Database Reference
In-Depth Information
a sliding bar user interface to choose between the different storage settings. Each of these categories provides a
set of predefined settings that can be customized even further by selecting the Custom Settings radio button and
clicking the Options button at the bottom of the dialog window.
Once you have clicked the Options button, a Storage Options dialog window appears, as shown in Figure
12-18
.
In this dialog window, you can choose between using MOLAP, HOLAP, or ROLAP using the “Storage mode”
dropdown box. This has the same effect as using the slider bar in the previous dialog box.
In the default configuration, each partition is set to the storage mode of OLAP without the proactive cache
option enabled.
Selecting “Enable proactive cache” checkbox (circled in Figure
12-18
) provides a way of updating data between
the data warehouse and the partition in an automated manner. For example, if you look at the settings in Figure
12-18
,
we have checked the box that allows SSAS to update the cache whenever data changes within the data warehouse.
Figure 12-18.
The Storage Options dialog window
The “Silence interval” option indicates the time SSAS should wait before it attempts to process the partition
after a transaction has been sensed within the data warehouse tables. The default is ten seconds. This means that
SSAS waits for ten seconds of no transaction activity before it attempts to process new data into your SSAS partition.
If transactions are happening so rapidly that a ten-second pause is not occurring, the “Silence override
interval” setting will force processing to begin after ten minutes (while in the default configuration).
Automated updates may seem like a pretty good idea. And in fact they are, if they are set up correctly.
But, there are a few things that need to be considered. Notification that the data has changed within the data
warehouse is done in one of three ways:
•
Setting up a job to poll for changes
•
Creating an application that polls for changes
•
Creating database triggers
For many developers, none of these options seem all that attractive. Because of this, many companies do not
enable the proactive caching with automatic updates, but instead rely on scheduled processing that occurs each
night or several times a day. In these cases, employees are notified that the report data is “up-to-date” based on
