Database Migration for SAP ERP on HANA: The Technical & Business Challenges
Bill Kozel spoke to SAP’s Juan Carlos Duarte about the challenges of migrating the database for ERP on HANA.
As the team charged with migrating SAP’s own SAP ERP database to SAP HANA, my colleagues and I needed to be mindful of not only the technical challenges, but also the business challenges. After all, our company’s SAP ERP software is fundamental to operations – as it is for thousands of companies around the world. So we need to arrive at a solution that would ensure uninterrupted performance – as well as a future of unlimited scalability path.
The Technical Challenges
The technical challenges arose from the fact that SAP ERP is a very large system with a very large database. Prior to moving to SAP HANA, our ERP database was distributed over 35 servers and consumed six terabytes of storage. Initially, we planned to export it to a single two-terabyte server but we quickly found that we needed more RAM to provide enough processing power for the future. At that time, there were no certified four-terabyte servers on the market but several manufacturers’ servers were certified later in the process, giving us good options in the 4TB range.
The Business Challenges
To meet the needs of our business, we needed to make absolutely sure that SAP’s operations could be fully handled by our hardware and software systems – with capacity to spare for any future acquisitions and future growth.
As in most organizations, SAP ERP is the single largest software application we operate. What’s more, SAP’s ERP database is extremely dynamic, constantly drawing on new sources of information. And whenever there is a merger or acquisition, we frequently need to import and integrate the new company’s ERP data into our database – even when that database has some unique characteristics.
For example, when SAP acquired SuccessFactors, the company’s ERP application contained information from specific processes that needed to be incorporated into our own SAP ERP on SAP HANA. One of these was a strong focus on HR applications, which meant that we needed to allot a great deal of capacity for employee records and related processes.
As we forecasted demand, we quickly discovered that even a four-terabyte server wasn’t going to be large enough for the near future, and that it would take just one year to outgrow its capacity. So, we combined two four-terabyte servers for a total of eight terabytes in a so called HANA Scale-Out architecture and we’re looking into that direction as our next move. This required a different technical landscape, but offers greater flexibility and scalability.
Why not simply use a single eight-terabyte server? Because as of this moment, the current models are not yet certified.
Prior to migrating to SAP HANA, our ERP instance ran on an IBM DB2 database. In the future, we plan to employ a Sybase database for its unique abilities in delivering solutions to mobile devices. In addition, we continue to work with manufacturers on certifying eight-terabyte servers – and larger – for the big, complex processes we use as part of our SAP ERP on SAP HANA instance.
Stay tuned for additional details in the weeks and months ahead!