Moving SAP CRM to SAP HANA: The Replication Process
Bill Kozel spoke to SAP System Analyst Hans Baessler about the replication process of moving CRM to HANA.
SAP CRM is essential to our company’s sales and customer service program – as it is to thousands of companies around the world. So when it was time to migrate SAP CRM to SAP HANA, our data migration team needed to be absolutely, positively sure that the transition would be smooth and seamless to our thousands of daily users. Two of the most critical migration processes were storage replication and system replication, which help ensure that SAP CRM will always be available to the people who rely on it. The revolutionary nature of SAP HANA requires new techniques for storage and system replication, and our methodology can help any organization that is adopting this exciting new platform. Here’s how we did it:
System replication ensures the continuous operation of the software application itself. For most applications, replication means transferring the software onto a physical storage medium, such as a disk drive. But because SAP HANA exists entirely in memory, all data is transferred to the software layer. We have configured two active SAP HANA software installations for SAP CRM. The primary node functions as a “sender” of data, and the secondary node functions as a “receiver” of data. This method is known as synchronous in-memory log replication. The primary node commits a transaction after it receives a reply that the log was received by the secondary system. This keeps both nodes fully synchronous. In the event of a failure in the primary node, an administrator instructs the secondary mode to switch from live replication mode to full operation. The secondary system then becomes the primary system by replaying the last transaction logs, and then starts to accept queries.
Storage replication ensures the continuous availability of the software application’s core database. To accomplish this, we operate two geographically separate data centers that “mirror” one another through ongoing, real-time updates. If the primary data center becomes unavailable for any reason, a “takeover” is initiated to point SAP CRM to the secondary data center. All data is then uploaded into SAP HANA. Since the data centers are continuously mirrored, the databases are identical, and performance will not be compromised.
We conducted extensive end-to-end testing of our two replication methodologies, and they performed well in a variety of scenarios – including database switching, IP address switching, and hostname switching.