Informatica Cloud DI Foundation cheat sheet
Informatica Cloud DI Foundation · Keep the key ideas within reach.
Cloud Data Integration
Informatica Cloud Data Integration (CDI) is a multitenant SaaS service: the Secure Agent runs on customer-hosted infrastructure, while the repository, task definitions, and scheduling metadata live in Informatica Cloud (IICS).
A Secure Agent is a lightweight Java process installed on Windows or Linux that executes mappings and workflows; each agent group contains one or more agents, and a task assigned to a group is dispatched to a member of that group.
Mapping Designer objects are built from Source, Transformation, and Target transformations connected in a linear DAG; the mapping must contain at least one source and one target, and every transformation port must be linked or the mapping validation fails.
A Mapping task runs a mapping on demand or on a schedule and can use source, target, and lookup parameters; a Taskflow orchestrates multiple tasks with conditional paths and notification steps, while a linear Taskflow runs steps sequentially.
Synchronization tasks are template-driven and offer insert, update, upsert, delete, and full-load modes; upsert requires a defined primary key or a user-specified key field, and delete mode requires the key field to match source records.
A Replication task copies one or more source objects to a target in bulk without transformation logic, supports incremental load via change capture, and is preferred for straight object copies; use a mapping when transformation is needed.
Elastic mappings push transformation logic to the source or target database (for example Snowflake or Redshift) when the database engine supports the operation, reducing data movement; unsupported expressions fall back to the Secure Agent.
Connections store endpoint, credential, and Secure Agent group information; a connection can be runtime-environment bound, so the same mapping can be parameterized to point at different connection objects per environment.
Task monitoring shows run status (Running, Success, Failed, Stopped) with row-level counters for source rows, target rows, and rejected rows; rejected rows are written only when an error-handling target or reject file is configured.