View and vote on the article here: Distributed Databases
Distributed Databases| Category | | | Summary | | | Body | Distributed Database Concepts
In the last decade of computing, networking and database technologies have made rapid progress. With the advancement of internet technologies and distributed computing, the computing industry is rapidly changing its paradigm from keeping data at a central location, to distributing data across the network. This led to the conjunction of networks and data storage to give rise to distributed databases.
Distributed processing or computing as such has many advantages. The foremost being the division of labor between various processing elements or 'nodes', as they are commonly referred to. Distributed databases in particular have advantages like sharing of data, reliability, availability and speedup of query processing. But, as usual, there are two sides of the coin. Some of the disadvantages of distributed databases are higher development and maintenance costs, significant processing overhead and complexity in ensuring co-ordination between various sites.
Two issues which are to be given consideration while designing distributed databases are Replication and Fragmentation. Replication means that the system maintains several copies of a relation, each at a different site. The advantage of replication is availability of data even if a site goes down, and increased parallelism. The disadvantage of replication is the increased overhead on update queries, since this would result in multiple sites needing update. Fragmentation means a relation is divided into several fragments, each of which is stored at a different site. There are two types of fragmentation, vertical and horizontal. In horizontal fragmentation, a relation is split by assigning each tuple of the relation to one or more fragments.
Any good distributed database management system should maintain transparency with respect to the end user. This means that the implementation details of how the relation is stored actually should not be a concern of the user of the distributed database. Transparency with respect to distributed databases is called network transparency. Issues regarding network transparency are of three types,
- With respect to naming of data items
- With respect to replication
- With respect to fragmentation
References :-
Database System Concepts by Silberschatz, Korth and Sudarshan
Fundamentals of Database Systems by Elmasri and Navathe
An Introduction to Database Systems by C J Date
|
|
This article was imported from the CyberArmy University site. (original author: rae)
There are no replies to this post yet.
|