SmartBench
Peeyush Gupta, Michael J. Carey, Sharad Mehrotra, oberto Yus
Abstract
This paper proposes SmartBench, a benchmark focusing on queries resulting from (near) real-time applications and longer-term analysis of IoT data. SmartBench, derived from a deployed smart building monitoring system, is comprised of: 1) An extensible schema that captures the fundamentals of an IoT smart space; 2) A set of representative queries focusing on analytical tasks; and 3) A data generation tool that generates large amounts of synthetic sensor and semantic data based on seed data collected from a real system. We present an evaluation of seven representative database systems and highlight some interesting findings that can be considered when deciding what database technologies to use under different types of IoT query workloads.