Traffic Aggregation Based RSA (TARSA) for Time Varying Sub-Wavelength Connections in Elastic Optical Networks

Bandwidth in the spectrum sliced-elastic optical networks (EONs) can be allocated flexibly as per the users’ demand. Further, the allocated bandwidth on a connection can also be adjusted as per the traffic conditions and network scenarios. Currently, due to the growth in data centres, cloud computing based applications, and users’ tendency of using high bandwidth services, backbone network traffic has become more heterogeneous and time varying. This time varying and heterogeneous traffic in EONs is handled by the spectrum expansion/contraction (SEC) policies by adjusting the allocated spectrum as per the traffic requirement during network operation. In the SEC policies, the spectrum gap between anchor frequencies of two spectrum adjacent connections is shared to follow the traffic variations. The performance of the SEC policies depends on this spectral gap. This spectral gap is also termed as maximum allowable expansion region. Lower value of maximum allowable expansion region results in increased connection blocking and inefficient utilization of transmitter capacity. As a solution to these problems, we propose an efficient traffic aggregation based routing and spectrum allocation (TARSA) algorithm which increases the expansion region between the adjacent connections and improves the performance of the network. Performance evaluation of the proposed TARSA algorithm has been undertaken on the NSFNET and USNET network topologies on the metrics of average blocking rate and the percentage transmitters saving. It has been observed that the proposed TARSA improves the performance by reducing average blocking rate and by minimizing the number of transmitters used.