1 / 13

Multi-tier application controller for utility computing center

Multi-tier application controller for utility computing center. CPS210 – Operating Systems class project Azbayar Demberel, Amre Shakimov, Harold Lim. Duke University 2008. Agenda. Motivation Design Implementation Evaluation Conclusion. Visitor trends in major retail web sites.

odessa
Download Presentation

Multi-tier application controller for utility computing center

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. Multi-tier application controller for utility computing center CPS210 – Operating Systems class project Azbayar Demberel, Amre Shakimov, Harold Lim Duke University 2008

  2. Agenda • Motivation • Design • Implementation • Evaluation • Conclusion

  3. Visitor trends in major retail web sites Source: http://www.akamai.com/html/technology/nui/retail/index.html MotivationDesign Implementation Evaluation Conclusion

  4. Cluster on demand: ORCA overview • A service-oriented toolkit for autonomic, on-demand computing • Substrate for servers/actors that interact to manage shared infrastructure • Factor the resource allocation policy/mechanism out of Cluster on Demand (COD) • Manages isolated virtual clusters (physical or virtual machines) within a single physical cluster • Illusion of own personal cluster, but actually sharing with other applications • Virtual clusters are dynamic; must adapt to changing conditions • Defines what it means for a cluster to add/remove machines, add/remove reservations • Resources are currently Xen Virtual Machines or Physical Machines MotivationDesignImplementation Evaluation Conclusion

  5. Orca controller design MotivationDesignImplementation Evaluation Conclusion

  6. Multi-tier controller design Web Tier Application Tier Database Tier CPU CPU CPU Tomcat controller Jboss controller Mysql controller Memory Memory Memory HQ server MotivationDesignImplementation Evaluation Conclusion

  7. Load variance on node 1 due to resource growth Number of requests served per second Motivation Design ImplementationEvaluationConclusion

  8. Memory usage: not very useful Free memory JVM free memory Motivation Design ImplementationEvaluationConclusion

  9. Discussion • Contribution • Separate policy & mechanism • Service level granularity: more extensible • Future work • Top level coordinator controller  stackable controller? • Advanced policy modules • Integrate with workload generator • Web portal plug-in Motivation Design Implementation EvaluationConclusion

  10. Thank You Questions or comments?

  11. Backup slides

  12. Poor resource utilization example 2

  13. Throughput saturation in apache web server Source: http://www.eecs.harvard.edu/~mdw/talks/seda-sosp01-talk.pdf MotivationDesign Implementation Evaluation Conclusion

More Related