Robust Index Selection for Stochastic Dynamic Workloads

Abstract Fast query processing is a primary goal of modern database systems. The use indexes crucial to reduce the execution times queries. Hence, it great interest determine an efficient selection for management system (DBMS). However, index problems are highly challenging as cause additional memory consumption and individual benefit influenced by others. In this paper, we consider accounting non-standard features, such (i) multiple potential workloads, (ii) different risk-averse objectives, (iii) multi-index configurations, (iv) reconfiguration costs, (v) anticipation dynamic workload scenarios. For problem extensions, propose specific model formulations, which can be solved efficiently using solver-based solution techniques. applicability performance our concepts demonstrated reproducible synthetic workloads well standard TPC-H TPC-DS-based benchmark workloads.