Block level query scrambling algorithm within distributed multi-database environment

Abstract

Remote data access from disparate sources across a wide area network such as the Internet is problematic due to the unpredictable nature of the communications source and the potential delays at the remote site. Current query plan scrambling techniques require all blocks from a relation be present to perform join operations based on an execution plan. Delays arising from the remote site can result in total abandoning of a plan and starting a new execution plan. In this paper we present an algorithm that uses block level query scrambling techniques to performs real time query scrambling on relations with remote site delays. The algorithm starts scrambling as soon as there are a few blocks available from the remote sites and continues scrambling as more blocks arrive. The algorithm utilises the changing of the query execution plan at the sub-tree level. We present an example that demonstrates the effectiveness of this algorithm.