1 / 29

An Efficient Threshold-Based Power Management Mechanism for Heterogeneous Soft Real-Time Clusters

An Efficient Threshold-Based Power Management Mechanism for Heterogeneous Soft Real-Time Clusters. Leping Wang, Ying Lu University of Nebraska-Lincoln, USA September 4, 2014. Outline. Motivation Related Work Problem Statement Threshold-based approach Evaluation Conclusion. Motivation.

azure
Download Presentation

An Efficient Threshold-Based Power Management Mechanism for Heterogeneous Soft Real-Time Clusters

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. An Efficient Threshold-Based Power ManagementMechanism for HeterogeneousSoft Real-Time Clusters Leping Wang, Ying Lu University of Nebraska-Lincoln, USA September 4, 2014

  2. Outline • Motivation • Related Work • Problem Statement • Threshold-based approach • Evaluation • Conclusion

  3. Motivation • Why power management (PM) for heterogeneous clusters • The power-related costs dominate the total cost of ownership of a cluster system • Most PM mechanisms are applicable to homogenous systems • Heterogeneous clusters are already widespread

  4. Motivation • Opportunities for PM in heterogeneous clusters • Turn off or hibernate idle servers • Dynamically scale operating frequency/voltage (DVS) for underutilized servers • Distribute more requests to power-efficient servers

  5. New Challenges • Decide not only how many but also which cluster servers should be used to process current requests, when necessary • Identifying the optimal load distribution for a heterogeneous cluster is a non-trivial task

  6. Related Work • PM in homogeneous systems • [Bianchini et al. 2004], [Bohrer et al. 2002], [Chase et al. 2001], [Chen et al. 2005], [Elnozahy et al. 2002], [Rajamani et al. 2003] • PM in heterogeneous systems • [Heath et al. PPoPP2005], [Rusu et al. RTAS2006]

  7. Related Work • Current PM approaches for heterogeneous clusters • Search-based algorithms • Extensive performance measurements • Long optimization process •  high customization costs upon new installations, server failures, cluster upgrades or other changes

  8. Goal and Components • Goal • Near-optimal power consumption • QoS (average response time guarantee) • Efficient algorithm • Three components • Vary-on/off • DVS with feedback control • Optimal workload distribution

  9. System Model 1.CPU-bounded server clusters (e.g. web server cluster) 2.One front-end server 3.N heterogeneous back- end servers

  10. Optimization Problem Cast the PM to an optimization problem • Objective: Minimize the total cluster power consumption J • QoS constraints: • Decisions on • Which servers should be used to process the current workload cluster , i.e., decide xi: 0 or 1 • How should the workload clusterbe distributed to active back-end servers, i.e., decide λi such that • According to i, back-end server set its CPU frequency fi

  11. Power and Capacity Models : The ith server’s throughput : The ith server’s performance coefficient • Power Model • Capacity Model : Total power consumption : The ith server’s on/off state : The ith server’s constant power consumption : The ith server’s operating frequency : The ith server’s dynamic power consumption

  12. Optimization Problem • According to the M/M/1 queuing model and our server capacity model, we have • To make , we know

  13. Optimization Problem • The optimization problem is formed as follows • Minimize: Subject to:

  14. Optimization Problem • No analytical method to get the closed-form solution on i and xi • Time complexity of search-based algorithm • Basic idea of our efficient PM • Use a heuristic method to decouple decisions on xiand i, then solve them separately to obtain near-optimal solutions.

  15. Threshold-Based Approach • An efficient PM heuristic • Efficient offline analysis: • Divides the possible workload range into N sub-ranges • For each sub-range, the PM decisions are derived offline • Online execution: Periodically, • Front-end server: workload clusteris predicted and depending on the range cluster falls into, the corresponding PM decisions will be followed • Back-end server: applies DVS mechanism to decide fi

  16. Offline Analysis • Order the heterogeneous back-end servers, i.e., generates a sequence, called ordered server list • Produce server activation thresholds 1, 2, … N such that if cluster  (k-1, k], it is optimal to turn on the first k servers of the ordered server list • Optimal workload distribution problem is solved for the N scenarios where cluster  (k-1, k], k=1, 2, …, N (time complexity: (N))

  17. Offline Analysis • When cluster  (k-1, k], the first k servers of the ordered server list are turned on and the optimization problem becomes • Minimize: Subject to: Solution: the optimal workload distribution i

  18. Algorithm • Our method, denoted as TP-CP-OP • Server Ordered ListOrder all back-end servers according to their Typical Power (TP) efficiencies • Server Activation Thresholds Consider both server Capacity constraints and Power efficiencies (CP) • Optimal Workload Distribution (OP)

  19. Dynamic Voltage Scaling Feedforward M/M/1 Based Controller i fi errori Feedback PI Controller fi ith Back-end Server + + - Ri

  20. Evaluation • A small cluster with 4 back-end servers • Continuous operating frequency ranged in (0, fi_max] • Discrete operating frequency levels in [fi_min , fi_max] • A large cluster with 128 back-end servers in 8 different types

  21. Evaluation • Synthetic workload and Real Workload • Desired average response time is set at 1s • Evaluation metrics: average response time and power consumption • Each simulation lasts 3000s • Power management in every 30s

  22. Evaluation • Baseline algorithms • Threshold-based approaches: AA−AA−CA, SP−CA−CA, EE-RT-HSC • Optimal power management solution OPT-SOLN obtained by a search-based algorithm

  23. Evaluation • Average Response Time

  24. Evaluation • Power Consumption

  25. Conclusion • A efficient power management algorithm for heterogeneous server clusters • Mathematical models based • Minimum performance profiling • Workload threshold based • Low algorithm time complexity • Balance overhead and optimal solution • Fewer number of server on/off changes • Near-optimal power consumption

  26. TechnicalReport • L. Wang and Y. Lu. Efficient power management of heterogeneous soft real-time clusters. Technical Report TR-UNL-CSE-2008-0004, University of Nebraska-Lincoln, 2008

  27. Questions or Comments? ? Thanks! Leping Wang, Ying Lu

  28. Evaluation Effect of Feedback Control

  29. Evaluation Effect of Feedback Control

More Related