Parallel Batch B+-tree Insertion on Multi-core Architectures

Multi-core computers have become more and more popular, which presents new opportunities for improving performance of database systems. How to exploit the potential performance on multi-core systems has become a key figure now. In database systems, B⁺-tree is the most widely-used index structure for speeding up performance of database systems. This paper implements the algorithm of parallel batch B⁺-tree insertion with multithreaded disk buffer. In our parallel batch B⁺-tree insertion algorithm (PBI), records can be inserted into index file concurrently by different threads. The key of the algorithm is to make traditional insertion into batch insertion. Our experimental results show that our system achieves a higher performance on multi-core systems.