Semantic Network Layering

Abstract: The stack in various forms has been widely used as an architectural template for networking systems. Recently the stack has been subject to criticism for a lack of flexibility. However, when it comes right down to it nobody has offered a truly compelling alternative. Various cross-layer optimizations have been proposed, but these optimizations are frequently hacks to achieve a particular goal and offer no direct insight into why the existing network stack is inadequate. We propose that a fundamental problem with the existing network stack is that it attempts to layer functionality that is not well-suited to layering. In this work we use a "bottom up" model of information computation, storage, and transfer and the "top down" goals of networking systems to formulate a modular decomposition of networking systems. Based on this modular decomposition we propose a semantic layered structure for networking systems that eliminates many awkward cross-layer interactions that arise in the canonical layered stack.