Distributed DBMS - Distribution Transparency

Distributed DBMS - Distribution Transparency

Transparency in distribution is the property of distributed databases from which users are hidden from the internal information of the distribution. The designer of the DDBMS may choose to fragment tables, replicate the fragments, and store them at different sites. However, because users are unaware of this information, like any centralized database, they find the distributed database simple to use.

The three aspects of transparency in distribution are –

• Location transparency
• Fragmentation transparency
• Replication transparency

Location Transparency

Location transparency means that any table(s) or fragment(s) of a table can be query by the user as if they were locally stored on the user's web. The fact that the table or its fragments are stored in a distributed database system on a remote site should be completely ignored by the end-user. The address and the access mechanisms of the remote site(s) are completely hidden.

DDBMS should have access to a modified and accurate data dictionary and a DDBMS directory containing details of data locations in order to implement location transparency.

Fragmentation Transparency

The transparency of fragmentation allows users to question any table as if it were unfragmented. Thus, it hides the fact that a fragment or union of some fragments is actually the table on which the user is querying. It also conceals the fact that at various sites the fragments are found.

This is very similar to SQL view users, where the user does not know they're using a table view instead of the table itself.

Replication Transparency

The transparency of replication ensures that database replication is concealed from the users. This allows users to query a table as if there is just a single copy of the table.

Transparency of replication is associated with transparency of concurrency and transparency of failure. The update is reflected in all the copies of the table if a user changes a data item. This activity should, however, not be known to the user. This is transparency for concurrency. Also, the user can still continue with his queries using replicated copies without any knowledge of failure in the event of a site failure. This is transparency for failure.

Combination of Transparencies

The designer should ensure in any distributed database system that all the stated transparencies are maintained to a substantial extent. The designer can choose to fragment, duplicate and store tables at different sites; the end user is not aware of all of them. Complete transparency of distribution, however, is a difficult task and requires considerable design efforts.