.NET PivotGrid and Dashboard: OLAP Performance Improvements

09 June 2015

OK, let me state the blindingly obvious for a moment: the performance of aggregation or sorting operations (or both!) are key factors when you’re working with OLAP data sources. So, with v15.1, we took the time to do some research and design and have managed to increase the speed of the data engine used in the .NET Pivot Grid and Dashboard by optimizing some basic data processing – anything from constructing queries to the parsing of returned data.

Sorting is now performed on the client side. This allows you to speed up the process of visualizing data in complex scenarios such as sorting by summary and sorting in Tabular mode. Moreover, as an added bonus, using client sorting gives you the capability to perform custom sorting with the CustomServerModeSort events.

To compare the new Pivot Grid’s performance with other products in OLAP mode, we investigated two scenarios. With both scenarios, the pivot data and row areas display fields from the Adventure Works cube and are deployed on a local machine.

Pivot Grid Performance Improvements v15.1

 

Scenario PivotGrid v14.2 PivotGrid v15.1 Competitor 3 Competitor 4
#1 6 sec 3 sec 5 sec 4 sec
#2 1 min 20 sec 19 sec 35 sec 1 min 40 sec

The chart and table shown illustrate the performance improvements of the Pivot Grid in OLAP mode in v15.1 when sending queries to the Adventure Works cube. The data shows a comparison with the old v14.1 version and a couple of our (unnamed) competitors in this space. Note that the ADOMD provider was used to establish a connection to the OLAP cube.

Here is the data we used for each of the two scenarios.

Scenario 1

Data area: Internet Sales Amount, Internet Order Quantity
Row area: Postal Code, Product
Row count: ~ 100 000

Pivot Grid Performance Improvements Scenario 1

Scenario 2

Data area: Sales amount, Total Product Cost
Row area: Business Type, City, Product
Row count: ~ 600 000
Sorting: Product field is sorted by Total Product Cost

Pivot Grid Performance Improvements Scenario 2

I’m sure you’ll agree with me that these basic performance improvements will enhance the snappiness of the data analytics parts of your applications.

no comments
No Comments

Please login or register to post comments.