SAP Process Integration (PI) for Dummies (previously called Exchange Infrastructure (XI))

Process Integration or PI forms one of the components of SAP’s NetWeaver along with other components such as Business Intelligence (BI), Master Data Management (MDM) and so on.

Earlier when PI was not available, data communication between multiple systems was very time consuming and very confusing to design, configure and track. Hence it was called a spider web connection. As seen in Figure 1 below, even an addition of one system to the network needed too much of coding and configuration to make the new system communicate with other systems in the network.  As seen in Figure 2 below, with PI coming in as a middleware, it avoided this extra work. Adding an extra system needed only 2 more extra connections, one from PI to new system and from new system to PI box.

Figure 1: Spider Web Connection

Figure 2: Advantage PI: A single point of integration for all systems, SAP and non-SAP

From the figure above, if all the systems were to know XML, then work will get so simple. This is where our PI middleware comes into picture. PI middleware has a component called ‘Adapter’ which helps in this data conversion from XML to any format and vice versa.

Some examples of middleware products other than PI in the market are SeeBeyond, TIBCO, etc.

Now, let us briefly see the various components of SAP PI:

Figure 3: Components of PI

Take a look at the Figure 3 above. Let’s briefly see what these important components of PI do:

a.       System Landscape Directory (SLD) : SLD contains the info about the set of systems in the organization and their details like system name (HR, MM, SD, etc.), products installed, their versions etc., SLD is not a component of PI but a part of NetWeaver. PI uses some components of SLD.

b.      Integration Repository or Enterprise Services Repository (IR or ESR):  The design time component of PI, the Source & target structures is defined in repository box (IR or ESR). Mapping is also done in IR box where mapping means converting source structure to target structure.

Key functions of the IR or ESR box are:

1.       Creating the Structures (Source and Target).

2.       Mapping the Structures.

c.       Integration Directory (ID): The configuration time component of PI.

Key functions of the ID box are:

1.       Helps in connecting 2 systems (Source & Target).

2.       Configuration of the scenario.

d.      Adapter Engine (AE): AE converts any data to XML and vice versa. In case of SAP system as target, Adapter will convert XML to IDOCS or RFC. Same way, if source is a SAP system, IDOC/RFC will be converted to XML by the AE module.

e.      Integration Engine (IE): IE is a runtime component of PI which helps in message transformation as per the configuration made in the Integration Directory.

f.        Business Process Engine (BPE): With the help of a graphical process editor it controls message correlations and how messages are processed as part of cross component Business Process Management (ccBPM). This will be helpful in creating cross-system process for processing messages.

g.       Partner Connectivity Kit (PCK): PCK is a development java library freely downloadable. You can write your own program if there is a new format of data other than supported ones like IDOC, RFC etc.,
Market Place adapters are 3^rd party programs that have to be certified by SAP. This is created using PCK or adapter specification.

h.      Proxy: It is an alternative for adapter.  There are 2 types of proxies i.e.; Java & ABAP. Sometimes adapters won’t be there for a particular application and instead we can use Proxy. Proxy thus avoids PCK.

To understand PI in the simplest manner, let is imagine 3 Boxes as seen in Figure 4 below. Source box and Target box can be SAP or Non SAP. PI middleware uses standard XML format for communication. Hence, Input and Output of PI box will be XML.

Figure 4: The 3 Boxes