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

다국어 초록 (Multilingual 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 metadataof the flows. However, due to the random nature of hash functions,ECMP often distributes the traffic unevenly, which makes ithard to utilize the links’ full capacity. Thus, we propose an adaptiveload balancing mechanism for multiple paths in data centers,dubbed MaxPass, to complement ECMP. A sender adaptively selectsand dynamically changes multiple paths depending on thecurrent network status like congestion. To monitor the networkstatus, the corresponding receiver transmits a probe packet periodicallyto 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.

참고문헌 (Reference)

1 A. Greenberg, "VL2: A scalable and flexible data center network" 51-62, 2009

2 T. Benson, "Understanding data center traffic characteristics" 65-72, 2009

3 R. Mittal, "Timely: Rtt-based congestion control for the datacenter" 45 : 537-550, 2015

4 A. Ford, "TCP extensions for multipath operation with multiple addresses, draft-ietf-mptcp-multiaddressed-09"

5 N. Bronson, "TAO: Facebook’s distributed data store for the social graph" 49-60, 2013

6 H. Zhang, "Resilient datacenter load balancing in the wild" 253-266, 2017

7 A. Dixit, "On the impact of packet spraying in data center networks" 2130-2138, 2013

8 S. McCanne, "Network simulator ns-2"

9 M. Honda, "Multipath congestion control for shared bottleneck" 357 : 378-, 2009

10 Y. Lu, "Multi-path transport for rdma in datacenters" 2018

11 A. R. Curtis, "Mahout : Low-overhead datacenter traffic management using end-host-based elephant detection" 1629-1637, 2011

12 J. Zhang, "Load balancing in data center networks: A survey" 20 (20): 2018

13 A. Roy, "Inside the social network’s(datacenter)network" 45 : 123-137, 2015

14 W. Bai, "Information-agnostic flow scheduling for commodity data centers" 455-468, 2015

15 C. Raiciu, "Improving datacenter performance and robustness with multipath TCP" 41 : 266-277, 2011

16 M. Al-Fares, "Hedera: Dynamic flow scheduling for data center networks" 2010

17 W. Wang, "Freeway: Adaptively isolating the elephant and mice flows on different transmission paths" 362-367, 2014

18 S. Wang, "Fdalb: Flow distribution aware load balancing for datacenter networks" 1-2, 2016

19 J. Perry, "Fastpass : A centralized zero-queue datacenter network" 44 (44): 307-318, 2015

20 Y. Cui, "FMTCP : A fountain code-based multipath transmission control protocol" 23 (23): 465-478, 2015

21 W. Bai, "Enabling ecn in multi-service multi-queue data centers" 537-549, 2016

22 M. Alizadeh, "Data center tcp (dctcp)" 41 (41): 63-74, 2011

23 I. Cho, "Credit-scheduled delay-bounded congestion control for datacenters" 239-252, 2017

24 H. Kung, "Credit-based flow control for atm networks : Credit update protocol, adaptive credit allocation and statistical multiplexing" 24 : 101-114, 1994

25 Y. Zhu, "Congestion control for large-scale rdma deployments" 523-536, 2015

26 M. Alizadeh, "Conga: Distributed congestion-aware load balancing for datacenters" 44 : 503-514, 2014

27 N. Katta, "Clove: How i learned to stop worrying about the core and love the edge" 155-161, 2016

28 L. Chen, "Auto: Scaling deep reinforcement learning for datacenter-scale automatic traffic optimization" 191-205, 2018

29 C. Hopps, "Analysis of an equal-cost multi-path algorithm"

30 M. Kheirkhah, "AMP: A better multipath tcp for data center networks"

31 C. Clos, "A study of non-blocking switching networks" 32 (32): 406-424, 1953

32 M. Al-Fares, "A scalable, commodity data center network architecture" 38 : 63-74, 2008