Middleware Support for a Self-Configurable Wireless Sensor Network

This paper describes a self-configurable middleware and a node execution platform to support autonomous sensor networks. We achieve self-configuration by scheduling and strategies similar to load balancing (mapping) that is integrated in our proposed middleware. On the node execution platform we decide on the fly between microprocessor and FPGA realization of hybrid tasks. We propose a combination of the middleware with the node execution platform to achieve a fully self-configurable system. If an error occurred in a self-configurable network it is able to overcome it that makes the network more reliable and stable. In the case a node fails all the tasks running on this node are mapped automatically by our middleware to other nodes inside the sensor network. To enable the migration and mapping it is important to identify the requirements of all nodes within the sensor network and their tasks.