Benchmarking IoT Context Management Platforms: High-level Queries Matter

Abstract

While IoT hardware and application silos achieved significant progress in recent years, the realization of full IoT potential is still limited by the lack of successful middleware platforms for horizontal IoT context sharing and integration, as well as by the lack of common standards on context access interfaces and APIs. In this paper, we propose and discuss the benefits of high-level context query languages over basic languages from performance and expressive power points of view. We argue why existing approaches to benchmarking of IoT middleware are not sufficient for emerging Context Management Platforms (CMP). We propose our benchmarking and reporting framework, which is based on queries’ richness as one of the main variables as well as the notation for a compact description of that variable, which can be used both as an input for the load generation and the final benchmarking report. We demonstrate and validate our proposals using Context Definition and Query Language (CDQL), which has been developed for the Context-as-a-Service (CoaaS) IoT platform as part of the EU Horizon-2020 project bIoTope.