A traffic scheduling strategy based on cost function for differentiated class of service in multi-domain optical networks

Abstract The emerging cloud computing, IoT, bulk-data store-and-forward datacenter applications are dominating the Internet traffic, which contains the commercial and business data services and behaves periodically. The traffic peak remains high during a certain time each day (known as peak traffic hours). This results in heavy load on the network, leading to higher level of connections blocking. In order to handle such a critical situation in evolving multi-vendor/multi-operator based multi-domain optical networks (MDONs), prioritization of connection requests based on the differentiated class of service (CoS) is a must. The differentiated CoS (DiffCoS) may be categorized as: CoS1-delay-tolerant, CoS2-deadline-driven, and CoS3-fixed scheduled (or hard time bounded) traffic. To avail the maximum benefit of DiffCoS traffic characteristics, proper planning and efficient resource provisioning during admission phase is essential for the network service providers. A traffic scheduling strategy based on cost function (TSSCF) for DiffCoS in MDON is proposed, in which the connection requests are arranged in decreasing order of the cost function, and the route is selected based on shortest path minimum domain criterion, as per which the shortest route traversing or passing through the minimum number of domains is chosen. The simulations are performed on three-domain ARPANET topology, and the results are evaluated for different intra and inter-domain request ratios i.e., 1:1, 3:2 and 7:3, under 1:1:1 ratio of various CoSs. The simulation results of the proposed TSSCF strategy show significant improvement on the metric of blocking probability (BP), individual CoS BP, resource utilization ratio, and a newly introduced metric called revenue index as compared to the existing strategies.