La Nina: framework in self-adaptive publish/subscribe middleware for wireless sensor networks

Abstract

This paper focuses on three key issues that the publish/subscribe (pub/sub) communication scheme faces for event routing in wireless sensor networks (WSNs): (1) balancing trade-offs among conflicting performance objectives such as data yield, data fidelity and power efficiency; (2) satisfying quality of service (QoS) requirements such as latency and (3) considering noise in evaluating event routing performance. To address these issues, this paper investigates self-adaptive event routing in TinyDDS, which is pub/sub middleware for WSNs. With its noise-aware and constraint-based evolutionary multiobjective optimisation framework, La Nina, TinyDDS autonomously adapts its routing parameters to dynamic network conditions by reducing the impacts of noise on performance evaluation and seeking the optimal trade-offs among performance objectives under given QoS requirements. Simulation results validate this ability of TinyDDS in large-scale, dynamic and noisy WSNs. TinyDDS is implemented lightweight enough to operate on resource-limited sensor nodes.