JudiShare: Judicious resource allocation for QoS-based services in shared wireless sensor networks

In shared wireless sensor networks (WSNs), multiple users request access to sensing resources, often with varying sampling rates and QoS requirements. To accommodate a request, appropriate sensing, computing and communication resources need to be allocated across the network. Traditionally, each request is mapped to a dedicated set of resources, even when the requests are similar. The need to reduce resource usage has led to data virtualization techniques that focus primarily on merging the requests ignoring the QoS requirements. In this paper, we present a QoS-aware resource allocation approach, JudiShare, that merges requests, where possible, if they are compatible in their requirements, providing judicious reuse of both sensing and communication resources through a mixture of data virtualization and Virtual Network Embedding (VNE). We show that JudiShare respects QoS requirements and reduces resource usage to up to 60%. This, in turn, allows up to 50% more requests to be accommodated onto the network, even when the network resources are fully utilized.