Setting Custom Storage Settings
Instead of using one of the standard storage settings, you can manually configure storage and proactive caching. Before you create custom storage settings, you may want to first click the Standard settings option and move the slider to the standard setting that most closely matches the configuration you want to use. Then, to create a custom configuration, click the Custom settings option and click Options.
Custom Storage Options
- You can specify whether changes in a data source trigger updates to the cache. To allow for a tolerable level of churn, you can specify a minimum silence interval after updates to the data source. You can also specify a silence interval override that updates the cache after a specified period if the interval between changes to the data source never reaches the minimum.
- You can specify whether to drop the outdated cache when updates occur. If you select this option, then when the specified latency is exceeded, the server switches to real-time relational OLAP (ROLAP) while it updates the cache. If you do not select this option, the server continues to query the stale multidimensional OLAP (MOLAP) cache while it builds the new one.
You can specify the latency interval that must occur between changes and dropping an outdated cache. This is the amount of time users may continue browsing data in an outdated cache before it is dropped. If changes occur and the cache is still being updated or processed at the end of this interval, queries will be redirected to ROLAP. - You can schedule forced updates of the cache if you want to periodically update the cached MOLAP objects regardless of changes to the data source. Real-time OLAP benefits vary with the size of your database and the latency period assigned by frequency of source data changes. You want users sending queries to a cache as often as possible, not to ROLAP.
If you select the Apply settings to dimensions check box, the same storage settings are applied to the dimensions related to the measure group. The dimension values are initially the same as the partition values.
See Also
Concepts
Partition Storage Modes and Processing
Designing Partition Storage and Aggregations