MaxPass: Credit-Based Multipath Transmission for Load Balancing in Data Centers

Minkyung Park, Sungmin Sohn, Kwangwook Kwon, and Ted Taekyoung Kwon

105 558.558-568


Abstract :Various applications require the data center networksto carry their traffic efficiently. The data center networks usuallyhave a hierarchical topology and exhibit distinct traffic patterns,which is different from the traditional Internet. These features havedriven the data center networks to reduce the flow completion time(FCT) and to achieve high throughput. One of the possible solutionsis balancing network loads across multiple paths by leveragingtransport mechanisms like equal-cost multipath (ECMP) routing.ECMP allows flows to exploit multiple paths by hashing the meta-data of the flows. However, due to the random nature of hash func-tions, ECMP often distributes the traffic unevenly, which makes ithard to utilize the links’ full capacity. Thus, we propose an adap-tive load balancing mechanism for multiple paths in data centers,dubbed MaxPass, to complement ECMP. A sender adaptively se-lects and dynamically changes multiple paths depending on thecurrent network status like congestion. To monitor the networkstatus, the corresponding receiver transmits a probe packet peri-odically to the sender; its loss indicates a traffic congestion. Wecarry out the quantitative analysis on the ns-2 simulator to showthat MaxPass can improve the FCT and the throughput.​ 

Index terms :Data center, equal-cost multipath (ECMP), fat-tree topology, multipath transmission, transport layer protocol.