Running a large company is impossible without having an ERP system in place, and SAP business software remains at the forefront in this category. But when companies are looking towards new technologies such as data lakes, machine learning or predictive analytics, SAP alone is just not enough. To keep up with tech trends, businesses have to face the challenges of integrating SAP with non-SAP technologies and embark on a crusade against data silos.
SAP as a must-have
SAP software has been steadily evolving since the 70s, from a mainframe-based suite of finance and logistics programs into today’s S/4HANA ERP suite. With a modern, top-notch, in-memory columnar database it offers full coverage of all major industries and business processes, from data entry to finance, legal, compliance, production planning, and HR. It’s the de facto choice for all major corporations on the planet to manage their business data.
For reporting purposes, SAP offers SAP HANA, SAP ERP’s own reporting capabilities, or SAP BW (Business Warehouse) for cross-system reporting and data warehousing. The integration between SAP’s own solutions is well thought out and easy to implement.
Challenges of integrating SAP with other technologies
However, most businesses run more than SAP alone. Approaches that involve data lakes or customer insight solutions require deep integration between SAP and non-SAP technologies. This is where the going gets tough. While SAP solutions are in general extremely mature, they often form data silos.
Accessing and extracting SAP data is challenging for various reasons:
- Complexity. Experienced SAP professionals will know only some of the most important 90,000 ERP database tables by heart, for just one or two modules. Simply identifying relevant data for a specific use case or scenario is challenging – most tables have hundreds of fields, some with obscurely abbreviated flags or values.
- Customization. More often than not, the SAP solution has been customized for customer specific requirements and business processes.
- Access. Accessing and extracting data – once it is clear what data is required – is technically difficult. Scenarios may require initial full extract followed by deltas, data federation or replication.
- Governance. Data governance and access restrictions are key topics that are impacted by new, tighter regulations and compliance constraints such as GDPR, BCBS 239, or HIPAA.
- Cataloging. Any data source should be integrated with a company-wide data catalog to simplify governance, finding relevant data sets, and implementing data management strategies including end-of-life scenarios.
Breaking down data silos
Silos are good for storing grain, not data. Isolated information becomes useless or even misleading, preventing organizations from moving their strategy forward and adapting to fast-paced market changes. Breaking down data silos is on a to-do list of every CIO aiming to sharpen their competitive edge. Enterprise-wide data lakes provide the basis for many high-value use cases, ranging from delivering complete customer insight to grow the business to connecting products and services for entirely new business models and revenue streams like predictive maintenance. Additionally, there are operational side effects that include reducing SAP TCO as well as giving a viable and effective way to migrate and decommission SAP systems.
All these benefits are achievable and real however many organizations are unsure how it impacts their current SAP deployments for one key reason: how can I do this in a supported manner? As a critical enterprise application, there is no room to just tinker around with it in a way that would invalidate SAP’s terms around approved data exchange and mechanisms. That’s why it’s important to work with an SAP-certified partner and their certified solutions. Datavard is that partner for Cloudera.
In subsequent blog posts, we will be looking in detail at topics such as SAP TCO reduction, Customer 360, Data Warehousing on Hadoop, and SAP decommissioning. How else would you like to use your SAP data?
The article was originally published on the Cloudera Vision Blog.