Extrapolating Trends in Distributed System Costs: The Future of Distributed Graphic Computing
At the beginning of the microprocessor revolution, Gordon Moore at Intel considered the economies of scale evident after a few years of increasing the density of transistors on processor chips and projected that transistor density would double every year for years to come. This observation became known as Moore's Law. Although it has been modified over the years, the fundamental advance of computing technologies have brought remarkable accessibility to computing power as a result of rapidly declining costs and improving performance. This paper examines the historical data for microprocessors, hard drives, network interface cards, graphics processors, random access memory, and datacommunications to identify basic patterns in the price to performance relationship. The resulting patterns are used to project what a distributed graphics computer system might look like in ten years and what new or improved human-technology interactions might be supported. As compared to current and past systems, the leading edge technologies are likely to be even more focused on graphics processing and network connectivity with wider access based on lower costs.
Keywords: Distributed Systems, Human-Technology Useability
Dr. G. Kent Webb
Professor, School of Business, San Jose State University