Algorithms for efficient processing of complex queries in node-partitioned data warehouses

In This work we propose and evaluate an algorithm for efficient processing of complex queries in a partitioned data warehouse. Partitioning allows cheap computer nodes to be able to host and process efficiently large data warehouses. In order for the system to be truly useful, significant speedup should be achieved for all query patterns. We describe the system, identify complex join queries and propose algorithms to deal efficiently with those queries. We also take a set of performance-constrained computer nodes and the query set from the decision support performance evaluation benchmark TPC-H as a study case and evaluation basis for our analysis.