1 / 20

Query Evaluation Techniques for Cluster Database Systems

Query Evaluation Techniques for Cluster Database Systems. Andrey V. Lepikhov , Leonid B. Sokolinsky South Ural State University Russia. Outline. Motivation Problem Statement Background Partial mirroring method Results Future work. Motivation. Top500 Cluster: 84.8% MPP : 14.8%

iona-cleveland
Download Presentation

Query Evaluation Techniques for Cluster Database Systems

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. ADBIS 2010 Query Evaluation Techniques for Cluster Database Systems Andrey V. Lepikhov, Leonid B. Sokolinsky South Ural State University Russia

  2. Outline • Motivation • Problem Statement • Background • Partial mirroring method • Results • Future work ADBIS 2010

  3. Motivation • Top500 • Cluster: 84.8% • MPP: 14.8% • Others: 0.4% ADBIS 2010

  4. Problem Statement • Not expensive parallel hardware needs not expensive parallel database management system • Today we have no such chip parallel database management system ADBIS 2010

  5. Background ADBIS 2010

  6. Exchange operator • p: port • ψ: distributing function ADBIS 2010

  7. Parallel plan for query Q = RS ADBIS 2010

  8. Query processing in cluster system ADBIS 2010

  9. The problem • Load balancing ADBIS 2010

  10. Partial mirroring method • Fragmentation strategy • Replicationstrategy

  11. Fragmentation strategy Source relation • Relation is divided into fragments distributed among cluster nodes • Each fragment is divided into sequence of segments with an equal length • Segment is the minimal unit of replication P0 . . . FRAGMENTATION P1 . . . . . . . . . Pn ADBIS 2010

  12. Replicationstrategy Ri ρj= 50% DiskDj DiskDi ADBIS 2010

  13. Load balancing method D1 D2 D1 D2 -scheduled to process - processed P1 -not used P1 P2 P2 t2 t1 D1 D2 D1 D2 P1 P2 P1 P2 t3 t4 ADBIS 2010

  14. Parallel agent with two input streams ADBIS 2010

  15. Load balancing algorithm ADBIS 2010

  16. Parameters of experiments ADBIS 2010

  17. Speedup versus replication factor ADBIS 2010

  18. Speedup versus skew factor θ • 0.68 corresponds to the ”80-20” rule (80 percents of tuples of the relation will be stored in 20 percents of fragments) • 0 corresponds to the uniform distribution ADBIS 2010

  19. Future Work • To incorporate the proposed technique of parallel query execution into open source PostgreSQL DBMS. • To extend this approach on GRID DBMS for clusters with multicor processors. ADBIS 2010

  20. Thank you ADBIS 2010

More Related