Cluster Computing

, Volume 19, Issue 3, pp 1555–1570 | Cite as

Workload-aware resource management for software-defined compute

  • Yoonsung Nam
  • Minkyu Kang
  • Hanul Sung
  • Jincheol Kim
  • Hyeonsang Eom
Article

Abstract

With advance of cloud computing technologies, there have been more diverse and heterogeneous workloads running on cloud datacenters. As more and more workloads run on the datacenters, the contention for the limited shared resources may increase, which can make the management of the resources difficult, often leading to low resource utilization. For effective resource management, the management software for the datacenters should be redesigned and used in a software-defined way to dynamically allocate “right” resources to workloads based on different characteristics of workloads so that they can decrease the cost of their operation while meeting the service level objectives such as satisfying the latency requirement. However, recent datacenter resource management frameworks do not operate in such software-defined ways, thus leading to not only the waste of resources, but also the performance degradation. To address this problem, we have designed and developed a workload-aware resource management framework for software-defined compute. The framework consists mainly of the workload profiler and workload-aware schedulers. To demonstrate the effectiveness of the framework, we have prototyped the schedulers that minimize the interferences on the shared computing and memory resources. We have compared them with the existing schedulers in the OpenStack and VMWare vSphere testbeds, and evaluated its effectiveness in high contention scenarios. Our experimental study suggests that the use of our proposed approach can lead to up to 100 % improvements in throughput and up to 95 % reductions in tail latency for latency critical workloads compared to the existing ones.

Keywords

Datacenter Resource management Cloud computing Virtualization Workload-awareness Memory intensity 

References

  1. 1.
    Linden, G.: Make data useful (2006)Google Scholar
  2. 2.
  3. 3.
  4. 4.
  5. 5.
    Gulati, A., Shanmuganathan, G., Holler, A.M., Ahmad, I.: Cloud scale resource management: challenges and techniques. In: Proceedings of the 3rd USENIX Conference on Hot Topics in Cloud Computing, pp. 3:1–3:6 (2011)Google Scholar
  6. 6.
  7. 7.
    Zhuravlev, S., Blagodurov, S., Fedorova, A.: Addressing shared resource contention in multicore processors via scheduling. In: ACM SIGARCH Computer Architecture News, vol. 38, pp. 129–142. ACM (2010)Google Scholar
  8. 8.
    Intel(r) 64 and ia-32 architectures software developer’s manualGoogle Scholar
  9. 9.
    Spec 2006 benchmark. https://www.spec.org/cpu2006
  10. 10.
    Kim, S., Eom, H., Yeom, H.Y.: Virtual machine consolidation based on interference modeling. J. Supercomput. 66(3), 1489–1506 (2013)CrossRefGoogle Scholar
  11. 11.
    Cheng, L., Wang, C.L.: vBalance: using interrupt load balance to improve i/o performance for smp virtual machines. In: Proceedings of the Third ACM Symposium on Cloud Computing, pp. 2:1–2:14. ACM (2012)Google Scholar
  12. 12.
    Gordon, A., Amit, N., Har’El, N., Ben-Yehuda, M., Landau, A., Schuster, A., Tsafrir, D.: Eli: bare-metal performance for i/o virtualization. ACM SIGPLAN Not. 47(4), 411–422 (2012)Google Scholar
  13. 13.
    Li, J., Sharma, N.K., Ports, D.R., Gribble, S.D.: Tales of the tail: hardware, OS, and application-level sources of tail latency. In: Proceedings of the ACM Symposium on Cloud Computing, pp. 1–14. ACM (2014)Google Scholar
  14. 14.
    Tu, C.C., Ferdman, M., Lee, C.T., Chiueh, T.C.: A comprehensive implementation and evaluation of direct interrupt delivery. In: Proceedings of the 11th ACM SIGPLAN/SIGOPS International Conference on Virtual Execution Environments, pp. 1–15. ACM (2015)Google Scholar
  15. 15.
    Xu, Y., Bailey, M., Noble, B., Jahanian, F.: Small is better: avoiding latency traps in virtualized data centers. In: Proceedings of the 4th annual Symposium on Cloud Computing, pp. 7:1–7:16. ACM (2013)Google Scholar
  16. 16.
    Little, J.D., Graves, S.C.: Little’s law. In: Building Intuition, pp. 81–100. Springer, New York (2008)Google Scholar
  17. 17.
  18. 18.
  19. 19.
    Atikoglu, B., Xu, Y., Frachtenberg, E., Jiang, S., Paleczny, M.: Workload analysis of a large-scale key-value store. In: ACM SIGMETRICS Performance Evaluation Review, vol. 40, pp. 53–64. ACM (2012)Google Scholar
  20. 20.
    Delimitrou, C., Kozyrakis, C.: iBench: quantifying interference for datacenter applications. In: Workload Characterization (IISWC), 2013 IEEE International Symposium on, pp. 23–33. IEEE (2013)Google Scholar
  21. 21.
    Mars, J., Tang, L., Hundt, R., Skadron, K., Soffa, M.L.: Bubble-up: increasing utilization in modern warehouse scale computers via sensible co-locations. In: Proceedings of the 44th annual IEEE/ACM International Symposium on Microarchitecture, pp. 248–259. ACM (2011)Google Scholar
  22. 22.
    Zhang, X., Tune, E., Hagmann, R., Jnagal, R., Gokhale, V., Wilkes, J.: Cpi 2: Cpu performance isolation for shared compute clusters. In: Proceedings of the 8th ACM European Conference on Computer Systems, pp. 379–391. ACM (2013)Google Scholar
  23. 23.
  24. 24.
    Hindman, B., Konwinski, A., Zaharia, M., Ghodsi, A., Joseph, A.D., Katz, R.H., Shenker, S., Stoica, I.: Mesos: a platform for fine-grained resource sharing in the data center. In: Proceedings of the 8th USENIX Conference on Networked Systems Design and Implementation, pp. 295–308 (2011)Google Scholar
  25. 25.
    Zaharia, M., Chowdhury, M., Das, T., Dave, A., Ma, J., McCauley, M., Franklin, M.J., Shenker, S., Stoica, I.: Resilient distributed datasets: a fault-tolerant abstraction for in-memory cluster computing. In: Proceedings of the 9th USENIX conference on Networked Systems Design and Implementation, pp. 15–28. USENIX Association (2012)Google Scholar
  26. 26.
    Delimitrou, C., Kozyrakis, C.: Paragon: Qos-aware scheduling for heterogeneous datacenters. ACM SIGARCH Comput. Archit. News 41(1), 77–88 (2013)Google Scholar
  27. 27.
    Delimitrou, C., Kozyrakis, C.: Quasar: resource-efficient and qos-aware cluster management. ACM SIGPLAN Not. 49(4), 127–144 (2014)Google Scholar
  28. 28.
  29. 29.
    Karanasos, K., Rao, S., Curino, C., Douglas, C., Chaliparambil, K., Fumarola, G.M., Heddaya, S., Ramakrishnan, R., Sakalanaga, S.: Mercury: hybrid centralized and distributed scheduling in large shared clusters. In: 2015 USENIX Annual Technical Conference (USENIX ATC 15), pp. 485–497 (2015)Google Scholar
  30. 30.
  31. 31.
    Yang, H., Breslow, A., Mars, J., Tang, L.: Bubble-flux: precise online qos management for increased utilization in warehouse scale computers. ACM SIGARCH Comput. Archit. News 41(3), 607–618 (2013)CrossRefGoogle Scholar
  32. 32.
    Lo, D., Cheng, L., Govindaraju, R., Ranganathan, P., Kozyrakis, C.: Heracles: improving resource efficiency at scale. In: Proceedings of the 42nd Annual International Symposium on Computer Architecture, pp. 450–462. ACM (2015)Google Scholar
  33. 33.
    Leverich, J., Kozyrakis, C.: Reconciling high server utilization and sub-millisecond quality-of-service. In: Proceedings of the Ninth European Conference on Computer Systems, pp. 4:1–4:14. ACM (2014)Google Scholar

Copyright information

© Springer Science+Business Media New York 2016

Authors and Affiliations

  • Yoonsung Nam
    • 1
  • Minkyu Kang
    • 1
  • Hanul Sung
    • 1
  • Jincheol Kim
    • 2
  • Hyeonsang Eom
    • 1
  1. 1.Department of Computer Science and EngineeringSeoul National UniversitySeoulSouth Korea
  2. 2.AI Tech Lab., Future Technology R&D Center, Corporate R&D CenterSK TelecomSeoulSouth Korea

Personalised recommendations