ROLAP, MOLAP and HOLAP

I always see a lot of confusion among most of my students with regards to these three terminologies. Let’s have a quick talk on them.

When we create Cube in SSAS, it consist of two things – Dimensions data and partitions.
Partitions hold measure group data.
By default for every measure group one partition will be created. We can configure measure group data to be stored into multiple partitions if required. We will discuss on advantage of creating multiple partition some time later. In this small article we will focus on three important terminologies ROLAP, MOALP and HOLAP.

Dimensions in SSAS cube supports two storage modes. ROALP and MOLAP.

  • ROLAP –In this storage mode, dimensional data is actually stored in the source table only. When we try to read any dimension from cube it internally fires a query to data source and data is returned.
  • MOLAP – In this storage mode, dimension data is get copied into cube every time cube is processed. In this case data will be directly queried from our Multidimensional database.

MOLAP is good for performance but need more disk space as data is getting duplicated.

Partitions in SSAS cube supports three storage modes. ROLAP, HOLAP and MOLAP.
When it comes to partitions, it holds measure group data plus it may hold some aggregations. May times in times we go and create on or more aggregations on each partition to increase the performance. If you are not aware about Aggregations then in simple words, it’s a pre-calculated data which we generate to increase performance.

  • MOLAP – Both source data and aggregation data will be stored in cube each time cube or partition is processed.
  • ROLAP – Source data will exist only in the data source. Aggregations defined in the partition will be stored as indexed views in source itself. So in simple words both source and aggregations will exist in data source.
  • HOLAP – Combination of both MOLAP and ROLAP. Source data will not be copied into cube and referred directly from data source whenever needed but aggregation will be calculated and stored every time cube is processed.

If you are from Mumbai you can meet us personally coming Saturday/Sunday that is 9 June at our Sakinaka office.

Comments

comments

Leave A Reply

%d bloggers like this: