1 / 32

Efficient Min-Max Graph Partitioning for Cloud Computing: A Novel Approach

This study delves into the problem of graph partitioning for cloud computing, focusing on communicating processes and bandwidth allocation among machines. Previous work in the field is discussed, along with related problems. The results showcase a good partition strategy utilizing disjoint covers with "good" sets, ensuring each vertex is covered adequately. The outline includes an introduction to the topic, details on the approach taken, and various covering methods for optimal graph partitioning. The study breaks down the challenges of min-max partitioning and emphasizes the ineffectiveness of any k-partition containing a specific vertex. The configuration, LP, and SSE approaches are also analyzed for effective partitioning. The study concludes with remarks on the effectiveness of different covering strategies in graph partitioning.

swillingham
Download Presentation

Efficient Min-Max Graph Partitioning for Cloud Computing: A Novel Approach

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. Min-Max Graph Partitioning

  2. The Problem

  3. . . . k 1 Motivation Cloud Computing n communicating processes Bandwidth B machines

  4. Previous Work

  5. A Related Problem

  6. A Related Problem

  7. Our Results

  8. Our Results Good Partition: Disjoint Cover with “good” sets (SSE’s)

  9. Our Results Good Partition: Disjoint Cover with “good” sets (SSE’s) LP: Each vertex covered to extent 1

  10. Our Results

  11. . . . k 1 Our Results Machines

  12. Outline • Introduction • Graph Partitioning (quick intro) • Our Approach • Coverings to Partition

  13. Graph Partitioning Approaches 0 1 (all vertices sit here)

  14. Graph Partitioning Approaches 0 1 (all vertices sit here)

  15. . . . k 1 Breaks down for min-max 0 Any k-partitioning is bad: Part containing vertex 0 LP/SDP can always cheat (by smearing vertex 0) (even with triangle inequalities, all kinds of separating constraints)

  16. Outline • Introduction • Graph Partitioning (quick intro) • Our Approach • Coverings to Partition

  17. Configuration LP

  18. SSE

  19. Configuration LP

  20. Covering to Partition

  21. Covering to Partition

  22. Covering to Partition

  23. Covering to Partition

  24. Covering to Partition

  25. Covering to Partition

  26. Covering to Partition

  27. Covering to Partition

  28. Covering to Partition

  29. Covering to Partition

  30. Covering to Partition

  31. Concluding Remarks C B

  32. Thank you

More Related