1 / 11

ICS 421 Spring 2010 Parallel & Distributed Databases 2

ICS 421 Spring 2010 Parallel & Distributed Databases 2. Asst. Prof. Lipyeow Lim Information & Computer Science Department University of Hawaii at Manoa. Shared-Nothing Architecture. CPU. CPU. CPU. CPU. Memory. Memory. Memory. Memory. Disk. Disk. Disk. Disk. Network.

gordy
Download Presentation

ICS 421 Spring 2010 Parallel & Distributed Databases 2

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. ICS 421 Spring 2010Parallel & Distributed Databases 2 Asst. Prof. Lipyeow Lim Information & Computer Science Department University of Hawaii at Manoa Lipyeow Lim -- University of Hawaii at Manoa

  2. Shared-Nothing Architecture CPU CPU CPU CPU Memory Memory Memory Memory Disk Disk Disk Disk Network Lipyeow Lim -- University of Hawaii at Manoa

  3. Logical Parallel DBMS Architecture • results • query Parallel DB layer Parallel DB layer • Catalog DB • DBMS • DBMS • DBMS • Data Fragments • Data Fragments Network Lipyeow Lim -- University of Hawaii at Manoa

  4. Horizontal Fragmentation: Range Partition Partition 1 Partition 2 • Range Partition on rating column • Partition 1: 0 <= rating < 5 • Partition 2: 5 <= rating <= 10 Lipyeow Lim -- University of Hawaii at Manoa

  5. Range Partition: Query Processing Partition 1 • Which partitions? • Better than non-parallel ? SELECT* FROMSailors S SELECT* FROMSailors S WHERE rating = 2 Partition 2 SELECT* FROMSailors S WHERE age > 30 SELECT* FROMSailors S WHERE rating < 2 and age < 30 Lipyeow Lim -- University of Hawaii at Manoa

  6. Horizontal Fragmentation: Hash Partition Partition 1 • Hash partitioning using hash function • Partition = rating mod 2 Partition 2 Lipyeow Lim -- University of Hawaii at Manoa

  7. Hash Partition: Query Processing Partition 1 • Which partitions? • Better than non-parallel ? SELECT* FROMSailors S SELECT* FROMSailors S WHERE rating = 2 Partition 2 SELECT* FROMSailors S WHERE age > 30 SELECT* FROMSailors S WHERE rating < 2 and age < 30 Lipyeow Lim -- University of Hawaii at Manoa

  8. Vertical Fragmentation/Partition Partition 1 • Vertical partitioning • Use sid as row identifier Partition 2 Lipyeow Lim -- University of Hawaii at Manoa

  9. Vertical Partition: Query Processing Partition 1 • Which partitions? • Better than non-parallel ? SELECT* FROMSailors S Partition 2 SELECTsname FROMSailors S SELECT* FROMSailors S WHERE rating = 2 SELECTsid FROMSailors S WHERE age > 30 SELECTsid FROMSailors S WHERE rating < 2 and age < 30 Lipyeow Lim -- University of Hawaii at Manoa

  10. Fragmentation & Replication • Suppose table is fragmented into 4 partitions on 4 nodes • Replication stores another partition on each node • What happens when 1 node fails ? 2 nodes ? • What happens when a row needs to be updated ? • Node 1 • Node 2 • Node 3 • Node 4 P4 P1 P2 P3 Network P1 P2 P3 P4 Lipyeow Lim -- University of Hawaii at Manoa

  11. What about joins ? Partition 1 SELECTR.sid, R.bid FROMSailors S, Reserves R WHERE S.sid=R.sid AND rating > 8 Sailors Reserves • Sailors: hash • part = rating mod 2 • Reserves: hash • part = sid mod 2 • Where to perform join ? • What data to ship ? Partition 2 Sailors Reserves Lipyeow Lim -- University of Hawaii at Manoa

More Related