Intelligent routing in congested Approximate Flow-Aware Networks

A new algorithm for intelligent routing in Approximate Flow-Aware Networks (AFAN) is presented and analyzed in the paper. In contrast to basic AFAN, where cross-protect routers block excessive traffic under congestion, it assumes a possibility of using other routes to destination nodes based on the idea similar to load balancing. The analysis is provided for the AFAN concept, which is the newest version of Flow-Aware Networks (FAN). The simulation experiments, discussed in the paper, were provided for basic AFAN and for the network with one of two congestion control mechanisms, the Remove and Accept Most Active Flows (RAMAF) and Simple Congestion Control Mechanism (SCCM). These solutions allow for fast acceptance of streaming flows independently of the volume of traffic load in a network. The results are presented for four different simulation topologies and prove the advantages of the proposed intelligent routing approach. Moreover, it is shown that the SCCM ensures better transmission properties than the RAMAF.