Cloud Data Integration
Cloud Data Integration covers how Informatica Cloud (IICS) moves and transforms data. The exam tests Secure Agent runtime choices, connection setup, synchronization versus mapping tasks, transformation behavior, parameterization, pushdown, and job monitoring.
Runtime and connections
- The Secure Agent is the execution process; install it on-premises or on a cloud host and place it in an agent group for availability. A runtime environment associates a task with an agent group.
- Connections hold endpoint, credential, and property details. Parameterize connection properties so the same mapping can run in test and production.
- A synchronization task is a point-to-point load with simple field mapping and filters. A mapping task runs a reusable mapping that can include complex transformations.
Transformations and run-time control
- Filter removes rows by condition. Expression computes ports. Lookup enriches rows and uses a cache plus a multiple-match policy. Joiner combines master and detail pipelines. Aggregator groups rows. Router sends rows to groups. Update Strategy marks rows with DD_INSERT, DD_UPDATE, DD_DELETE, or DD_REJECT.
- Pushdown optimization sends eligible logic to the database; source, target, and full pushdown differ in how much is pushed. Unsupported transformations prevent full pushdown.
- Mapping parameters are fixed at run time; variables can change and persist values between runs.
Worked example
Load Salesforce Accounts changed since the last run into Oracle. Define mapping parameter $$LastRunTime, filter with LastModifiedDate > $$LastRunTime, and add Update Strategy with DD_UPDATE keyed on AccountId. Run it as a mapping task. Reasoning: the filter limits rows, Update Strategy makes reruns idempotent, and the mapping task is needed because the custom transformation logic is not available in a synchronization task. After success, store the job start time for the next run.
Traps
Do not assume every transformation pushes down; check the pushdown compatibility. A lookup with multiple matches can fail or return the wrong row unless the policy is set. DD_REJECT discards the row. A Secure Agent that is down, or lacks network access to the endpoint, fails the task even if the connection is valid.