Cache-Based Aggregate Query Shipping: An Efficient Scheme of Distributed OLAP Query Processing

Our study introduces a novel distributed query planrefinement phase in an enhanced architecture of distributed queryprocessing engine (DQPE). Query plan refinement generates potentiallyefficient distributed query plan by reusable aggregate query shipping(RAQS) approach. The approach improves response time at the cost ofpre-processing time. If the overheads could not be compensated byquery results reusage, RAQS is no more favorable. Therefore a globalcost estimation model is employed to get proper operators: RR\_Agg,R\_Agg, or R\_Scan.For the purpose of reusing results of queries with aggregate function in distributedquery processing, a multi-level hybrid view caching (HVC) scheme isintroduced. The scheme retains the advantages of partial match andaggregate query results caching. By our solution, evaluations withdistributed TPC-H queries show significant improvement on averageresponse time.