Improved concurrency control techniques for multi-dimensional index structures
K. V. Ravi Kanth, D. Serena, A.K. Singh · 2002
Multi-dimensional index structures such as R-trees enable fast searching in high-dimensional spaces. They differ from uni-dimensional structures in the following aspects: index regions in the tree may be modified during ordinary insert and delete operations; and node splits during inserts are quite expensive. Both these characteristics may lead to reduced concurrency of update and query operations. We examine how to achieve high concurrency for multi-dimensional structures. First, we develop a new technique for efficiently handling index region modifications. Then, we extend it to reduce/eliminate query blocking overheads during node-splits. We examine two variants of this extended scheme: one that reduces the blocking overhead for queries, and another that completely eliminates it. Experiments on image data on a shared-memory multiprocessor show that these schemes achieve up to 2 times higher throughput than existing techniques, and scale well with the number of processors.