Robust routing paradigms exploiting path diversity: modeling, analysis and design

In this paper we consider feedback-based routing exploiting path diversity that jointly uses best path selection and flow control for optimality and stability. The contributions of the present paper are manifold. We show numerical evidence that routing performance are insensitive to rate coordination along available routes in practical scenarios. Route uncoordination arises in case of in-network per flow rate sharing imposed by link schedulers and is known to be suboptimal. However, we show the gap is squeezed out on medium sized networks with different kind of realistic traffic demands. Furthermore, we design a distributed routing algorithm whose performance are evaluated and compared with TRUMP and TEXCP, two recently proposed multi-path routing algorithms. On a US-like backbone network, with and without in-network scheduling our algorithm proves to work as better as the other two while being much simpler in a real protocol implementation. Modeling and analysis of such algorithms is performed through fluid models based on ordinary differential equations. We finalize our design into a real protocol evaluated in an experimental network.