Dissertation > Excellent graduate degree dissertation topics show

The Research of Energy Conservation Mechanism for Server Clusters in Network

Author: LiuBin
Tutor: YangJian
School: University of Science and Technology of China
Course: Control Theory and Control Engineering
Keywords: Server cluster Energy-saving Load -line prediction RLS algorithm LMS algorithm
CLC: TP393.05
Type: Master's thesis
Year: 2011
Downloads: 49
Quote: 0
Read: Download Dissertation

Abstract


With the development of science and technology, the computer is an increasingly wide range of applications in various industries, many industries increasing computer performance requirements. Large-scale server clusters bring convenience for research, work, and life at the same time, it also brings a serious problem - a huge energy consumption. The energy consumption for the server cluster is a major problem, not only because it is a single or a group of server energy consumption also directly affect the cooling requirements of the system, spare equipment cooling demand, as well as the demand for standby power generation equipment. Although many local governments spent a lot of money to solve the supply problem, but basically all of the power generation technologies on the environment have huge side effects. Both from the perspective of the Internet, or from the point of view of society as a whole, the cluster energy issues have become a very real and serious problem. More mature strategy, such as the scheduling algorithm for system-level energy-saving strategy of dynamic power management (DPM), dynamic voltage and frequency adjustment strategy (DVFS), a strategy of dynamic voltage scaling (DVS). Several strategies applications in the system-level energy level was very successful, but is not suitable for use in a server cluster level. Applicable at the cluster level energy-saving strategy is the hot issues of the industry's research, the the earlier strategy appears proportional integral derivative feedback control strategy (PID) and load centralized strategy (LC). With the complexity of the structure of the cluster system, cluster-business diversification, both the effectiveness of energy-saving strategies in reducing. Dynamic cluster configuration is dynamically adjusted according to the network load server scale, the minimum system power consumption to achieve optimal service performance. This paper presents a dynamic cluster configuration strategy based on the forecast, the method according to the historical information network service requests using the minimum mean square error (LMS) and recursive least squares (RLS) to forecast the future time service request, according to the load request cluster The processing capability decision server scale of change dynamically adjust the server cluster, the computer to turn on and off. In addition, for compute-intensive server clusters cluster allocation strategy. Provide guaranteed QoS services based on the ultra-negative rate, we abstract the energy saving problem for constrained optimization problem, that maintain ultra-negative rate is below a desired threshold, the minimized active service node number. Super negative rate is estimated using a mathematical tool is a large deviation theory / close status of the opening of the service node in the cluster by this algorithm to Decision. Based on the the large deviation algorithm of the decision-making method requires only concerned about the current cluster load, load business without having to care about historical and statistical information. Another advantage of this strategy is the iterative method to adjust the work of the state of the server, rather than directly to determine the number of working state service node. Simulation experiments, we use the the Hebrew University the Parallel Workloads system of Jerusalem user trace data, the feasibility and advantages of a real network cluster users access data inspection scheduling policy.

Related Dissertations

  1. Several Problems of Energy Saving Operation in Air Conditioning System of Large Scale Public Buildings,TU831.6
  2. Research on University’s Energy Saving System Facing the Construction of Green University,G647
  3. Applied Research of Emulsified Fuel on S195 Diesel Engine,TE626.24
  4. A Research and Development for Single Chip Micyoco AVR Usage,TP368.12
  5. Location Assisted Wi-Fi Discovery on Mobile Devices for Energy Saving,TN929.53
  6. Saw motor energy-saving controller for tilting sand stone processing DSP-based research and development,TM343
  7. Research on Evaluation of Coal Enterprise Energy-Saving Mechanism,F206
  8. Study on Green Finance Promotes Industrial Restructuring,F832;F205
  9. Traffic Adaptive Energy-Saving Algorithm for Wireless Transceivers,TN859
  10. Study on the Electrical Parameter Measure System of the Motor Energy-saving Device,TM306
  11. Stainless Steel Bright Annealing Furnace with Waste Heat Recovery Energy-saving Technology Research,TG155.1
  12. Second-order Circulation Statistics-based Array Signal Processing,TN911.7
  13. On a New Type of Building Walls-photoelectric Curtain Wall,TU228
  14. The Research and Implementation of Energy-efficient RAID,TP333
  15. Research of Optimal Design of Power System in High Speed Railway,U223
  16. Study on a Bathroom Heating Plan of Utilizing Renewable Energy Source,TU833
  17. Sustainable Construction Design Analysis in Macau,TU201.5
  18. Phase-coded signal pulse technology research and implementation,TN957.51
  19. 14-bit 100 MB samples / sec pipelined ADC Research and Design,TN792
  20. A Study on the Green Building Evaluation System,F205
  21. Research on Control System of Hot-reclaimed Asphalt Pavement Maintenance Vehicle Based on PLC,U418.6

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Computer network > General issues > Network equipment
© 2012 www.DissertationTopic.Net  Mobile