An Efficient Mice Flow Routing Algorithm for Data Centers Based on Software-Defined Networking

Abstract

A significant problem affecting the overall performance of Data Center Networks (DCNs) based on Software-Defined Networking (SDN) is the delay introduced by controllers to mice flows. In the literature, there exist approaches facing this problem by compiling and installing paths for mice and elephants dynamically. However, such approaches have shortcomings related to the large number of routing rules that the switches must handle, which may also lead to extensive delays to mice flows. In this paper, we propose MiceDCER, an algorithm that efficiently routes mice flows in SDN-based DCNs by assigning internal Pseudo-MAC (PMAC) addresses to the edge switches and hosts. Aiming at reducing the number of routing rules, MiceDCER installs wildcard rules based on the information carried by the Address Resolution Protocol (ARP) packets. Our evaluation reveals MiceDCER significantly reduces the number of rules installed in switches and, therefore, contributes to reducing the delay in SDN-based DCNs.