An evolutionary based heuristic approach towards energy efficient software defined data centers

Abstract

Today’s world mainly comprises of service computing, where everything is available as a service which in turn have led to exponential increase the user service demands. In order to meet the increasing service demands of the users, there has been rapid development in cloud computing. Due to these huge demands of cloud services many Data Centers (DC) have been deployed. This have led to rapid increase in the number of Data Centers and number of servers within the DC, while increasing the consumption of energy by these DCs. The main issue which is to be considered is the energy consumption by these Data Centers. A sincere consideration to the solution of the problem mentioned above is required for optimization of energy consumption without violating the QoS degradation. This paper proposes a management strategy for an efficient energy network that will lead to the fulfillment of demands in network traffic in SD-DCNs. The proposed model addresses three main issues, the subset of switch which is to be activated by using selective-switch-activation approach, an efficient routing for a multipath and for all the flow to be scheduled, and Installation of forwarding rule in SDN switches. The above-mentioned issues combined together are taken into consideration and formulated in the form of ILP problem. Finally, a GA based heuristic approach is formulated for solving the problem so as to tackle the computational complexity of the same. The presented simulation result lays the efficacy of the proposed algorithm.