Edge router congestion control with congestion estimation

This paper deals with the problem of congestion control in a heterogeneous network scenario. Specifically, it introduces the concept of dynamic congestion estimation. The algorithm runs in the edge routers (ER) (the routers collecting the traffic between two different networks) with the aim of avoiding congestion in the network routers. ERs communicate via the network itself, which consequently introduces a variable feedback delay in the control loop. The algorithm is based on the Smith´s principle; the main novelty is that the algorithm adapts to the delay and rate variations, since it exploits on-line estimates obtained via probe packets.