Graph-based redundancy removal approach for multiple cross-layer interactions

Research has shown that the availability of cross-layer information from different protocol layers enable adaptivity advantages of applications and protocols which significantly enhance the system performance. However, the development of such cross-layer interactions typically residing in the OS is very difficult mainly due to limited interfaces. The development gets even more complex for multiple running cross-layer interactions which may be added by independent developers without coordination causing (i) redundancy in cross-layer interactions leading to a waste of memory and CPU time and (ii) conflicting cross-layer interactions. In this paper, we focus on the former problem and propose a graph-based redundancy removal algorithm that automatically detects and resolves such redundancies without any feedback from the developer. We demonstrate the applicability of our approach for the cross-layer architecture CRAWLER that utilizes module compositions to realize cross-layer interactions. Our evaluation shows that our approach effectively resolves redundancies at runtime.