Towards intelligent datacenter traffic management : using automated fuzzy inferencing for elephant flow detection

Effective traffic management has always been one of the key considerations in datacenter design. It plays an even more important role today in the face of increasingly widespread deployment of communication intensive applications and cloud- based services, as well as the adoption of multipath datacenter topologies to cope with the enormous bandwidth requirements arising from those applications and services. Of central importance in traffic management for multipath datacenters is the problem of timely detection of elephant flows flows that carry huge amount of data so that the best paths can be selected for these flows, which otherwise might cause serious network congestion. In this paper, we propose FuzzyDetec, a novel control architecture for the adaptive detection of elephant flows in multipath datacenters based on fuzzy logic. We develop, perhaps for the first time, a close loop elephant flow detection framework with an automated fuzzy inference module that can continually compute an appropriate threshold for elephant flow detection based on current information feedback from the network. The novelty and practical significance of the idea lie in allowing multiple imprecise and possibly conflicting criteria to be incorporated into the elephant flow detection process, through simple fuzzy rules emulating human expertise in elephant flow threshold classification. The proposed approach is simple, intuitive and easily extensible, providing a promising direction towards intelligent datacenter traffic management for autonomous high performance datacenter networks. Simulation results show that, in comparison with an existing state-of-the-art elephant flow detection framework, our proposed approach can provide considerable throughput improvements in datacenter network routing.Published versio