1 / 51

Improving Datacenter Performance and Robustness with Multipath TCP

Improving Datacenter Performance and Robustness with Multipath TCP. Costin Raiciu † , Sebastien Barre ‡ , Christopher Pluntke † , Adam Greenhalgh † , Damon Wischik † , Mark Handley †. Before Start. 我想要先用我的話說一次這篇 Paper 想做的事情 : 這邊 Paper 的重點在於, single path 用在 TCP 上面的時代已經過去了, multipath TCP 才是主流。

temima
Télécharger la présentation

Improving Datacenter Performance and Robustness with Multipath TCP

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Improving Datacenter Performance and Robustness withMultipath TCP Costin Raiciu†, Sebastien Barre‡, Christopher Pluntke†, Adam Greenhalgh†, Damon Wischik†, Mark Handley†

  2. Before Start • 我想要先用我的話說一次這篇Paper想做的事情: • 這邊Paper的重點在於,single path用在TCP上面的時代已經過去了,multipath TCP才是主流。 • MultipathTCP又主要有三種topologies: fattree, VL2, Bcube。 • 這篇paper又分析了在不同情況下multipath確實比single path優秀的地方。還有一些影響performance的因素的分析。 • 最後是有沒有其他可能再發展的topology

  3. OUTLINE • Introduction • Data center networking • Topology. • Routing. • Path Selection. • Congestion Control. • Multipath TCP in summary • MPTCP in data center • Examples of Benefits • Analysis • Influence of Topology • Number of Subflows • Influence of the Traffic Matrix • Evolving topologies with MPTCP • Analysis • Discussion about DHFT • Summary

  4. Introduction • Traditional single path TCP has many significant problems, ill-suited for such network. (intra-data center traffic.) • Ex. Bottleneck, can’t scale up, congestion. • We propose using Multipath TCP as a replacement for TCP in such data centers.

  5. Introduction • Traditional data center topology. …

  6. Introduction • 4components to a data center architecture: • 1.physical topology • 2. routing • 3. path selection • 4. congestion control of traffic on the selected path.

  7. OUTLINE • Introduction • Data center networking • Topology. • Routing. • Path Selection. • Congestion Control. • Multipath TCP in summary • MPTCP in data center • Examples of Benefits • Analysis • Influence of Topology • Number of Subflows • Influence of the Traffic Matrix • Evolving topologies with MPTCP • Analysis • Discussion about DHFT • Summary

  8. Data center networking---Topology • Traditional data center topology. …

  9. Data center networking---Topology • That has many problems and can be improved a lot like we have discussed before. • So, fat-tree is one kind of solutions.

  10. Fat Tree Topology [Fares et al., 2008; Clos, 1953] K=4 Aggregation Switches K Pods with K Switches each Racks of servers

  11. Data center networking---Topology • And we also have VL2 and Bcube these two kinds of solutions.

  12. BCube Topology [Guo et al, 2009] BCube (4,1)

  13. Data center networking---Topology

  14. VL2 Topology [Greenberg et al, 2009, Clos topology] 10Gbps … 10Gbps 20 hosts

  15. Data center networking---Routing • We cannot expect the host itself to know which of these paths is the least loaded, so the routing system must spread traffic across these paths. • Use randomized load balancing, where each flow is assigned a random path from the set of possible paths.

  16. Data center networking---Path Selection • randomized load balancing有他一定的問題,就是有一定的機率某些path會有超大流量、同時會有一些path沒什麼(極低或零)流量。 • The use of a centralized flow schedulerhas been proposed.

  17. Multipath TCP: Congestion Control [NSDI, 2011]

  18. Data center networking---Congestion Control • Our hypothesis is that given sufficiently many randomly chosen paths, MPTCP will find at least one good unloaded path, and move most of its traffic that way. In so doing it will relieve congestion on links that got more than their fair share of ECMP balanced flows.

  19. Multipath TCP in summary • By making the window increase depend on the total window size, subflowsthat have large windows increase faster than subflows with small windows.

  20. OUTLINE • Introduction • Data center networking • Topology. • Routing. • Path Selection. • Congestion Control. • Multipath TCP in summary • MPTCP in data center • Examples of Benefits • Analysis • Influence of Topology • Number of Subflows • Influence of the Traffic Matrix • Evolving topologies with MPTCP • Analysis • Discussion about DHFT • Summary

  21. MPTCP in data center • Threemain benefits: • 1. Better aggregate throughput • 2.Better fairness • 3.Better robustness

  22. MPTCP in data center • Depend on 4 requirements: • 1. The congestion control scheme used. • 2. The physical topology. • 3. The traffic matrix generated by theapplications. • 4. The level of load in the network.

  23. Examples of Benefits • Throughput: • Using MPTCP in a FatTree network, we can see the more subflows we have, the more throughput we have. • The reason is simple, n flows share one link, each flow need to achieves 1/n capacity.

  24. Examples of Benefits

  25. Examples of Benefits • Fairness: • Every host’s throughput is shownranked in order of increasing throughput. Is is clear that not onlydid the utilization improve with MPTCP, but also the fairness improved.

  26. Examples of Benefits • 這是一個累積百分比的圖,可以看出single TCP有些pathperform 真的很差(很閒的意思),而MPTCP則大幅改善此情況。

  27. Analysis • Then, we have questions: • How many subflows are needed? • How does the topology affect results? • How does the traffic matrix affect results?

  28. Analysis • We need to view in these aspects to answer the question: • Influence of Topology • Number of Subflows • Influence of the Traffic Matrix

  29. Influence of Topology • We compare VL2 and Bcube.

  30. Influence of Topology • Why Bcube is relative low? • I guess, just guess, is because in Bcube, pathes between two hosts are not in same length which lead to worse flows management.

  31. Number of Subflows • We weresurprised that eight subflows were needed to achieve 90% throughput. • Why eight, and on what does this depend?

  32. Number of Subflows • Results:

  33. Influence of the Traffic Matrix

  34. OUTLINE • Introduction • Data center networking • Topology. • Routing. • Path Selection. • Congestion Control. • Multipath TCP in summary • MPTCP in data center • Examples of Benefits • Analysis • Influence of Topology • Number of Subflows • Influence of the Traffic Matrix • Evolving topologies with MPTCP • Analysis • Discussion about DHFT • Summary

  35. Evolving topologies with MPTCP • In single homed topologies: • Hosts links are often bottlenecked. • ToR switches failure also lead to crash tens of hosts.

  36. Evolving topologies with MPTCP • 以FatTree為例子

  37. Evolving topologies with MPTCP • So, how do we improved? • There are two solutions: perfect switch and dual-homed FatTree.

  38. Evolving topologies with MPTCP • Perfect switch: • Serves as a good control experiment, giving an upper bound on what any network core might provide using single links to the hosts.

  39. Evolving topologies with MPTCP • Dual-homed FatTree (DHFT): • is to the host and four connect the linksbetween the two layers of switches. • If we remove one portper host from the core and use it to connect the second interfaceon each server, the network requires the same numberof switch ports.

  40. Analysis • QUESTION: • Will perfect switch really improve performance?

  41. Analysis • Testing throughput over MPTCP, perfect switch TCP, normal TCP.

  42. Analysis • QUESTION: • Will perfect switch really improve performance? • ANSWER: • Yes, it is.

  43. Analysis • QUESTION: • Dual-homed FatTree really work? • Does it have any problem?When may troubles occur?

  44. Analysis • Testing relative throughput over MPTCP, TCP DHFT, TCP perfect switch.

  45. Analysis • QUESTION: • Dual-homed FatTree really work? • Does it have any problem?When may troubles occur? • ANSWER: • Yes, it works only when load is low. When load is high, throughput will go down near to TCP.

  46. Discussion about DHFT • DHFT’s worst case performance is 75% and best case is around 200%. • 在流量不穩定的情況下,DHFT + MPTCP有可能會有較差的表現產生。 • Beyond performance, DHFT improves robustness: any lowerpodswitch failure does not cut-off an entire rack of servers. • DHFT is not optimal by any measure, but it shows that we cancreate topologies with better performance if we assume MPTCP isthe transport protocol.

  47. OUTLINE • Introduction • Data center networking • Topology. • Routing. • Path Selection. • Congestion Control. • Multipath TCP in summary • MPTCP in data center • Examples of Benefits • Analysis • Influence of Topology • Number of Subflows • Influence of the Traffic Matrix • Evolving topologies with MPTCP • Analysis • Discussion about DHFT • Summary

  48. Summary • “One flow, one path”thinking has constrained datacenter design • Collisions, unfairness, limited utilization • Multipath transport enables resource pooling in datacenter networks: • Improves throughput • Improves fairness • Improves robustness

  49. Reference • conferences.sigcomm.org/sigcomm/2011/slides/s266.ppt‎

More Related