Detecting network anomalies in Ethernet/MPLS/IP networks using Ethernet OAM performance data

Abstract

The effective monitoring of big data and telecom networks is becoming more and more crucial nowadays. Network anomalies can cause a decrease in performance , prevent the normal functioning of a network or put the network out of service completely. Deployment and monitoring cost of large networks are costly so network operators are constantly looking for solutions to reduce this cost. One approach employed was the enlargement of the Ethernet technology scope for use in WAN networks that run on gigabit speeds. The Ethernet became a technology used in WAN networks and now known as carrier network. The lack of Operations, Administration and Maintenance (OAM) functionalities for its management was a problem leading to the creation of Ethernet OAM which is an operation, administration and management tool for Ethernet networks. In this thesis we evaluate the accuracy by which the Ethernet OAM can help detect anomalies in a network. we simulate a gigabit Ethernet Network and introduce common network anomalies to investigate this. A comparison is done between the loss and delay experienced by the traffic in the network and the different rates of the Ethernet OAM stream present in the network. Results obtained show that Ethernet OAM can effectively be used at a certain rate to detect anomalies in a network. Finally, some suggestions for future work is given.