Proactive Caching and Real-Time Updates - Microsoft SQL Server 2008 Analysis Services - page 438

. ForceRebuildInterval = -1 (infinite)

In this scenario, Analysis Services doesn't react to a stream of updates separated by an
interval of fewer than 10 seconds. Nonetheless, it will start reprocessing the cache 10
minutes after it receives the first notification. This scenario works well when batch
updates are pushed by the ETL process into your data warehouse.

Medium-Latency MOLAP Scenario

The values for properties in this scenario are as follows:

. SilenceInterval = 10 sec

. SilenceOverrideInterval = 10 min

. Latency = 4 hours

. ForceRebuildInterval = -1 (infinite)

This scenario is similar to the automatic MOLAP scenario, but after four hours the MOLAP
cache expires and the object switches to the ROLAP mode.

Low-Latency MOLAP Scenario
The values for properties in this scenario are as follows:

. SilenceInterval = 10 sec

. SilenceOverrideInterval = 10 min

. Latency = 30 minutes

. ForceRebuildInterval = -1 (infinite)

This scenario forces switching to ROLAP mode after 30 minutes. If the processing opera-
tion takes longer than 30 minutes, users will see a drop in performance.

Real-Time HOLAP Scenario

Real-time HOLAP (hybrid online analytical processing) scenarios are valid for setting
proactive caching only to partitions, not to dimensions. The values for properties in this
scenario are as follows:

. SilenceInterval = 0 sec

. SilenceOverrideInterval = N/A

. Latency = 0 sec

. ForceRebuildInterval = -1 (infinite)

In this scenario, Analysis Services sets the storage mode for the partition as HOLAP. The
partition data is in ROLAP mode, and the aggregations are in MOLAP mode. MOLAP
aggregations are rebuilt immediately after a notification arrives because the
SilenceInterval property is set to 0 .