Copy Of Parallel Database Systems

Technical Seminar Presentation

Parallel Database Systems

PARALLEL DATABASE SYSTEMS Under the guidance of :

Mr. Anisur Rahman

AMRITA MISHRA

CS200198086

1

Parallel Database Systems

Technical Seminar Presentation

INTRODUCTION A parallel database system (PDBS) is a DBMS implemented on a parallel computer which is made of a number of nodes (processors and memories) connected by a fast network within a cabinet. It strives to exploit modern multiprocessor architectures using software-oriented solutions for data management. AMRITA MISHRA

CS200198086

2

Parallel Database Systems

Technical Seminar Presentation

OBJECTIVE Problems of conventional DBMS - high disk access time. - very large databases cant be supported within a single system. PDBS is the only viable solution for increasing the I/O bandwidth through parallelism & for storing huge databases in a single system. AMRITA MISHRA

CS200198086

3

Parallel Database Systems

Technical Seminar Presentation

ADVANTAGES OF PDBS High Performance – Increased throughput (inter-query parallelism) & decreased response time (intra-query parallelism). High Availability – Using data replication. Extensibility – Linear scaleup and Linear speedup.

AMRITA MISHRA

CS200198086

4

Parallel Database Systems

Technical Seminar Presentation

PARALLEL DBMS ARCHITECTURE Shared Memory  Advantages – Simplicity, Load Balancing.  Problems – Cost, Limited Extensibility, Low Availability.

AMRITA MISHRA

CS200198086

5

Parallel Database Systems

Technical Seminar Presentation

Shared Disk Advantages – Cost, Extensibility, Load Balancing, Availability. Problems – Higher Complexity, Potential Coherence Problems.

AMRITA MISHRA

CS200198086

6

Parallel Database Systems

Technical Seminar Presentation

Shared Nothing Advantages – Cost, Extensibility, Availability. Problems – Complex, Addition of new nodes requires reorganizing the database.

AMRITA MISHRA

CS200198086

7

Parallel Database Systems

Technical Seminar Presentation

PARALLEL DBMS TECHNIQUES DATA ALLOCATION – Methods that spread the database across the system’s disks to ensure efficient parallel I/O.  Partitioning (Fragmentation) – 3 strategies

# Round Robin – i th tuple to partition (i mod n) for n partitions. # Hashing – Apply hash function to some attribute to give partition no. # Range Partitioning – Distribute tuples based on value(ranges) of some attribute. AMRITA MISHRA

CS200198086

8

Parallel Database Systems el

Technical Seminar Presentation

USES OF DATA FRAGMENTATION  Maximize system performance. Minimize response time (through intra-query parallelism). Maximize throughput (through inter-query parallelism).

Problems: Skewed data distributions lead to non-uniform partitioning & hurt load balancing. AMRITA MISHRA

CS200198086

9

Parallel Database Systems

Technical Seminar Presentation

QUERY PARALLELISM Inter query parallelism – parallel execution of multiple queries generated by concurrent transactions.  Intra query parallelism – Two kinds: #Inter-operator – Executing several operators of the query tree on several processors in parallel. #Intra-operator – Same operator is executed by many processors, each one working on a subset of the data.

AMRITA MISHRA

CS200198086

10

Parallel Database Systems

Technical Seminar Presentation

PARALLEL ALGORITHMS

AMRITA MISHRA

CS200198086

11

Technical Seminar Presentation

Parallel Database Systems

AMRITA MISHRA

CS200198086

12

Technical Seminar Presentation

Parallel Database Systems

AMRITA MISHRA

CS200198086

13

Parallel Database Systems

Technical Seminar Presentation

PARALLEL JOIN EXAMPLE Insert into C Select * from A , B where A.x = B.y ;

AMRITA MISHRA

CS200198086

14

Technical Seminar Presentation

Parallel Database Systems

AMRITA MISHRA

CS200198086

15

Parallel Database Systems

Technical Seminar Presentation

CONCLUSION  The availability of fast microprocessors & small inexpensive disks is an ideal platform for parallel database systems.  Employing data partitioning, intra-operator parallelism etc., converting an existing DBMS into a highly parallel one is relatively straightforward.

AMRITA MISHRA

CS200198086

16

ational Institute of Science & Technology

Technical Seminar Presentation

Parallel Database Systems

THANK YOU!!!

AMRITA MISHRA CS200198086 17