0% found this document useful (0 votes)
46 views

Parallel System

This document discusses different architectures for parallel database systems. It describes the limitations of a shared memory architecture, including that memory access becomes a bottleneck as more CPUs are added. It then describes a shared disk multiple CPU architecture, where each CPU has its own memory but can access a shared set of disks. A locking scheme is required to ensure data integrity when multiple CPUs want to access the same data. Key advantages are improved load balancing, scalability and fault tolerance compared to shared memory architectures.

Uploaded by

Chetan Kumavat
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
46 views

Parallel System

This document discusses different architectures for parallel database systems. It describes the limitations of a shared memory architecture, including that memory access becomes a bottleneck as more CPUs are added. It then describes a shared disk multiple CPU architecture, where each CPU has its own memory but can access a shared set of disks. A locking scheme is required to ensure data integrity when multiple CPUs want to access the same data. Key advantages are improved load balancing, scalability and fault tolerance compared to shared memory architectures.

Uploaded by

Chetan Kumavat
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
You are on page 1/ 11

Parallel Database System

Shared Memory

Limitations of shared memory Architecture: Memory access uses a very high-speed mechanism that is difficult to partition without losing efficiency .Thus the design must take special precautions that the different CPUs have equal access to the common memory . Also ,the data retrieved by one CPU should not be unexpectedly modified by another CPU acting in parallel . Since the communication bus or interconnection network is shared by all the CPUs ,the shared memory architecture is not scalable beyond 80 or 100 CPUs in parallel . The bus or the interconnection network becomes a bottleneck as the number of CPUs increases. The addition of more CPUs causes CPUs to spend time waiting for their turn on the bus to access memory.

Shared-disk Multiple CPU Parallel Database Architecture : ----------------------------------------------------------------------In a shared disk system , multiple CPUs are attached to an interconnection network and each CPU has its own memory but all of them have access to the same disk storage or ,more commonly ,to a shared array of disks .The scalability of the system is largely determined by the capacity and throughput of the interconnection network mechanism.Since memory is not shared among CPUs ,each node has its own copy of the operating system and the DBMS. It is possible that ,with the same data accessible to all nodes ,two or more nodes may want to read or write the same data at the same time. Therefore , a kind of global (or distributed ) locking scheme is required to ensure the preservance of data integrity . Sometimes ,the shared-disk architecture is also referred to as a parallel database system

This figure shows a schematic diagram of a shared-disk multiple CPU architecture. Advantages of Shared-disk architecture: Shared-disk architecture is easy to load-balance ,because data does not have to be permanently divided among available CPUs. Since each CPU has its own memory ,the memory bus is not bottleneck It offers a low cost solution to provide a degree of fault tolerance.In case of a CPU or memory failure, the other CPUs take over its task; since the database is resident on disks that are accessible from all CPUs It has found acceptance in wide applications.

The interconnection networks for shared nothing architecture are usually designed to be scalable .Thus adding more CPUs and more disks enables the system grow (or scale) in a manner that is proportionate to the power and capacity of the newly added components

Page 531 missing

Page 535 missing

Page 538 missing

You might also like