Index selection for NoSQL database with deep reinforcement learning

Abstract With the development of big data technology, management complex applications has become more and resource intensive. In this paper, we propose an automatic approach (DRLISA) to achieve NoSQL database index selection. For different workloads, automatically select its corresponding indexes parameters which can totally improve performance. Our DRLISA establishes optimal by building a deep reinforcement learning model is able adapt dynamic change workloads. We conducted our experiments in five aspects (the impact manipulation, operation count, comparison with random selection, existing method robustness DRLISA) using open source benchmark, YCSB. The experimental results showed that high efficient recommendation under