SAP Integration Suite

The Integration Suite provides everything to integrate in the cloud era: build integration flows (CPI), manage and secure APIs (API Management), connect to third-party apps (Open Connectors), exchange events (Event Mesh), and design B2B/EDI integrations, all as managed cloud services with pre-built content.

• Cloud Integration (CPI): iFlows.
• API Management: publish, secure, monitor APIs.
• Open Connectors: pre-built third-party connectors.
• Event Mesh & B2B/EDI capabilities.

An integration flow, or iFlow, is a pipeline drawn on a canvas. A message arrives at a sender adapter, passes through a series of steps, and leaves through a receiver adapter. Everything in Cloud Integration is a variation on that shape.

The steps fall into a few kinds. Adapters speak the protocol at each end: HTTPS, SFTP, OData, IDoc, JMS and others. Mappings convert one structure into another, and message mapping is where most of the work sits. Routing sends a message down different paths depending on its content. Enrichment calls out for extra data mid-flow. Scripting in Groovy handles what the graphical steps cannot.

Two things make it different from a program that does the same job. The flow is deployed as a unit, so promoting it between tenants is one action. And every message is traced, so the monitoring view shows what arrived, what each step did to it, and where it stopped.

Everything is in the browser. The Design workspace holds integration packages, which group iFlows, mappings and value mappings for one scenario. The Monitor workspace shows deployed artefacts and message processing, which is where you spend most of your time once something is live. Settings holds security material: the credentials, keys and certificates the adapters use.

The API Business Hub carries prepackaged integration content that SAP and partners publish. For common scenarios, such as replicating employees from SuccessFactors into the ERP core, that content is a supported starting point rather than something to build. Copying a package into your own workspace and adjusting it is normal and expected.

Transport between tenants is by exporting and importing packages, or through a configured transport mechanism. There is no SPRO and no transport request in the classic sense, which is the adjustment for anyone arriving from on-premise. See SAP integration for how this sits beside the older tooling.

Twenty minutes in a trial or training tenant, and it is the daily work of an integration consultant.

1. Open the Monitor workspace and filter message processing by status Failed over the last day.
2. Select one and read the error. Most fall into three groups: the receiver rejected the payload, the connection or credential failed, or a mapping hit data it did not expect.
3. Open the message processing log and step through the flow. The log shows which step it reached, which is what separates a mapping problem from a connectivity problem.
4. If tracing is enabled, read the payload at each step and find where the field went wrong.
5. Fix and redeploy the iFlow, then use the retry where the scenario supports it. Whether a retry is safe is a design question, not a monitoring one.

Integration Suite is SAP’s strategic integration platform, use it for new cloud/hybrid integration, API management and event-driven scenarios. It is central to modern SAP landscapes and BTP.

The reverse is worth saying plainly too. High volume asynchronous document exchange between two SAP systems that already speak IDoc is not obviously improved by moving it, and a working PI or PO landscape does not need replacing on principle. Integration Suite earns its place on new cloud and hybrid scenarios, on API management, and wherever prepackaged content already covers the job. See Cloud Integration and API Management.

Cloud Integration is the part people mean when they say Integration Suite, and the other capabilities are where a lot of the value sits.

API Management puts a managed layer in front of an API. It handles the things you would otherwise build into every consumer: authentication, rate limiting, quotas, caching and analytics on who is calling what. It also gives you a developer portal, so an API can be published to people you do not personally onboard.

Open Connectors supplies normalised connectors to third party applications, so integrating with a common cloud service is configuration rather than reading its API documentation from scratch.

Event Mesh carries the event driven style: publish that a business object changed and let subscribers react, without the publisher knowing who they are.

Integration Advisor and trading partner management address business-to-business exchange, where the formats are industry standards rather than anything SAP designed, and where much of the effort is in agreeing the mapping with a partner.

Graph and the data side present connected data across products through one API, which reduces the number of point integrations somebody has to write.

Whether you need any of them follows from the style question rather than the product. An organisation exposing APIs to partners needs API Management. One reacting to business events needs Event Mesh. One doing neither should not buy either just because they are in the same suite. See SAP PI and PO for the on-premise middleware many of these landscapes are migrating from.

If your background is on-premise middleware, most of what you know transfers and the working rhythm does not.

The concepts map closely. An integration flow is doing what an integration scenario did. Message mapping is still message mapping, and the graphical mapper is recognisable. Adapters play the same role. Value mappings replace the lookup tables you maintained before.

What changes is everything around the build. There is no dual stack, no Enterprise Services Repository and Integration Directory split, and no NetWeaver Administrator: design and configuration live in one browser workspace. Deployment is per artefact rather than per channel, so promoting a change is smaller and more frequent. Releases arrive on SAP's schedule, so an adapter can gain capabilities without you upgrading anything.

The habit that needs unlearning is scale of change. On-premise projects batch changes into releases because moving them is heavy. Here a single iFlow can be redeployed in a minute, which is an advantage until nobody can say which version is running in which tenant. Keeping packages under version control and promoting deliberately is the discipline that replaces the transport queue.

• Wrong tool for the integration style.
• Ignoring monitoring/error handling.
• Point-to-point instead of governed integration.

• Scripting what a standard step already does. Groovy is available and every line of it is something you now maintain and test on each release.
• Leaving tracing on in production. It is a debugging aid, it stores payloads, and those payloads often contain personal data.
• Building from scratch without checking the Business Hub. Prepackaged content is supported and updated, and a bespoke copy of it is neither.

• Storing credentials in the flow instead of the security material store. They belong in one place, per tenant, so promoting a package does not carry a secret with it.
• One enormous iFlow. Several small flows are easier to monitor, redeploy and reason about than one that does everything for a scenario.

Reading a failed message is the start, and building the flow with its mapping, error handling and monitoring is the part you do in the course.