High-Performance Parallel Processing Aerospace Information Server

This paper focuses on development and implementation of a high-performance parallel processing information server for web based education and provision of aerospace information services. An innovative model of software architecture is presented to effectively utilize the computational power of a distributed server platform. It fulfills on-demand Internet-based client requests for astronomical image browsing and aerospace information exchange. In our design, a tuple-space programming paradigm is used to enable parallel processing of the image browsing requests. Additionally, a hashing algorithm is applied for fast searching of astronomical images in the database. Due to the different image resolutions and various transmission rates, multiple worker threads perform progressive, on-demand image decompression and wavelet transformations[1][2]. The server architecture supports thread, core, and/or processor-level parallel processing for high computing performance. The study also identifies and overcomes architectural bottlenecks in memory devices.