Guest blog written by Mary Kwan, HP Global Business Manager, SAP Solutions, Business Critical Systems
Since delivering the industry’s first SAP BW Enhanced Mixed Load (BW-EML) benchmark on HANA, HP and SAP have now taken another major step toward delivering an enterprise-class, in-memory, centralized database for all SAP applications. HP is the first technology partner to be certified by SAP for its HP AppSystems for SAP HANA portfolio with the disaster tolerant (DT) architecture. This is tremendously important for customers – particularly those who contemplate using HANA as the central database for their ERP system as well as their BW system and/or other analytical applications. It allows them to host a second HANA system – typically at a different site – that can take over for their production system in the event of a significant outage (for example, a power failure). With this certification, HP is now the only company in the world to deliver a certified DT solution for HANA.
HP’s DT solution makes use of an integrated data replication solution using the X9300 IBRIX Network Storage System and Continuous Access (CA) software, a feature of the HP P6000 series disk array included as part of the shared storage infrastructure in HP’s scale out HANA architecture. CA provides site-to-site replication for data written to the P6000 series disk array. The first solution certified by SAP makes use of CA in synchronous mode, which provides robust, metropolitan area, data replication for distances up to 50 kilometers or 30 miles. Testing is now underway for an asynchronous solution – also based on CA – that will provide data replication for data centers that are 500+ kilometers or 300+ miles. HP anticipates the certification for the asynchronous solution to follow soon.
In synchronous mode, CA offers a Recovery Point Objective (RPO) of zero. As HANA writes data (e.g. log file updates) to the P6000 series disk array, CA immediately mirrors the write to the disk array at the remote site. The data write is not marked as complete until it is complete at the remote site as well as the primary site. This means that the state of the shared storage at the primary and secondary sites is exactly the same. In the event of a failure at the primary site, only the in-flight transactions will be lost when resuming operations at the secondary site.
This solution also offers a Recovery Time Objective (RTO) that is very close to zero. When the system administrator declares that a disaster has occurred and initiates the failover, the HP Storage Management Server (SMS) changes the status of the replicated data volumes so that the servers at the remote site have read/write access. HP X9300 Network Storage Systems perform an orchestrated failover and begin presenting the data to the HANA servers at the remote site. At this point, the HANA database is able to perform a normal recovery from the latest savepoint, which should take very little time, since these are typically scheduled to occur every fifteen minutes.
Now that the certification is complete, HP is working to make the solution generally available to customers. To learn more, please contact your local HP Sales representative or visit us at www.hp.com/go/sap/hana. (See SAP note #1755396 for details regarding this SAP certification.)
For more details, click on the video link below and learn what HP's VP of Engineering, Martin Whittaker has to say about the new performance benchmark held by the HP AppSystem for SAP HANA and how it translates into tangible benefits for end users.