Ways of data integration between ServiceNow and SAP and how do I go about differentiating them?

jiakheelim · 2 hours ago

Anyone who's been long enough within the Integration space can relate intimately to the saying of ‘There are many ways to skin a cat..’. There are multiple approaches to the same problem statement of data connectivity between two systems, with terms like APIs, CDC or even File-Based Transfer protocols (SFTPs) constantly being floated around. What are the differences between them and when should you use which?

An overview of the Integration estate within ServiceNow

When integration is being mentioned in the context of ServiceNow, the first thing that should come to mind is Workflow Data Fabric. Workflow Data Fabric, as the name suggests, is the integrated platform data layer that is designed to connect to all your data, wherever it resides, for the purposes of powering AI Agents and Workflows with the necessary context all in a secure and governed manner. It is a Data Fabric that is purpose built for and only by ServiceNow AI Platform. Of which, the different means of connectivity between ServiceNow and SAP then comes from the various modules within Workflow Data Fabric.

Connectivity – Ways of Data Integration between ServiceNow and SAP

Here are the ways to which the connectivity can be established:

Integration Hub Spokes
Zero Copy Connector for SQL
Zero Copy Connector for ERP

Integration Hub Spokes

Integration Hub Spokes are a “logical grouping” of related Actions, Connectors and Flow Logic to enable integration between ServiceNow Platform and third-party applications. It is intended to be low-code no-code and primarily REST API based integrations (even though it supports other styles as well, but it will not be discussed in this article).

With various workflows that a customer has purchased with ServiceNow comes with some out-of-the-box spokes to connect to third party systems. For, a couple of prime examples would be the SAP S4 HANA RFC / SAP S4 HANA OData Spoke with prebuilt endpoints to actions such as ‘Create Goods Receipt, Create Invoice and Authorize Payment’.

Of course, there are ways to which you can built your own Integration Hub Spoke by defining REST API related configurations and parameters, namely HTTP method, authentication protocol and pagination treatment. These all come under Integration Hub transactions eventually.

Zero Copy Connector (ZCC) for SQL

One of the most misunderstood parts of Workflow Data Fabric is that there are two ZCC functionalities – one for Structured Query Language (SQL) and the other exclusively for Enterprise Resource Planning (ERP) software(s).

In the setup of the ZCC for SQL, S/4 HANA and ECC have available connectors that ServiceNow has native support for. Think JDBC type connectivity API. Through this means of data connectivity, it is a read-only where ServiceNow fetches data from the S4 HANA/ECC source and makes this data available to the platform in the form of Data Fabric tables. It is a different type of table in ServiceNow where the table’s metadata definition is stored in ServiceNow Platform but the external records live outside of ServiceNow and when queried, lives in memory in read-only mode. While this process is transparent to users when they are using it, it is important to note the table type and the various implications it has for the platform.

Zero Copy Connector for SQL shines most when you provide your AI Agents need to run heavily analytical queries against the ERP to be returned with context / data that is important for answering any question. Due to its one-way nature, it is used largely for analytical use cases on the platform.

Zero Copy Connector (ZCC) for ERP

There is most confusion around ZCC for ERP. We addressed one of the points of confusion above as to whether there is one or multiple ZCCs on the platform. The name ZCC for ERP is rebranded from what long-time practitioners of ServiceNow know as ERP Canvas / ERP Data Hub. With the name change, the features have not reduced, but in fact, have increased with the newly added support for IDoc.

This piece does not aim to go through the engineering behind ZCC for ERP. There is an excellent community article out there that does that (providing the link here). Instead, what we will do is to lay out high level the ideas of how it is Zero Copy and where ZCC for ERP shines best.

The value of ZCC for ERP goes beyond just integration, but its real value lies in simplification and abstraction of complexity. In terms of good practices with integration, what it does is to integrate with SAP in ways of best practice. For example, for REST-based integration, ServiceNow retrieves the CSDL document when establishing connectivity to SAP S4 HANA or ECC. This CSDL document describes the data models and metadata representation available in the source and what the source chooses to expose to the connecting client. This greatly reduces the amount of work and changes the perception from point-to-point integration to more so what’s already available and what’s not.

When working with the CSDL document, it is metadata oriented as we talk about the entity types, entity fields and relationships across them, and that’s the “Zero Copy” part of it. All of this is handled under the hood and made is easy for end users to interact with them, with simplicity being one of the key drivers in our design decision.

Secondly, ZCC for ERP provides bi-directional integration options and the option to either have Remote Tables or Extraction Tables built on top of the Data Models that you have in ZCC for ERP. The options do exist, and the choice of using which and for what reasons can be thought of between performance, data volumes and frequency of invoking the APIs.

The last important point to note is that connectivity to multiple ERPs through the ZCC for ERP can be done, and will not be exhaustive only to SAP’s ERPs in the near future.

Credits to @leofrancia who played a big contributing role as well.