Designing the Data Structure › Choose to Distribute or Replicate Data › Distributed Data Alternative

Distributed Data Alternative

Distribution of data is the division of data between a number of servers.

The complexity of the division can range from simply placing different types of data (tables, or even individual columns) on separate servers to partitioning data of the same type between servers according to where it is created or needed.

The following list explains advantages of distributing data are:

• Brings complex or iterative maintenance of data closer to the user
• Provides for faster retrieval through dedicated database processors

The following list explains disadvantages of distributing data are:

• Can create extensive network traffic
• May not always be accessible as quickly as needed

Data can be distributed by using a database management system to automatically take control of the process.

Note: Controlling the distribution of data that is outside the scope of the DBMS can be achieved by specifying locations in procedure logic. Remember to consider the risks of failing to maintain data consistency, integrity, and security. This approach should be considered only if you are familiar with these issues, as well as the techniques for controlling distributed data.

Mechanisms distributing data depend on the currency requirement for the data:

• If the solution demands immediate, up-to-date data, a database management system should be used.
• If, however, the solution is a decision support system and requires data distribution on a monthly basis, it may be quite safe for an application or personal productivity software (such as a spreadsheet) to maintain the data locally.