Performance Tuning of Informatica mapping | Interview Question

 One of the most frequently asked Informatica interview question is about performance tuning. 

• How do you optimize mapping?

• How do you identify which part of mapping is taking more time to run?

• How do you optimize a transformation in your mapping?

• How do you improve the performance of a mapping?

---

First of all, what is performance tuning?

• To identify and resolve bottlenecks 
• Performance bottleneck can occur at target, source, mapping, and session.

In this post, we will see where the bottleneck is caused, how to identify and eliminate them.

1) TARGET BOTTLENECK

• Caused while Integration Service writes to a target database.

How to identify Target Bottleneck?

• Configure to write a flat-file. If the performance increases then you have a target bottleneck.

• This means more time is spent on writer thread than transformation or reader thread.

How to eliminate Target Bottleneck?

• Configure index and key constraints on the target.

2) SOURCE BOTTLENECK

• Caused while reading from the source database.

How to identify Source Bottleneck?

• If the session reads from a flat file, then probably no source bottleneck.

• If the source connection is relation then add the filter to the source, perform a read test.

• This means more time is spent on the reader thread than transformation or writer thread.

How to eliminate Source Bottleneck?

• Configure index and key constraints on the source.

3) MAPPING BOTTLENECK

• Caused while transforming data.

How to identify Mapping Bottleneck?

• Check the session log to find how time is spent on the transformation thread.

• This means more time is spent on the transformation thread than the reader or writer thread.

How to eliminate Mapping Bottleneck?

• By optimizing the transformation properties.

4) SESSION BOTTLENECK

• If the above bottlenecks are eliminated then you are probably facing the session bottleneck.

How to identify Session Bottleneck?

• Check the information about each transformation input, output, and error row details.

How to eliminate Mapping Bottleneck?

• By optimizing the session properties.