Using a Tuple Space Web Service for Parallel Processing in Bioinformatics

This paper discusses the use of a Tuple Space Web Service for the construction of parallel processing applications in the field of bioinformatics. This system is based on Linda, which is a coordination language for parallel and distributed processing. The use of a coordination language greatly simplifies the parallelization of bioinformatics applications. Since this system is conceived as a web service, it provides language portability, has a simple and easily understandable architecture and does not depend on a dedicated network of workstations. To illustrate the utility of the tuple space web service, a DNA string searching application has been developed. The results of this application show very favorable performance benefits arising from the use of the web service on a network of commodity workstations. The speedup measured for the application on 30 processors was 29.1 times the one measured for one processor, which means that the execution time was reduced by 97.1%.