Real-time ontology-based context-aware situation reasoning framework in pervasive computing

Currently, the use of mobile applications for smart environments have multiplied, and they play a fundamental role in people’s daily lives. While there are many applications based on a microservices architecture that can simultaneously handle various context data provided by heterogeneous connected objects, this is not the case with mobile devices for real-time composite user-centric urgent situations. In fact, it becomes necessary to rethink a new way to minimize both the ratio of missed situations and the response time by combining competitive microservices with a priority-based parallel reasoning strategy. In this paper, we introduce a novel flexible, modular and hierarchical loosely coupled framework which can be employed for parallel composite situations that rule management and reasoning. To provide such reasoning, we proposed an innovative parallel reasoning process, which involved a set of filtered and factorized user’s situations rules, priority-based on the emergency of the events and situations to speed up the queued urgent situation rules, and parallel situations identification across collaborating microservices. We compared the proposed approach with existing reasoning approaches of the smart domains use cases in terms of expired event ratio, missed situation ratio, event detection throughput and latency. The experiments show that the proposed approach achieved low expired event and missed situation ratios, going up to 0.07 and 0.1 respectively. In addition, results show that the proposed approach can detect events in a high workload with an average detection latency of 2.4ms, also that the event detection throughput is improved to 82.4% thanks to a new shared lightweight thread approach in comparison with pipelining-based approach.