Dynamic code morphing in network embedded systems

The use of mobile code in embedded, resource limited systems like Wireless Sensor Networks (WSN) is an opportunity and a challenge at the same time. The opportunity lies in the dynamic re-tasking and run-time adaption that can considerably extend the functional envelope of the deployed hardware. The price to pay, however, is additional communication overhead that results in shorter lifetime and decreased performance. That is why the system's scarce resources must be utilized with even more efficiency than usual. But optimization methods applied at design-time lead to case-restricted solutions, or the dismissal of mobile code solutions altogether. In this work we challenge ourselves with an integrated design of a system that addresses the increase of communication overhead by online code compression. The main task of the proposed method is to extract semantics from the transmitted mobile code at run-time and to tie it to the on-node holder, a dictionary of some type, avoiding costly code (re-)transmissions. By doing so, the actual code representation is brought to a near optimal form for each specific task covering both, the dynamic re-tasking and the reduction of communication overhead. The distinctive feature of the method is that it can adapt to the changes in code structure, code content, and regional usage patterns at run-time without interruption of system operation. The low computational complexity of the method allows to use it in resource-constrained devices like WSN and to implement time-sensitive applications.