Real-Time Application Domain Visionaries

Abstract

The gulf separating the real-time research, product development, and application development communities appears to be the largest in the entire computing field. This gulf has an unfortunate tendency to obscure the research opportunities of most value to the real-time application domains. It is an unsurprising consequence of factors historically intrinsic to each of those communities.Almost all academic, and much industrial, research in the area of real-time computing - and especially distributed real-time computing - suffers from two handicaps. First is the necessity for unusually extensive application domain-specific knowledge and understanding compared with what is required in most other computer science and engineering research. Second is the directly contrary reality that few real-time researchers have the option of obtaining such knowledge and understanding, because they lack access to non-trivial, deployed, real-time application environments - particularly to distributed ones (e.g., in process control or discrete manufacturing plants, defense systems, telecommunication intelligent network architectures) - and to the developers and users of those environments.The real-time application domain practitioners (real-time computer product and application developers and users) contribute to this gulf as well. First, the historical focus on traditional small, static, centralized, sampled-data subsystems has often subjected real-time computers to severe constraints on hardware cost, size, weight, and power. And although software development and maintenance costs per line of source code are generally one or two orders of magnitude greater for real-time software than for non-real-time software, the smallest scale instances of these subsystems actually have negligible costs for software compared with hardware. These two factors tend to discourage many of the application developers and users from believing that they need or can afford technology advances from the software and real-time research communities. Second, hardware parsimony makes it necessary, and application simplicity makes it possible, for the application programmers to do the majority of the system's resource management - most of it a' priori, and as little as possible at run time.The increasing importance and deployment of larger scale, more complex, more distributed real-time computer systems makes this gulf an increasingly greater obstacle to their cost-effectiveness and even viability. The purpose of this session is to assist in reducing this gulf by bringing to ISORC2K's researcher audience some of the most prominent visionaries in some of real-time computing's most important application domains.