logo

SCIENTIA SINICA Informationis, Volume 51 , Issue 5 : 735(2021) https://doi.org/10.1360/SSI-2019-0098

An energy-efficient incentive mechanism for resource recycling in multi-tenant datacenters

More info
  • ReceivedMay 13, 2019
  • AcceptedAug 13, 2019
  • PublishedApr 23, 2021

Abstract


Funded by

国家重点研发计划(2017YFB1001703)

国家自然科学基金(61722206,61761136014,61520106005,61802449)

国家高层次人才特殊支持计划


References

[1] Deng W, Liu F, Jin H. Harnessing renewable energy in cloud datacenters: opportunities and challenges. IEEE Network, 2014, 28: 48-55 CrossRef Google Scholar

[2] Gao P X, Curtis A R, Wong B, et al. It's not easy being green. In Proceedings of ACM SIGCOMM Conference on Applications, technologies, architectures, and protocols for computer communication, Helsinki, 2012. 211--222. Google Scholar

[3] Zhou Z. On the energy efficiency of green geo-distributed datacenters. Dissertation for Ph.D. Degree. Wuhan: Huazhong University of Science and Technology, 2017. Google Scholar

[4] Song J, Sun Z Z, Liu H, et al. Research advance on energy consumption optimization of hyper-powered data center. Chinese J Comput, 2018, 41: 2670--2678. Google Scholar

[5] Armbrust M, Fox A, Griffith R. A view of cloud computing. Commun ACM, 2010, 53: 50-58 CrossRef Google Scholar

[6] Xu Z, Li C. Low-entropy cloud computing systems. Sci Sin-Inf, 2017, 47: 1149-1163 CrossRef Google Scholar

[7] Ma J, Sui X F, Sun N H. Supporting Differentiated Services in Computers via Programmable Architecture for Resourcing-on-Demand (PARD). SIGPLAN Not, 2015, 50: 131-143 CrossRef Google Scholar

[8] Delimitrou C, Kozyrakis C. Quasar: resource-efficient and QoS-aware cluster management. In: Proceedings of the 19th International Conference on Architectural Support for Programming Languages and Operating Systems, 2014. 127--144. Google Scholar

[9] Sun X, Ansari N, Wang R. Optimizing Resource Utilization of a Data Center. IEEE Commun Surv Tutorials, 2016, 18: 2822-2846 CrossRef Google Scholar

[10] Lin C, Tian Y, Yao M. Green network and green evaluation: mechanism, modeling and evaluation. Chin J Comput, 2011, 34: 593-612 CrossRef Google Scholar

[11] Lin M, Wierman A, Andrew L L, et al. Dynamic right-sizing for power-proportional data centers. IEEE/ACM Transactions on Networking (TON), 2015, 21: 1378-1391. Google Scholar

[12] Zhou Z, Liu F, Zou R. Carbon-Aware Online Control of Geo-Distributed Cloud Services. IEEE Trans Parallel Distrib Syst, 2016, 27: 2506-2519 CrossRef Google Scholar

[13] Shaolei Ren , van der Schaar M. Dynamic Scheduling and Pricing in Wireless Cloud Computing. IEEE Trans Mobile Comput, 2014, 13: 2283-2292 CrossRef Google Scholar

[14] Feng C, Xu H, Li B. An Alternating Direction Method Approach to Cloud Traffic Management. IEEE Trans Parallel Distrib Syst, 2017, 28: 2145-2158 CrossRef Google Scholar

[15] Islam M A, Mahmud H, Ren S, et al. Paying to save: reducing cost of colocation data center via rewards. In: Proceedings of the 21st International Symposium on High Performance Computer Architecture (HPCA), San Francisco, 2015. 235--245. Google Scholar

[16] Greenberg A, Hamilton J, Maltz D A. The cost of a cloud. SIGCOMM Comput Commun Rev, 2008, 39: 68-73 CrossRef Google Scholar

[17] Neely M J. Stochastic Network Optimization with Application to Communication and Queueing Systems. Synthesis Lectures Communication Networks, 2010, 3: 1-211 CrossRef Google Scholar

[18] Zhou Z, Liu F M. An Energy-Efficient Incentive Mechanism for Resource Recycling in Multi-Tenant Datacenters. Technical Report, 2019. https://1drv.ms/b/s!Ar9mS_s-frkZgd0FE2XOinBcZ7A1Yg. Google Scholar

[19] Yao Y, Huang L B, Sharma A B. Power Cost Reduction in Distributed Data Centers: A Two-Time-Scale Approach for Delay Tolerant Workloads. IEEE Trans Parallel Distrib Syst, 2014, 25: 200-211 CrossRef Google Scholar

[20] Roh H, Jung C, Lee W, et al. Resource pricing game in geo-distributed clouds. In: Proceedings IEEE International Conference on Computer Communications, Torino, 2013. 1519--1527. Google Scholar

  • Figure 1

    (Color online) An illustration of the proposed incentive mechanism for resource recycling in multi-tenant datacenters

  • Figure 2

    (Color online) An illustration of the proposed online algorithm adaptively adjusting resource allocation based on the queue backlog $Q(t)$

  • Figure 3

    (Color online) The workload data of a Google cluster (a) and the real-time electricity price data of New York city (b)

  •   

    Algorithm 1 Online pricing algorithm for resource recycling in multi-tenant datacenters

    At the beginning of each time slot $t$, observing the current queue backlog $Q(t)$, backup resource price $R(t)$, backup resource capacity $C(t)$ and the response function $F_i(p(t),t)$ for each tenant $i$;

    Determine the optimal resource recycling $p(t)$ and backup resource usage $x(t)$, by solving the optimizationproblem (9;

    At the end of each time slot $t$, observe the amount of newly arrived batch workload $A(t)$ and update the queue backlog $Q(t+1)$ according to the following queuing dynamics: $$Q(t+1)=Q(t)- \sum_{i \in \mathcal{N}(t)} F_i(p(t),t)-x(t) + A(t).$$

qqqq

Contact and support