Informatica CAI R42 Professional key terms

Informatica's cloud service for building and running API and process integrations, combining real-time service orchestration, event-driven processing, and human-task workflows on a shared cloud runtime.

A reusable design-time asset that defines an integration or orchestration flow in CAI, holding steps, connections, and error handling that can be published as a service or invoked by other processes.

A published endpoint created from a process or guide that external clients or other processes call over HTTP, exposing defined input and output parameters and a fixed invocation URL.

A single unit of work inside a process, such as a service call, assignment, decision, loop, or human task, connected by transitions that determine execution order.

A directed connection between two process steps that passes execution and data forward, optionally carrying a condition so the flow branches based on runtime values.

A branching step that evaluates one or more conditions and routes the flow down the matching outgoing transition, letting a single process handle multiple runtime scenarios.

A process invoked from within another process so common logic can be reused and maintained in one place; inputs and outputs are mapped at the call site.

The design pattern of catching errors raised during step execution and routing them to recovery steps, so a process can retry, compensate, log, or return a controlled error.

The CAI tool for building guided, human-in-the-loop experiences where users answer questions and the guide branches to the right outcome, often used for intake and triage.

A workflow step that pauses automated execution and assigns work to a person or group, resuming the process when the assignee submits a response or approval.

A human-task pattern where a reviewer accepts or rejects a request, and the process then continues down an approved or rejected path with the decision recorded.

A reusable design-time definition of how to call an external system, holding the endpoint, authentication, and request/response shapes so processes can invoke it consistently.

A configured, secured link to an external application or data source, holding credentials and endpoint details that connectors and process steps reference at runtime.

An HTTP-based interface style using resources, verbs, and status codes; CAI consumes and publishes REST endpoints, mapping JSON payloads to process input and output fields.

A web service using XML envelopes and a WSDL contract; CAI can call such services by importing the contract and mapping request and response elements to process fields.

An XML contract describing a SOAP service's operations, messages, and data types, used in CAI to generate the connector shape and validate request and response structures.

The structured text body exchanged with REST services, consisting of objects, arrays, and typed values that CAI parses into process fields and serializes on output.

An XML schema that defines the allowed elements, order, and data types for XML documents, used in CAI to validate and shape hierarchical process data.

The design-time act of connecting a source field to a target field, optionally with transformations, so data flows correctly between steps, services, and process variables.

A formula evaluated at runtime that combines fields, functions, and operators to compute a value, used in mappings, conditions, and assignments inside a process.

A named value scoped to a process run that stores input, intermediate results, or output, and can be read and written by steps throughout execution.

A declared value a caller must or may supply when invoking a service or subprocess, defining the contract's inbound shape and enabling validation at the boundary.

A declared value a process returns to its caller on completion, forming the response contract that client applications and downstream steps rely on.

The act of publishing a designed process to the runtime environment so it becomes callable, including version selection and any required connection or agent configuration.

The hosted or agent-based execution context where deployed processes run, providing compute, connectivity to on-premises systems, and logging of each run.

A lightweight Informatica program installed in a customer network that executes jobs and brokers connectivity to on-premises systems that the cloud runtime cannot reach directly.

An architecture where cloud-hosted orchestration works together with on-premises agents and systems, keeping data behind the firewall while still using cloud design and monitoring.

An integration style where a business event, such as a new order or updated record, triggers processing asynchronously rather than waiting for a scheduled or direct request.

A configured listener that receives published events from a topic or source system and starts a process for each matching event, decoupling producers from consumers.

A buffered channel that holds messages until a consumer processes them, providing decoupling, load leveling, and retry behavior between producing and consuming systems.

A property where repeating the same request produces the same result without duplicate side effects, essential when retries or at-least-once message delivery can occur.

Rules that define how many times and how often a failed step or service call is attempted before the process declares failure and hands off to fault handling.

A configured limit on how long a step or service call may wait for a response before it is treated as failed, protecting the process from hanging indefinitely.

The overall design decision for how a process detects, logs, retries, compensates, and reports failures, chosen per integration based on criticality and business impact.

The CAI workspace where operators inspect process runs, view step-level logs and payloads, and diagnose failures using status, timestamps, and error messages.

Declarative logic that encodes organizational policy, such as eligibility or pricing thresholds, which a process evaluates to drive decisions without hard-coding in every step.

An authorization framework where a client obtains a time-limited access token from an identity provider and presents it to call protected APIs, configured within connections.

Managing multiple releases of a process or service so existing callers keep working while new logic is tested and promoted, with a chosen version active at runtime.

A front-door layer that routes, authenticates, throttles, and logs API requests before they reach backend services, often placed in front of published CAI endpoints.