Autonomic management of heterogeneous sensing devices with ECA rules

Facing to the great number of wireless sensor/actuator devices, smart applications have an increasingly important requirement: autonomy. Besides autonomy in terms of energy, these devices have also to be functionally autonomous: they have to be self-discovered, self-(re-)configured and selfhealed. This paper presents a self-manageable autonomic platform for heterogeneous sensing systems. Based on a service oriented architecture and an Event-Condition-Action model, management services collect information from managed elements (e.g., sensors, proxies, gateways) and perform management actions according to user defined ECA rules. We define a generic event model that allows us covering a large set of events. Events are captured and processed by rule evaluators which then perform corresponding management actions on the sensors. We show with experimental results that the mechanism brings little overhead in terms of energy consumption and execution time.