Experimental Investigation of the EIGRP Routing Protocol Properties with Respect to Fault Tolerance and Load Balancing

Abstract

The article presents an experimental study of the proprietary EIGRP protocol in the GNS3 environment to analyze its properties regarding support for multipath, fault-tolerant routing strategies, and the impact of these solutions on load balancing across primary routes with different metrics. The exper-iments were conducted on a network topology for several sets of initial data regarding interface/route metrics and variance coefficient values, which influenced the selection of optimal paths, the set of primary and backup routes for ensuring the fault tolerance of the information and communication network (ICN), as well as the load balancing between them both under normal operating conditions and in the event of network equipment failures. The results of the experimental study confirmed the fast convergence and adaptation to network state changes caused by variations in route metrics and ICN topology due to component failures. The adaptation decisions were accompanied by a revision of the optimal paths, changes in the set of primary and backup routes, and adjustments in load balancing between the primary routes to improve the quality of service. It was demonstrated that the EIGRP protocol balances load by automatically recalculating the ratio of share counter values for each primary path in the routing table, which is inversely proportional to the metric ratios of optimal and non-optimal routes. This is done to equalize the load across routes with different metrics, level packet delays across different paths, and minimize packet jitter caused by multipath routing. Monitoring of routing and topology tables enables real-time tracking of changes to the set of optimal and non-optimal, primary and backup paths, along with their characteristics and statuses. The loop-prevention mechanism in EIGRP also prevents the use of high-metric paths for load balancing.