Abstract
In data center networks (DCNs), the presence of long-lived TCP flows tends to bloat the switch buffers. As a consequence, short-lived TCP-incast traffic suffers repeated losses that often lead to loss recovery via timeout. Because the minimum retransmission timeout (minRTO) in most TCP implementations is fixed to around 200 ms, interactive applications that often generate short-lived incast traffic tend to suffer unnecessarily long delays waiting for the timeout to elapse. The best and most direct solution to such problem would be to customize the minRTO to match DCNs delays; however, this is not always possible; in particular in public data centers where multiple tenants, with various versions of TCP, co-exist. In this paper, we propose to achieve the same result by using techniques and technologies that are already available in most commodity switches and data centers and that do not interfere with the tenant’s virtual machines or TCP protocol. In this approach, we rely on the programmable nature of SDN switches and design a SDN-based incast congestion control (SICC) framework, that uses a SDN network application in the controller and a shim-layer in the host hypervisor, to mitigate incast congestion. We demonstrate the performance gains of the proposed scheme via real deployment in a small-scale testbed as well as ns2 simulation experiments in networks of various sizes and settings.
Similar content being viewed by others
Notes
We define the traffic inertia of a flow as the worst case amount of traffic that the flow can inject in the network before congestion can be detected. This is typically the maximum amount of traffic that could be lost before the loss is detected and recovery invoked.
Typically, the throughput of internal networking stack is 50–100 Gb/s. This is fast enough to handle tens of concurrent VMs sharing a single or several physical links. Hence, the window update function added to the vswitch would not hog the CPU nor affect the achievable throughput.
The hardware switch was not used because its OF-DPA implementation does not follow OF1.5 specifications [22] which allows for matching on TCP flags. The support for OF1.5 is introduced into OvS starting from version 2.3.
References
Abdelmoniem AM, Bensaou B (2015) Efficient switch-assisted congestion control for data centers: an implementation and evaluation. In: Proceedings of the IEEE international performance computing and communications conference (IPCCC)
Abdelmoniem AM, Bensaou B (2015) Incast-aware switch-assisted TCP congestion control for data centers. In: Proceedings of the IEEE global communications conference (GlobeCom)
Abdelmoniem AM, Bensaou B (2015) Reconciling mice and elephants in data center networks. In: Proceedings of the IEEE CloudNet conference
Abdelmoniem AM, Bensaou B, Abu AJ (2017) SICC: SDN-based incast congestion control for data centers. In: Proceedings of the IEEE international conference on communications (ICC)
Akyildiz IF, Lee A, Wang P, Luo M, Chou W (2014) A roadmap for traffic engineering in SDN-OpenFlow networks. Comput Netw 71:1–30
Alizadeh M Data Center TCP (DCTCP). http://simula.stanford.edu/alizade/Site/DCTCP.html
Alizadeh M, Greenberg A, Maltz DA, Padhye J, Patel P, Prabhakar B, Sengupta S, Sridharan M (2010) Data center TCP (DCTCP). ACM SIGCOMM CCR 40:63–74
Alizadeh M, Javanmard A, Prabhakar B (2011) Analysis of DCTCP: stability, convergence, and fairness. In: Proceedings of the ACM SIGMETRICS
Apache.org: Apache HTTP server benchmarking tool. http://httpd.apache.org/docs/2.2/programs/ab.html
Benson T, Akella A, Maltz DA (2010) Network traffic characteristics of data centers in the wild. In: Proceedings of the ACM SIGCOMM
Chen W, Ren F, Xie J, Lin C, Yin K, Baker F (2015) Comprehensive understanding of TCP Incast problem. In: Proceedings of the IEEE INFOCOM
Chowdhury M, Zhong Y, Stoica I (2014) Efficient coflow scheduling with varys. In: Proceedings of the ACM SIGCOMM, pp 443–454
Feamster N, Rexford J, Zegura E (2013) The road to SDN. Queue 11:20–40
iperf: The TCP/UDP bandwidth measurement tool. https://iperf.fr/
Jouet S, Perkins C, Pezaros D (2016) OTCP: SDN-managed congestion control for data center networks. In: Proceedings of IEEE/IFIP network operations and management symposium (NOMS)
Kandula S, Sengupta S, Greenberg A, Patel P, Chaiken R (2009) The nature of data center traffic. In: Proceedings of the ACM IMC
Karakus M, Durresi A (2017) A survey: control plane scalability issues and approaches in software-defined networking (SDN). Comput Netw 112:279–293
Lu Y, Zhu S (2015) SDN-based TCP congestion control in data center networks. In: Proceedings of IEEE IPCCC
Mckeown N, Anderson T, Peterson L, Rexford J, Shenker S, Louis S (2008) OpenFlow: enabling innovation in campus networks. ACM SIGCOMM CCR 38:69–74
NS2: the network simulator ns-2 project. http://www.isi.edu/nsnam/ns
Open Networking Foundation (2013) SDN architecture overview. Tech. rep., Open Networking Foundation
opennetworking.org: OpenFlow v1.5 Specification. https://www.opennetworking.org/sdn-resources/openflow
openvswitch.org: Open virtual switch project. http://openvswitch.org/
Panda A, Scott C, Ghodsi A, Koponen T, Shenker S (2013) CAP for networks. In: Proceedings of the ACM HotSDN workshop
Rijsinghani A (1994) RFC 1624—computation of the internet checksum via incremental update. https://tools.ietf.org/html/rfc1624
Ryu Framework Community: Ryu: a component-based software defined networking controller. http://osrg.github.io/ryu/
Vasudevan V, Phanishayee A, Shah H, Krevat E, Andersen DG, Ganger GR, Gibson GA, Mueller B (2009) Safe and effective fine-grained TCP retransmissions for datacenter communication. ACM SIGCOMM CCR 39:303–314
Wu H, Feng Z, Guo C, Zhang Y (2013) ICTCP: Incast congestion control for TCP in data-center networks. IEEE/ACM Trans Netw 21:345–358
Zhang J, Ren F, Tang L, Lin C (2015) Modeling and solving TCP incast problem in data center networks. IEEE Trans Parallel Distrib Syst 26(2):478–491
Acknowledgements
This work is supported in part under Grants: HKPFS PF12-16707, FSGRF13EG14, REC14EG03, and FSGRF14EG24.
Author information
Authors and Affiliations
Corresponding author
Additional information
The source code, simulation, and experiments is available at https://ahmedcs.github.io/SICC/
This paper is an extended version of [4].
Rights and permissions
About this article
Cite this article
Abdelmoniem, A.M., Bensaou, B. & Abu, A. Mitigating incast-TCP congestion in data centers with SDN. Ann. Telecommun. 73, 263–277 (2018). https://doi.org/10.1007/s12243-017-0608-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s12243-017-0608-1