The Application Server uses in-memory data caches for improved performance. By keeping data in memory and not being forced to query data from the Application Server Database every time it is requested, the Application Server can process data magnitudes faster than without caching. With time these caches will accumulate more and more data and therefore the performance of the system will be optimized at the cost of system memory.
Since the building of caches takes some time some of these caches will be persisted to disc on shutdown of the Application Server and read from disk into memory again on startup. See table 1 for a list of caches that are persisted to disc. Depending on the sizes of these caches, writing caches to disc and loading caches will take some time. When starting the Application Server it will only accept requests as soon as all persisted caches are loaded back into memory.
The path where caches persisted can be defined during the setup and is stored in the instance configuration file. We highly recommend to keep the caches on a local disc on the Application Server to speed up Application Server startup and shutdown. Since caches hold only information that is available in the Application Server Database or is calculated on data read from the database necessary to backup those persisted caches.
Cache |
Notes |
Effective permission cache |
If a permission is queried from the system (e.g. does User1 have read permission on Feature1) this query might result in complex queries to the database and therefore take some time to complete. This cache stores permissions that were evaluated once and provides fast access to permissions without complex queries to data base for future queries. |
View member cache |
This cache that stores the effective members of a view. The cache is refreshed by the Application Server based on a schedule. |
Table 1: In-memory caches the Application Server persists to disk on shutdown.
