In this lesson, we explore the best practice recommendations for optimizing planning model performance in SAP Analytics Cloud.
Scenario: You’re tasked with building a high-performing planning model in SAP Analytics Cloud. Your model needs to handle large datasets efficiently, providing quick response times for users. However, you've noticed some performance concerns, impacting the overall user experience.
Exploring Best Practice for Performance in a Planning Model
In this video, we take a closer look at the model health checklist. This checklist will help you identify and address potential bottlenecks in your SAP Analytics Cloud planning model’s performance.
Summary
The model health checklist helps you to systematically evaluate your model's health and identify potential bottlenecks and potential areas for improvement. Remember, there's no single ‘best’ model design; optimization depends on your specific data and requirements. The optimal model design is context-dependent and requires iterative refinement.
The model health checklist includes:
- Dimension granularity validity check: It’s important to carefully examine the granularity of each dimension, that is, the number of members and depth of a hierarchies within dimensions. Do you need all of the detail? Dimension granularity is covered in a separate lesson.
- Dimension count: Try to keep the number of dimensions (including standard dimensions like Account, Date, and Version) between 8 and 12.
- Public/private version data load: Strategies for optimizing version data load is covered in a separate lesson..
- Formulas: Use advanced formulas within data actions, triggering the on-demand execution of formulas, reducing the system load over formulas executed in real-time.
- Exception aggregations: A natural behavior of a multi-dimensional engine is to sum up data along multiple hierarchies. Two dimensions with hierarchies that are 9 levels deep (with 10 leaf members) creates 100 potential aggregations. Exception aggregations, such as averages or min/max values, are calculated in a second pass, creating additional processing steps that query the underlying data and longer runtimes. Try moving these calculations to advanced formulas for better performance.
- Eliminate unnecessary zeros: While occasionally useful, excessive zeros can impact performance. We’ll show you how to reduce unnecessary zeros in a lesson later in this unit.
- Security/data access profile: Restrictive data access profiles generally improve performance, however, it may have an impact on the functional administrator, as modeling rights will be refined depending on a power user’s privileges. For this reason, it’s important to carefully plan and implement data access profiles to balance security and performance.
Additional Information
To view details on the planning model diversity illustration chart and modeling checklist, go to the SAP Community blog post SAP Analytics Cloud for Planning: Modeling checklist.