Exploiting Parallelism to Accelerate Keyword Search on Deep-Web Sources

Increasingly, biological data is being shared over the deep web. Many biological queries can only be answered by successively searching a number of distinct web-sites. This paper introduces a system that exploits parallelization for accelerating search over multiple deep web data sources. An interactive, two-stage multi-threading system is developed to achieve task parallelization, thread parallelization, and pipelined parallelization. We show the effectiveness of our system by considering a number of queries involving SNP datasets. We show that most of the queries can be accelerated significantly by exploiting these three forms of parallelism.