An Elastic Hybrid Sensing Platform: Architecture and Research Challenges

The dynamic provisioning of hybrid sensing services that integrates both WSN and MPS is a promising, yet challenging concept. It does not only widen the spatial sensing coverage, but it also enables different types of sensing nodes to collaboratively perform sensing tasks and complement each other. Furthermore, it allows for the provisioning of a new category of services that was not possible to implement in pure WSN or MPS networks. Offering a hybrid sensing platform as a service results in several benefits including, but no limited to, efficient sharing and dynamic management of sensing nodes, diversification and reuse of sensing services, as well as combination of many sensing paradigms to enable data to be collected from different sources. However, many challenges need to be resolved before such architecture can be feasible. Currently, the deployment of sensing applications and services is a costly and complex process, which also lacks automation. This paper motivates the need for hybrid sensing, sketches an early architecture, and identifies the research issues with few hints on how to solve them. We argue that a sensing platform that reuses the virtualization and cloud computing concepts will help in addressing many of these challenges, and overcome the limitations of today's deployment practices.