Future Generation Computer Systems 29 (2013) 1969–1980 Contents lists available at ScienceDirect Future Generation Computer Systems journal homepage: www.elsevier.com/locate/fgcs Towards greener data centers with storage class memory ✩ In Hwan Doh a , Young Jin Kim a , Eunsam Kim a , Jongmoo Choi b , Donghee Lee c , Sam H. Noh a,∗ a Department of Computer Engineering, Hongik University, Seoul, Republic of Korea b Division of Information and Computer Science, Dankook University, Yongin, Republic of Korea c School of Computer Science, University of Seoul, Republic of Korea highlights • New Storage Class Memory (SCM) technology makes it possible to turn systems on and off with minimal overhead. • Energy savings are possible by turning off idle servers while satisfying performance requirements. • The SOONN server prototype takes roughly 1.17 s to be turned on. • The ZEUS data center prototype consumes energy in proportion to user requests with configurable service of quality. • SCM should not be viewed as just a replacement of either RAM or storage. article info Article history: Received 28 February 2011 Received in revised form 29 February 2012 Accepted 20 May 2013 Available online 3 June 2013 Keywords: Power management Servers Idleness SCM NVRAM abstract Studies have shown that much of today’s data centers are over-provisioned and underutilized. Over- provisioning cannot be avoided as these centers must anticipate peak load with bursty behavior. Under- utilization, to date, has also been unavoidable as systems always had to be ready for that sudden burst of requests that loom just around the corner. Previous research has pointed to turning off systems as one solution, albeit, an infeasible one due to its irresponsiveness. In this paper, we present the feasibility of using new Storage Class Memory (SCM, which encompasses specific developments such as PCM, MRAM, or FeRAM) technology to turn systems on and off with minimum overhead. This feature is used to control systems on the whole (in comparison to the previous fine-grained component-wise control) in finer time scale for high responsiveness with minimized power lost to idleness. Our empirical study is done by executing real trace-like workloads on a prototype data center of embedded systems deploying FeRAM. We quantify the energy savings and performance trade-off by turning idle systems off. We show that our energy savings approach consumes energy in proportion to user requests with configurable quality of service. Finally, based on observations made on this data center, we discuss the requirements for real deployment. © 2013 Elsevier B.V. All rights reserved. 1. Introduction This study is motivated by two interests: the need for greener data centers and the development of new memory technology. Green computing is a topic of growing interest and various solutions have been proposed. In this paper, we consider reducing the so many underutilized and/or idle systems that have been provisioned for the uncommon peak loads and yet, consume a ✩ A preliminary version of this paper was published in the Proceedings of ACM International Conference on Computing Frontiers 2010 held in Bertinoro, Italy. This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MEST) (No. 2012R1A2A2A01045733). ∗ Corresponding author. E-mail address: samhnoh@hongik.ac.kr (S.H. Noh). significant amount of power doing almost nothing. To resolve this problem, we consider the use of new Storage Class Memory (SCM) technology, which has characteristics similar to DRAM and yet is non-volatile [1]. SCM such as PCM has recently become a topic of interest due to its scalability and energy efficiency [2–7]. Though originally planned to replace non-volatile storage, as its name implies, it is now being considered as a replacement of DRAM, which is now facing scaling limitations. Aside from a proof-of-concept feasibility study using SCM, an important point we drive in this paper is that SCM should not be considered just as a simple replacement of DRAM, but that new features and functionalities should be looked for as SCMs have dual characteristics of non-volatility and random byte addressability. Making use of the byte addressability characteristics for non- volatile storage performance enhancements have recently been studied [3,8]. Our work, to the best of our knowledge, is the first 0167-739X/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.future.2013.05.012