M-ECho: A Middleware for Morphable Data-Streaming in Pervasive Systems

The end-to-end performance of pervasive mobile systems is commonly dictated by the availability of resources at the weakest link. However, a number of run-time adaptations or morphing steps can be performed to tune the system performance. In this paper, we present M-ECho, a middleware for system morphing. M-ECho is designed with focus on data streaming applications, specifically in the field of pervasive mobile systems. We consider an autonomous robotics application comprising of a set of cooperating mobile robots to demonstrate and evaluate M-ECho's system morphing capabilities. Optimizations are based upon metrics of average instantaneous power consumption at a single node (local) as well as the power consumed by all participants (global). Experimental results show that M-ECho is able to achieve improved end-to-end performance with its dynamic code morphing techniques.