21.12.2013 Views

READIT - 2009 - Indira Gandhi Centre for Atomic Research

READIT - 2009 - Indira Gandhi Centre for Atomic Research

READIT - 2009 - Indira Gandhi Centre for Atomic Research

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Symmetric Multi Processing (SMP) Servers and Clusters, come under highly<br />

available and highly reliable systems <strong>for</strong> fewer prices. SMP provides the parallel processing<br />

mechanism with shared disk space and shared memory concept. Each processor reaches the<br />

shared memory through a common bus. Communication between the processors occurs<br />

through common memory. Disk controllers are accessible to all processors. The knowledge<br />

warehousing technique can be employed in the SMP clusters. Since a variety of users and<br />

knowledge users are involved in the library system the architecture should be capable to hold<br />

all types of parallel transactions. SMP clusters Provides high concurrency. User can run many<br />

concurrent requests with the better workload balances. SMP gives scalable per<strong>for</strong>mance to<br />

fulfill the user or analyst need.<br />

Core 1<br />

Core 2<br />

Core 1<br />

Core 2<br />

Core 1<br />

Core 2<br />

L2 Cache<br />

L2 Cache<br />

L2 Cache<br />

CPU CPU CPU<br />

Common Bus<br />

Knowledge<br />

Shared Memory<br />

Warehouse<br />

data<br />

SMP Architecture<br />

Access gateway<br />

Depositing<br />

Query<br />

Processing<br />

Downloading<br />

Knowledge warehouse <strong>for</strong> Library<br />

A knowledge staff in the library can deposit or store the exact files to the warehouse<br />

hard disk through a gateway. The files can be of different <strong>for</strong>mats like document files, video,<br />

audio and image files of different sizes. Query processing is <strong>for</strong> analysis and report<br />

generation. This is done by the management to take the necessary decision. A user or users at<br />

the desktop can download a required file. These are some of the actions which are taking<br />

place regularly in the knowledge warehousing. All these operations may occur<br />

simultaneously and the requests are sent to the CPU scheduler. The scheduler checks the<br />

availability of the free CPU and the job is allocated to that CPU. So the jobs are served in a<br />

parallel manner there will not be a big queue in the scheduler.<br />

Most of the current database software can parallelize a large number of operations.<br />

These operations include mass loading of data, full table search, queries with exclusion<br />

conditions, multiple queries, selection with distinct values, aggregation, sorting, creation of<br />

tables using sub queries, creating and rebuilding indexes, inserting rows into a table from<br />

other tables, enabling constraints, star trans<strong>for</strong>mation and so on. Remarkable list of<br />

operations can be processed in parallel with the DBMS.<br />

SMP cluster provides parallel processing which in turn improves per<strong>for</strong>mance <strong>for</strong><br />

query processing, data loading and downloading compared to the single processing. The<br />

scalability option, allows the addition of CPUs and memory modules without any change to<br />

the existing application. Fault tolerance enables the availability of database even when some

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!