Programmable Re-tasking of Wireless Sensor Networks Using WISEMAN

In this paper we present a flexible middleware platform for re-tasking Wireless Sensor Networks (WSNs) that we coin WISEMAN. Based on our previous experiences with mobile agents in computer networks, we developed a lightweight interpreter of text-based codes that enables their deployment in order to implement diverse WSNs tasks. WISEMAN occupies 19Kbytes of TinyOS embedded code, and 3 Kbytes of memory to operate in commercially available sensor nodes. We examine different agent migration methodologies, and present performance evaluations to gauge their efficiency in terms of delay and bandwidth with aims to determine which approach works best depending on the intended agent application. Our results indicate that WISEMAN agents can migrate as fast as 235 mS per-hop, which is comparable to existing approaches, while supporting the necessary code execution flexibility needed for the rapid implementation and deployment of WSN re-tasking programs.