Modeling of traffic congestion and re-routing in a service provider network

This paper is a first step towards a mathematical and simulative environment to formally model the effect of router/link failures on the dynamics of TCP and UDP packet flows belonging to different end-user services (i.e. http, ftp, mailing and video streaming) transiting through a Service Provider backbone network. Our model takes into account network topology, routing (a simplified version of OSPFv2 on Ipv4), TCP (a full model of the TCP-SACK protocol) and different end-user services. This framework is easily extendable and generalizable to other protocols and to take into account QoS for packets priority in the routers. The final aim of this research line is to provide a framework to analyze quantitatively (using an appropriate QoS metric for each service) and ad-hoc (with respect to different end-user services) the effect of network failures, in order to appropriately design the network redundancy and implement diagnosis of future failures (supporting service and network troubleshooting) and/or prognosis (supporting network trend analysis) of failures by using feedback and feedforward statistical information on the users' behavior.