Databases Reference
In-Depth Information
Client
Application
Data
Warehouse
Analysis
Services
OLTP System
ETL
Data
Load
Client
Application
FIGURE 24.1
OLTP data goes through ETL and the data warehouse before it is loaded into
Analysis Services.
Data Latency and Proactive Caching
Proactive caching is the part of the infrastructure of Analysis Services that enables you to
define rules to detect changes in the relational database and rules for automatic updates of
objects when Analysis Services is notified that a change has been received. To implement a
low-latency system without built-in server functionality, you must implement custom
logic that detects changes in the data warehouse and sends processing commands to
update Analysis Services objects. Such an application triggers processing of Analysis
Services objects every time new data comes into the relational database. That requires
unnecessary processing and consumes a lot of Analysis Services resources.
Analysis Services features proactive caching technology to implement a low-latency
system. Internally, the implementation of proactive caching can be seen as the retention
of two versions of each object: the current one and a new one (a multidimensional online
analytical processing [MOLAP] cache). After the new version has been processed, it
becomes the current one. See the example in Figure 24.2.
Analysis Services
Data Warehouse
Dimension1
Notification
Notification
Data
Notification
Data
New Version of
Dimension1
Dimension
Ta b l e
Data
Changes
Partition
Ta b l e
Notification
Notification
Data
Notification
Data
Partition1
Data
New Version of
Parition1
FIGURE 24.2
Proactive caching receives notifications and builds new version of dimensions
and partitions.
Search WWH ::
Custom Search