Opportunistic Power Savings with Coordinated Control in Data Center Networks

ABSTRACT

In this paper, we propose a holistic power management of the data center network (DCN) via coordination among distinct controllers with different functionality and scope. This includes - a local controller (LC) that has visibility only at individual switch/router level, a global controller (GC) that has a global view of the network, and a hint based topology aware user request assignment controller (RAC) that controls placement of the external requests on the endpoint hosts. The key function of these controllers from energy management perspective is to properly direct and consolidate network traffic to maximize low power (or "sleep") opportunities for the network interfaces. We show that the coordinated "hints" provided by the GC are vital to correct the myopic view of LCs and RAC for both avoiding congestion and maximizing network sleep opportunities. We show that these mechanisms can reduce power consumption by upto ~45% in the common fat-tree based DCNs using the low power idle (LPI) feature of the Ethernet.

References

1. Caida data server. http://data.caida.org/datasets/passive/passive-oc48/20020814-160000.UTC/pcap/.
2. Abts et al. Energy proportional datacenter networks. SIGARCH Comput. Archit. News, 38(3):338--347, June 2010.
3. Al-Fares et al. Hedera: Dynamic flow scheduling for data center networks. In USENIX NSDI, pages 19--19, 2010.
4. Blanquicet et al. An initial performance evaluation of rapid phy selection (rps) for energy efficient ethernet. In IEEE LCN, pages 223--225, 2007.
5. R. E. Brown, R. Brown, E. Masanet, B. Nordman, B. Tschudi, A. Shehabi, J. Stanley, J. Koomey, D. Sartor, P. Chan, and E. I. U. E. P. A. et al. with Alliance to Save Energy, ICF Incorporated. Report to Congress on Server and Data Center Energy Efficiency: Public Law 109--431. 2007.
6. Christensen et al. Ieee 802.3 az: the road to energy efficient ethernet. IEEE Communications Magazine, 48(11), 2010.
7. Erickson et al. Using Network Knowledge to Improve Workload Performance in Virtualized Data Centers. PhD thesis, Stanford University, 2013.
8. Gao et al. Energy-aware load balancing in heterogeneous cloud data centers. In ACM ICMSS, pages 80--84, 2017.
9. Greenberg et al. The cost of a cloud: Research problems in data center networks. SIGCOMM Comput. Commun. Rev., 39(1):68--73, 2008.
10. M. Gupta et al. Using low-power modes for energy conservation in ethernet lans. In IEEE INFOCOM, pages 2451--2455, 2007.
11. H. Jin et al. Joint host-network optimization for energy-efficient data center networking. In IEEE IPDPS, pages 623--634, 2013.
12. K. Kant. Multistate power management of communications links. In Proc. of COMSNET, 01 2011.
13. D. Li et al. Joint power optimization through vm placement and flow scheduling in data centers. In IEEE IPCCC, pages 1--8, 2014.
14. Liu et al. Greencloud: A new architecture for green data center. In ACM ICAC-INDST, pages 29--38, 2009.
15. P. Mahadevan et al. On energy efficiency for enterprise and data center networks. IEEE Communications Magazine, 49(8):94--100, 2011.
16. Meng et al. Improving the scalability of data center networks with traffic-aware virtual machine placement. In IEEE INFOCOM, pages 1154--1162, 2010.
17. Mostowfi et al. A simulation study of energy-efficient ethernet with two modes of low-power operation. IEEE Communications Letters, 19(10):1702--1705, 2015.
18. Mostowfi et al. An analytical model for the power consumption of dual-mode eee. Electronics Letters, 52(15):1308--1310, 2016.
19. M. Murugan et al. On the interconnect energy efficiency of high end computing systems. Sustainable Computing: Informatics and Systems, April 2012.
20. Nedevschi et al. Reducing network energy consumption via sleeping and rate-adaptation. In USENIX NSDI, pages 323--336, 2008.
21. Paya et al. Energy-aware load balancing and application scaling for the cloud ecosystem. IEEE Transactions on Cloud Computing, 2015.
22. L. R. Rabiner et al. An introduction to hidden markov models. IEEE ASSp Magazine, 1986.
23. Raghavendra et al. No power struggles: Coordinated multi-level power management for the data center. In ACM ASPLOS, pages 48--59, 2008.
24. Riley et al. The ns-3 network simulator. Modeling and tools for network simulation, pages 15--34, 2010.
25. Sohan et al. Characterizing 10 gbps network interface energy consumption. In IEEE LCN, pages 268--271, 2010.
26. Sondur et al. Implementing data center energy management capabilities in ns3. In publication.
27. C. Thaenchaikun et al. Augmenting the energy-saving impact of ieee 802.3az via the control plane. In IEEE ICCW, pages 2843--2849, 2015.
28. Travostino et al. Seamless live migration of virtual machines over the man/wan. Future Gener. Comput. Syst., 22(8):901--907, 2006.
29. Verma et al. pmapper: power and migration cost aware application placement in virtualized systems. In ACM/IFIP/USENIX International Conference on Middleware, pages 243--264, 2008.
30. X. Wang et al. Carpo: Correlation-aware power optimization in data center networks. In IEEE INFOCOM, pages 1125--1133, 2012.
31. K. Zheng et al. Joint power optimization of data center network and servers with correlation analysis. In IEEE INFOCOM, pages 2598--2606, 2014.

Index Terms

• Computing methodologies

  • Modeling and simulation

    • Simulation support systems

      • Simulation tools

• Hardware

  • Power and energy

    • Power estimation and optimization

      • Enterprise level and data centers power issues

• Networks

  • Network performance evaluation

    • Network simulations