Y.-J. Kang, T.-K. Kim, K.-E. Yang, S.-H. Jung, and W.-S. Cho (Korea)
Database Cluster System, Data Warehouse, OLAP, QueryDecomposition Technique
Performance of OLAP queries becomes a critical issue as the amount of data in the data warehouses increases rapidly in most areas of business. We proposed a high performance database cluster system called HyperDB to solve the performance issue. HyperDB is an OLAP engine running on multiple Linux PCs connected with high speed network, and each PC participates in the computation of complex OLAP queries. We propose a query decomposition technique to solve skew problem among queries on top of HyperDB. Histogram for the data in the partition key has been adopted for balanced query response time. Performance experiments under the TPC-R benchmark shows better performance benefit than an existing approach.
Important Links:
Go Back
