Protecting Production Databases

Sybase Data Federation is designed to help database administrators insulate their production databases from risk in several ways, including: providing access to data through stored procedures; limiting access to all but specific, predefined queries; caching results; enabling access through data services and generated views; and by using two-phase commit for distributed transactions.

Because users and applications have access to the data from the original source, you might be concerned about how this will affect production systems. After all, database administrators need to ensure that production applications they support will perform well. Data Federation is specifically designed to help database administrators insulate their production databases from risk in several ways:

By providing access to data through stored procedures.
Stored procedures are totally under the control of the DBA and have been optimized for the database.
By limiting access to all but specific, predefined queries.
Database administrators provide the queries used by users and application developers, and can define and test queries that will perform efficiently before making them available.
By caching results.
Access to relational data does not have to be dynamic as long as users and developers can retrieve reasonably fresh data. Data Federation enables database administrators to cache results—either on-demand or on a pre-determined schedule—so users never need direct access to the database. Results can be cached as frequently as required to meet the needs of data consumers. Caching has no impact on applications; no special coding is involved in using cached data.
By enabling access through data services and generated views.
When users need data in nonrelational forms, DBAs can provide generated views or data services, then grant access to those objects. The views or data services can be updated on a schedule, or updated when the database operation is updated.
By using two-phase commit for distributed transactions.
A developer can build a data service that uses the two-phase commit protocol to execute a group of database operations as a single transaction. This ensures that related operations that perform work on multiple independent databases will leave all the databases in a consistent state.