ICC Performance Optimization 3 — Questions and Answers
Question 1: Which Informatica IICS Advanced Cluster configuration option helps reduce data shuffling costs when performing large aggregations on cloud data?
- Disabling broadcast joins
- Enabling partitioning by the aggregation key (Correct answer)
- Setting the session commit interval to 1
- Using a Router transformation before the Aggregator
Correct answer: Enabling partitioning by the aggregation key
Partitioning the data by the aggregation key ensures all rows for a group land on the same node, eliminating cross-node data shuffles.
Question 2: What effect does enabling 'high precision' on a numeric port in an IICS mapping have on performance?
- It accelerates numeric calculations by using native double types
- It can slow performance because it uses Informatica's internal high-precision decimal engine (Correct answer)
- It enables pushdown for all arithmetic functions
- It forces string-to-number casting at the source qualifier
Correct answer: It can slow performance because it uses Informatica's internal high-precision decimal engine
High precision uses a software decimal library rather than native hardware floating-point, adding CPU overhead and reducing throughput.
Question 3: When should you use a 'Sorter' transformation placed before an 'Aggregator' in an IICS mapping?
- Always, to improve readability
- When the Aggregator is not configured to use sorted input, and the source data is unsorted (Correct answer)
- Only when writing to a flat file target
- When the mapping uses pushdown optimization
Correct answer: When the Aggregator is not configured to use sorted input, and the source data is unsorted
A Sorter before the Aggregator enables 'sorted input' mode, which lets the Aggregator process groups sequentially without building a full in-memory cache.
Question 4: In an IICS Mass Ingestion task loading data to a data lake, which setting most directly controls how many files are processed in parallel?
- Target commit size
- Maximum concurrent connections / parallelism degree (Correct answer)
- Secure Agent heap size
- Source row limit
Correct answer: Maximum concurrent connections / parallelism degree
The parallelism / maximum concurrent connections setting governs how many file transfer threads run simultaneously, directly impacting ingestion throughput.
Question 5: Which IICS transformation should be avoided in the critical path of a high-volume mapping because it blocks the pipeline until all input rows are received?
- Expression transformation
- Aggregator transformation (without sorted input) (Correct answer)
- Router transformation
- Lookup transformation with connected output
Correct answer: Aggregator transformation (without sorted input)
The Aggregator must collect all rows for a group before it can emit results, creating a blocking point that halts downstream processing.
Question 6: What is the primary benefit of using a 'persistent cache' for a Lookup transformation in IICS when the mapping runs on a recurring schedule?
- It encrypts the lookup data at rest
- It reuses the pre-built cache from a previous run, skipping the cache rebuild step (Correct answer)
- It forces the lookup to use a database index
- It enables dynamic port generation
Correct answer: It reuses the pre-built cache from a previous run, skipping the cache rebuild step
A persistent cache saves the lookup index to disk after the first run so subsequent runs load the existing cache instead of re-querying the source.
Question 7: When optimizing an IICS mapping that writes to Snowflake, which target load strategy generally provides the best bulk-insert performance?
- Row-by-row insert using JDBC
- Staging data in S3/Azure Blob and using Snowflake COPY INTO command (Correct answer)
- Using the Update Strategy transformation with upsert mode
- Writing through a Stored Procedure transformation
Correct answer: Staging data in S3/Azure Blob and using Snowflake COPY INTO command
Staging files in cloud storage and issuing a COPY INTO command leverages Snowflake's bulk-load engine, far outperforming individual row inserts.
ICC Performance Optimization 3
Which Informatica IICS Advanced Cluster configuration option helps reduce data shuffling costs when performing large aggregations on cloud data?