Wednesday, 13 December 2017

Award Winner 2013

Dynamic Server Provisioning for Data Center Power Management

by Anshul Gandhi

Press Release
Extended Abstract

Data centers play an important role in today's IT infrastructure. However, their enormous power consumption makes them very expensive to operate. Sadly, much of the power used by data centers is wasted because of poor capacity management, leading to low server utilization.

In order to reduce data center power consumption, researchers have proposed several dynamic server provisioning approaches. However, there are many challenges that hinder the successful deployment of dynamic server provisioning, including: (i) unpredictability in workload demand, (ii) switching costs when setting up new servers, and (iii) unavailability of data when provisioning stateful servers. Most of the existing research in dynamic server provisioning has ignored, or carefully sidestepped, these important challenges at the expense of reduced benefits. In order to realize the full potential of dynamic server provisioning, we must overcome these associated challenges.

This thesis provides new research contributions that explicitly address the open challenges in dynamic server provisioning. We first develop novel performance modeling tools [1,7,8] to estimate the effect of these challenges on response time and power. In doing so, we also address several long-standing open questions in queueing theory, such as the analysis of multi-server systems with switching costs [1,8]. We then present practical dynamic provisioning solutions [2-6,9] for multi-tier data centers, including novel solutions that allow scaling the stateful caching tier [4], and solutions that are robust to load spikes [2,3]. Our implementation results using realistic workloads and request traces on a 38-server multi-tier testbed demonstrate that dynamic server provisioning can successfully meet typical response time guarantees while significantly lowering power consumption.

While this thesis focuses on server provisioning for reducing power in data centers, the ideas presented herein can also be applied to: (i) private clouds, where unneeded servers can be repurposed for "valley-filling" via batch jobs, to increase server utilization, (ii) community clouds, where unneeded servers can be given away to other groups, to increase the total throughput, and (iii) public clouds, where unneeded virtual machines can be released back to the cloud, to reduce rental costs.

[1] Anshul Gandhi, Sherwin Doroudi, Mor Harchol-Balter, and Alan Scheller-Wolf. “Exact Analysis of the M/M/k/setup Class of Markov Chains via Recursive Renewal Reward”. In Proceedings of Sigmetrics 2013, pages 153-166.
[2] Anshul Gandhi, Mor Harchol-Balter, Ram Raghunathan, and Michael Kozuch. “AutoScale: Dynamic, Robust Capacity Management for Multi-Tier Data Centers”. In Transactions on Computer Systems, Volume 30, Issue 4, Article 14.
[3] Anshul Gandhi, Timothy Zhu, Mor Harchol-Balter and Michael Kozuch. “SOFTScale: Stealing Opportunistically For Transient Scaling”. In Proceedings of Middleware 2012, pages 142-163.
[4] Timothy Zhu, Anshul Gandhi, Mor Harchol-Balter and Michael Kozuch. “Saving Cash by Using Less Cache”. In Proceedings of HotCloud 2012.
[5] Anshul Gandhi, Mor Harchol-Balter and Michael Kozuch. “Are sleep states effective in data centers?”. In Proceedings of IGCC 2012, pages 113-122.
[6] Anshul Gandhi, Yuan Chen, Daniel Gmach, Martin Arlitt and Manish Marwah. “Minimizing Data Center SLA Violations and Power Consumption via Hybrid Resource Provisioning” (Best Paper Award). In Proceedings of IGCC 2011, pages 49-56.
[7] Anshul Gandhi, Varun Gupta, Mor Harchol-Balter and Michael Kozuch. “Optimality Analysis of Energy-Performance Trade-off for Server Farm Management”. In Performance Evaluation, Volume 67, Issue 11, pages 1155-1171.
[8] Anshul Gandhi, Mor Harchol-Balter and Ivo Adan. “Server Farms with Setup Costs”. In Performance Evaluation, Volume 67, Issue 11, pages 1123-1138.
[9] Anshul Gandhi, Mor Harchol-Balter, Rajarshi Das, and Charles Lefurgy. “Optimal Power Allocation in Server Farms”. In Proceedings of Sigmetrics 2009, pages 157-168.

Full Text

Hosted by CMU