Performance Optimization Flashcards

What is the purpose of the 'Optimizer Level' setting in an IICS mapping session properties?

Answer: It enables automatic optimizations such as early selection, push-into, and cost-based reordering of transformations

The Optimizer Level instructs IICS how aggressively to apply automatic rewrite rules—higher levels enable more transformations like pushdown and cost-based reordering.

An IICS synchronization task is performing poorly because it recalculates derived fields for every source row even when the source data is unchanged. Which technique addresses this?

Answer: Using CDC (Change Data Capture) to process only modified rows

CDC identifies only rows that changed since the last run so unchanged rows never enter the mapping, eliminating wasted computation.

Which IICS Secure Agent JVM tuning parameter should you increase when a mapping that processes large XML or JSON payloads throws 'OutOfMemoryError'?

Answer: -Xmx (maximum heap size)

-Xmx sets the JVM's maximum heap, providing more memory for parsing and holding large in-flight XML/JSON documents.

When designing an IICS pipeline to load a 500 GB file from Amazon S3 into Redshift, which approach yields the fastest load time?

Answer: Splitting the file into multiple smaller files and using Redshift COPY with parallel manifest loading

Redshift COPY with a manifest file over multiple S3 objects parallelizes across all compute nodes, saturating I/O bandwidth and minimizing load time.

In IICS, what does enabling 'constraint-based loading' on a target allow the Secure Agent to do?

Answer: Enforce referential integrity by loading parent tables before child tables automatically

Constraint-based loading ensures parent records exist before child records are inserted, preventing foreign-key violations without manual ordering.

A developer notices that a Lookup transformation configured as 'connected' is issuing one database query per incoming row. What change eliminates this row-by-row query pattern?

Answer: Enable lookup caching (static or dynamic) so all lookups are served from memory

With caching enabled, the Lookup loads all qualifying rows into memory once at session start and resolves subsequent lookups locally without database round-trips.

Which IICS Cloud Data Integration feature allows a single mapping to automatically scale compute resources up or down based on data volume at runtime?

Answer: Advanced cluster (Spark-based) execution with auto-scaling worker nodes

Advanced cluster mode runs mappings on an elastic Spark infrastructure that can dynamically add or remove executor nodes based on workload demand.