Saving space and time using index merging

摘要

Managing digital information is an integral part of our society. Efficient access to data is supported through the use of indices. Although indices can reduce the cost of answering queries, they have two significant drawbacks: they take additional storage space and their maintenance can become a bottleneck. We address these challenges by introducing search data structures that reduce the need for storing redundant data among indices. Our experimental results with the main-memory version of these data structures show that our approach can reduce by half the storage space and can improve performance, where the highest performance improvement is achieved for workloads with high update ratios. Our experimental results with the secondary-storage version of the data structures show that our approach produces a solution that can outperform both IBM DB2 and Microsoft SQL Server on the popular TPC-C workload.