informatica performance tuning

21
Performance Tuning* Check ‘Collect Performance data’ option in the session properties of a session in Server Manager

Upload: anushakalli

Post on 24-Nov-2014

322 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Informatica Performance tuning

Performance Tuning*

Check ‘Collect Performance data’ option in the session properties of a session in Server Manager

Page 2: Informatica Performance tuning

Performance Tuning

Page 3: Informatica Performance tuning

Performance TuningSource Qualifier

  Target   Evaluation

Buffer Input efficiency

BufferOutput efficiency

Buffer Input efficiency

Buffer Output efficiency

 

High Low High Low Source database slow, eliminate read bottleneck

Low Low/High Low High Target database slow, eliminate write bottleneck

Low High High Low DTM slow, optimize session or mapping

Efficiency PercentageHigh= 80-100%Low= 0-20%

Page 4: Informatica Performance tuning

Performance Tuning

The counters help identify• Read/Write/DTM bottlenecks

• Caching problems

• Transformation errors

• Shared memory allocation problems

To avoid read bottleneck• Reduce the number of records processed by filtering the

source

• Index the source

• Replace default query with an optimized query in SQL override of Source Qualifier

Page 5: Informatica Performance tuning

Performance Tuning*

To avoid write bottleneck• Utilize SQL loader facility of the database• Drop indexes before load and rebuild after

the load• Increase the database block size

To avoid DTM bottleneck• Reduce the amount of detail in the session

logs

Page 6: Informatica Performance tuning

Performance Tuning

Page 7: Informatica Performance tuning

Performance Tuning*

To avoid DTM bottleneck

• Run parallel sessions in Concurrent batches

• Use incremental aggregation for mappings that use aggregation

Page 8: Informatica Performance tuning

Performance Tuning

Page 9: Informatica Performance tuning

Performance Tuning*

To avoid DTM bottleneck• Optimize mapping• Optimize session

For cache problems• Increase index and data cache

Page 10: Informatica Performance tuning

Performance Tuning

Click

Page 11: Informatica Performance tuning

Performance Tuning*

For shared memory allocation problems

• Increase the shared memory size

Page 12: Informatica Performance tuning

Performance Tuning

Page 13: Informatica Performance tuning

Performance Tuning

To optimize sessions

• Increase shared memory size

• Increase buffer block size for very large row sizes

• Consider the commit interval. More commits=slower performance

• Disable decimal arithmetic if possible

Page 14: Informatica Performance tuning

Performance Tuning

Page 15: Informatica Performance tuning

Performance Tuning

To optimize mapping

• Utilize single pass reads, use SQL override

• Place filters, aggregators as close to source as possible

• Trim spaces before comparing strings

• Operators are faster than functions,i.e.,|| vs Concat

• Cache lookup table columns

Page 16: Informatica Performance tuning

Performance Tuning

Optimizing Transformations

Page 17: Informatica Performance tuning

Performance Tuning

Filter Transformation

Use the Filter transformation early in the mapping.

Use the Source Qualifier transformation to

filter.

Page 18: Informatica Performance tuning

Performance Tuning

Joiner TransformationPerform joins in a database when possible. Join sorted data when possible. For an unsorted Joiner transformation,

master---> fewer rows.For a sorted Joiner transformation,

master source --->fewer duplicate key values

Page 19: Informatica Performance tuning

Performance Tuning

Lookup Transformation

Cache small lookup tables(<300 MB)Use a persistent lookup cache for static lookup tables. Override the ORDER BY statement for cached lookups

Page 20: Informatica Performance tuning

Performance Tuning

Stored Procedure Transformation

Reduce the row count. Create an expression

Page 21: Informatica Performance tuning