How to make OLAP Cube faster

How to make SSAS Cube faster

In order to make the cube much faster both on processing and querying one should define the data type as Analysis Services does not treat all the data types in the same way below is the brief description that help during the design of the data mart.

Column Type

How to make Cube Processing Faster:-

Whenever a cube is processed Analysis Services send queries to the relational database as it requires information for this processing, in this scenario Analysis Services generate some complex queries that heavily depended to our data mart model.

We should keep in mind that if we have built our dimension using complex cubes they might cause SQL server to poor query performance likewise a complex condition of where clause in the view may be the reason for this poor SQL query performance.

Joining Dimensions

In snowflake it is the common practice to resolve the many-to-many relationship that required joining between the relevant tables, if these joins are based neither on Primary keys nor there is any performance problems but if not will resulted in bad performance.

Handling Reference Dimensions:-

Snowflake schema often deals with reference dimension required to reduce the size of the tale by separating the dimension in to sub tables. Analysis Services request the SQL Sever to resolve joining of intermediated tables if joins are problem with the dimension processing then it is a serious problem because it causes the SQL Server to write large amount of data to its temporary database before providing information to Analysis Services.

Order By Clause:-

Another fact that causes poor performance is the measure group that consists of a Distinct Count measure as the Analysis Services calculate the distinct count and the query contain a order by clause, because we are expecting from SQL Server to sort the column that is not the part of clustered index, normally a clustered index is performed on primary key this will result in temporary table and the query will take lot of time to be completed. 

Difference between the Data Source view and Views:-

• In Analysis Services Data Sources View (DSV) is an area or interface that allow you to create an interface between the database and the Analysis Services, here we can manipulate joins, define named queried and calculations.
• Views can be stored according our need and requirement.
• AS views are stored in the database so any change and dependency will easily be traced while in DSV these dependencies are hidden and cannot be seen by the database administrator.
• It is Easy in views to change column name and data type.
• Easy to calculate any sort of calculations.
• As views are of plain text hence are feasible to search specific column.
• Quickly to update
• Easy to assign null values
• Although the relational model is much complex but a view can expose a star schema