Many emerging applications rely on asynchronous data distribution, where several embedded computing notes act as sensors that feed data to other parts of the system. An important class of these systems is distributed real-time embedded (DRE) systems where the applications have timing constraints that must be correct, and the data have temporal intervals during which they are valid. This project addresses technology for asynchronous real-time data distribution in DRE systems, focusing on the substantial challenges of severely constrained resources in complex dynamic networks. It is developing new models of real-time scheduling theory, and implementations of the algorithms in a prototype sensor network. The models and algorithms being developed in this project will broaden the applicability of existing solutions developed for less-constrained environments. The project is also formalizing a taxonomy to describe the DRE data-distribution problem space. This taxonomy will provide researchers with a common tool to classify and apply future work in this field.
This project is being evaluated through theoretical analysis of algorithm goodness and correctness, through simulation, and through transition to the Coast Guard's Automated Identification System (AIS) application to sensor network applications at Raytheon. Through applications such as these, this project is expected to have an impact on many applications that employ real-time data distribution in sensor networks. The project is also contributing to Computer Science education in this emerging area, and through publication of algorithms and open-source release of the prototype software that adds to the base of widely applicable technology for DRE applications.
