The Unwired Server cache is the replicated data store component of the cache database (CDB) and is the integration point for synchronization and data refresh. It manages synchronized data between Unwired Server and device applications, and refreshed data between Unwired Server and the enterprise information system (EIS). Administrators can control cache behaviours in Sybase Control Center.
The cache performs a number of important functions, including:
- Maintaining a local copy of enterprise data.
-
Managing updates between the CDB and the EIS servers (data refresh).
- Managing updates between CDB and device application data (synchronization), even in environments where there are thousands of simultaneous synchronizations.
- Partitioning of data – partitions for MBO data, for example, based on client parameters. When a device application passes a client parameter used for both synchronization and data refresh, the CDB:
- Tracks rows under different partitions based on the synchronization parameter values. A synchronization parameter maps to an attribute that acts as a filter or variable that lets you limit the data that is returned to the device to rows in the table based on a supplied value.
- Keeps track of which partitions each client is interested in from prior synchronizations. For example, Unwired Server knows that client one is only interested in rows containing the "ABC" parameter value, while client two cares only about rows that contain "def".
Many of the complexities of maintaining synchronization and data refresh are transparent to the administrator, however, you can:
- Configure a dedicated CDB to run within a cluster. A cluster can have any number of Unwired Servers, but only one CDB. If you have a dedicated CDB within a cluster, it should be the first node of the cluster.
- Increase the number of worker threads dedicated to the CDB as the number of Unwired Server instances in a cluster increases.
- Modify CDB port number.
- Monitor synchronization and data refresh performance.