Using Parallel database systems we can overcome some problems in the conventional DBMS. The main problems in conventional DBMS are disk access time is high and it does not support the very large databases within a single system. The ultimate solution to overcome these problems is parallel database systems. The I/O bandwidth can be increased through parallelism and huge databases can be stored in a single system.
The parallel database system is a DBMS which is implemented on a parallel computer which is made of number of processors and memories connected by a fast network within a cabinet. For data management PDMS aims to achieve the modern microprocessor architectures using software oriented solutions.
The advantages of parallel database systems are high performance, high availability, and extensibility. Through inter-query parallelism and intra-query parallelisms the throughput is high and response time is less so that the performance is high. Data is available due to data replication.
The architecture of parallel dbms consists of Shared memory, shared disk, and shared nothing. In this we have some advantages and problems also. The advantages of shared memory are simplicity and load balancing and problems in that are cost, low availability and limited extensible like this we have the advantages
The techniques used in parallel DBMS:
The techniques that are used in the parallel DBMS are partitioning or fragmentation and query parallelism. In fragmentation 3 strategies are used those are round robin, hashing and range partitioning. The advantages of fragmentation are increased system performance, maximize through and minimize response time. The problem in fragmentation is the data is segmented non uniformly. We have two types of query parallelisms- intra and inter query parallelisms.
Download Parallel Database Systems PPT Paper Presentation.