Modeling multi-path routing and congestion control under FIFO and fair queuing

Multi-path routing is a valuable on-line technique to deal with unpredictable and variable traffic patters, mostly for intra-domain TE, multi-homing, wireless mesh networks, metropolitan access networks, and has been shown efficient for a large spectrum of future traffic scenarios. In this paper we analyze the performance of MIRTO, TEXCP and TRUMP, three recently proposed multi-path routing algorithms. Modeling of such algorithms is performed through fluid models, based on ordinary differential equations (ODEs). On a US-like backbone network, with and without in-network fair queuing schedulers, TEXCP and TRUMP show faster convergence times while MIRTO, that relies on simpler feedbacks, consumes less network resources.