (led by Motorola), will be launching hundreds of new satellites in the next few years. Starting from the basics of satellite orbital mechanics, Lo develops the role that some surprising approaches, such as ergodic theory, can help with constellation-design problems.
We realize full well that we have left out important areas of research and development and invite readers to fill in the gaps we have created in this theme issue with future articles, essays, or departments.