Register

Reducing Saves in Data Hubs

Data Hubs typically include large lists or numerous lists; both contribute to longer save durations.
Model saves block other model actions, so we want to reduce the frequency of saves as much as possible.

Reduce the Frequency of Saves

First, put the data hub model into Deployed Mode. Learn more here:
Enabling and Disabling Deployed Mode - Anaplan Technical Documentation

Deployed mode will reduce save frequency by a factor of 10—when the model is used by a small number of users or a single integration user.

What Causes Model Saves?

Calculations and model changes cause model saves, but two common triggers for saves in data hubs include the following:

  • Summaries.
  • List Restructuring.

Top-level members on large flat lists will create a large and slow calculation on line items dimensioned by this list that have a summary applied. For example, 100 line items dimensioned by a list with 10 million items will mean 1 billion cells will be aggregated to those top-level cells.
This will occur if items are added to that list or data is imported to those line items.
5.07-09 Avoid using a Top Level for large lists of transaction data 
Re: 1.05-07 Avoid Top Level for large flat lists 

List restructuring—adding or removing list items—can cause recalculations that contribute towards triggering a model save. A way to reduce this impact is to enable a feature called Pre-Allocation.

Pre-Allocation in Lists (and Impacts to Model Performance) 

This is done by adding a role (that has edit permissions) to the list. The role does not need to have users assigned to it. Do this with lists that are frequently changed.

Summary

To help reduce save frequency in models, complete the following:

  1. Enable deployed mode.
  2. Avoid summaries.
  3. Do not use top-level members on large lists.
  4. Enable pre-allocation.
The content in this article has not been evaluated for all Anaplan implementations and may not be recommended for your specific situation.
Please consult your internal administrators prior to applying any of the ideas or steps in this article.
Comments

Nice. But why wouldn't you have your datahub as deployed? Just means having a lot of production lists...

About the Author