Due to the distributed and resource constrained nature of wireless sensor networks, their design proves to be difficult.
We present a resource management framework, which integrates a data-centric lightweight operating system with a publish/subscribe middleware.
In this framework, the main system abstraction is data for both local and networked processing.
The resultant system software is extended with a quality-aware adaptation mechanism, which configures system timeliness according to the actual application requirements.
A feedback-based control mechanism is used to iteratively tune the resultant data granularity in order to fit user requirements.
Our design is evaluated by simulations and the concepts were also implemented in our sensor network testbed.
|Name||Lecture Notes in Computer Science|
|Conference||9th International Conference on Principles of Distributed Systems, OPODIS|
|Period||12/12/05 → 14/12/05|
|Other||December 12-14, 2005|
- CAES-PS: Pervasive Systems