The purpose of model concurrency testing is to simulate the expected real world concurrent user interaction with an Anaplan model and scheduled events (imports, exports, etc.) during typical daily usage and peak periods, such as quarter end.
Why is model concurrency testing important?
The model performance that users experience is critical to successful implementations and delighted customers. Model bottlenecks can be identified and resolved to ensure optimal model performance and avoid post go-live performance issues. Confidence that a model will perform optimally at the expected user concurrency and data volumes for the various user journeys/business processes is a key go-live success criteria.
Can globally distributed users be simulated?
Yes. The tests can be conducted to replicate user interaction from any global region so that network latency is factored into the model concurrency results.
What information is required?
To create the automated scripts to simulate concurrent model usage the following key inputs will be required:
Project milestone dates
Number of model users, user roles, and selective access details
Anticipated number of concurrent users
Step by step details of the key user journeys (how the users are going to interact with the model) with the acceptable time for each step to be processed/completed by the model.
Expected split of users between each user journey and number of times each process will be run in an hour
Sanitized version of the model with data volumes representative of expected production scale