Dissertation > Excellent graduate degree dissertation topics show

Research on Optimal Scheduling of Jobs Approach in Cloud

Author: HuiXinHua
Tutor: SongBaoYan
School: Liaoning University
Course: Computer Software and Theory
Keywords: Cloud Computing job scheduling hadoop MapReduce Multi-layer management
CLC: TP3
Type: Master's thesis
Year: 2011
Downloads: 313
Quote: 2
Read: Download Dissertation

Abstract


Cloud computing is a new business computing model, so far, Google, Yahoo, Microsoft, IBM, Amazon and some other famous cooperation has proposed their cloud computing application, and take development of cloud computing as the one of the most important strategies in the future. Cloud computing is the development of parallel computing, grid computing, distributed computing and which provide for the needs of users data resources and computing services on the Internet. The influence of customer requirements and resource distribution for the job scheduling is very large in cloud environment. Quality of Service for users is paid more attention in commerciality of cloud computing。In this paper, customer requirements were classified through customer’s preferences. The basic classifications of job include several aspects, such as completion time of operation, bandwidth, cost, stability of demand and other aspects. Meanwhile, monitor the nodes in the cloud environment, inspection node resource properties, such as distance from the node to the data resources required, CPU, memory, cost, bandwidth, failure rate, etc. Classification of jobs by users and resources for evaluating the results of the node are mapping. According to the results of mapping the node select the appropriate implementation of operations as Manager.In some cloud environments, such as hadoop, take use of MapReduce mode scheduling, often using the Master/Worker level management model. Master node handled so much the relatively heavy work that the number of jobs requst was very limited。In the paper, we add Manager node into the system that we implement the Master/Manager/Worker three level management model for improving system throughput and improving the efficiency of the cloud environment management model .we proposed method of MultiLe(multi-management scheduling optimization)and support MultiLe’s architecture.There are three roles in the MapReduce framework, such as user, Master, Worker and so on, and we add another roles Manager.Users submit jobs to the Master node.Through the needs of user for jobs and the information of node resource, Master node map both of them to access to the node Manager dynamically. Master node assigned the job to the Manager node.Manager splits the job into map and reduce tasks, assigned the tasks to the Worker to perform . Then split into Map Manager and Reduce Operating tasks. Although the system complexity increases by setting Manager, which can better meet the needs of business and the load of Master node was reduced effectively.Cloud computing system so as to enhance the throughput and improve the efficiency of the cloud environment.

Related Dissertations

  1. The Research of Dynamic Trust Model on Cloud Computing Platform,TP309
  2. Several studies for scheduling problem,O157.5
  3. Cloud-based digital library service model,G250.76
  4. Research and Application of Map/Reduce Based Distributed Log Analyzer,TP311.52
  5. Design and Implementation of Online Shopping Prototype System Based on Hadoop,TP311.52
  6. Cloud computing technology and its application in e-commerce logistics center design and operation of the application,F724.6;F252
  7. Design and Implementation based the Google platform promotional modules commodity module,TP311.52
  8. Research and Implementation of job scheduling algorithm in the Linux cluster environment,TP301.6
  9. The Operation and Maintenance of ITIL Based on Cloud Computing,TP311.52
  10. Web application system design based on Google's cloud computing platform and,TP393.09
  11. An Intrusion Detection System for High-Speed Networks,TP393.08
  12. Research and Implementation on a Distributed Service Registry Based on HADOOP Platform,TP393.09
  13. Load Balance Approach to Save Power on Cloud Datacenter,TP308
  14. Resarch of Task-level Data Processing Based on Multicore CPU and Test of Its Performance on Cluster Platform,TP274
  15. Linux-based platform LAN Cloud Monitoring System Analysis and Implementation,TP311.52
  16. New Container Terminal mixed loading optimization scheduling strategy,U691.3
  17. Aircraft assembly moving assembly line job scheduling optimization,V262.43
  18. Cloud-based mobile data storage backup system,TP309.3
  19. Cloud data storage technology research homeostasis,TP333
  20. Cloud-based computing and immune reconfigurable embedded systems research network,TP368.1
  21. Cloud storage system for mass data,TP333

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology
© 2012 www.DissertationTopic.Net  Mobile