SAP HANA is the preferred database for all future SAP applications. Columnar databases and the in memory capabilities of SAP HANA make it an excellent platform for all SAP applications. Virtualized SAP HANA (SAP HANA) provides significant advantages over Physical HANA implementations by providing flexibility and agility in operating a HANA environment
SAP HANA environments have a large memory footprint with the majority of data in memory. The changes to the memory are constantly being replicated to disk and there can be significant disk activity on the system in spurts. In addition when the system is restarted or there is any high availability event there is a massive requirement for data that needs to be quickly loaded into memory from disk. Due to these reasons, there is a requirement in HANA to have a highly performant IO subsystem. All Flash Storage can be a great asset if used as shared storage for virtualized HANA implementations as they can provide excellent IO performance. This paper looks at the benefits of using virtualization that is backed by All Flash storage for SAP HANA.
Another challenge in SAP HANA environments is the large memory requirement for the HW. In the real world, not all data needs to be in memory. By reducing the amount of data resident in memory there are many potential benefits from a cost and efficiency perspective. SAP HANA since SPS09 has introduced dynamic tiering that can help optimize memory utilization and move the less actively used data to extended tables on disk. In addition SAP HANA SPS10 provides capabilities for multi-tenant capabilities combined with dynamic tiering.
The new paper Virtualizing SAP HANA leveraging All Flash Storage seeks to explore use cases for SAP HANA on All Flash storage to improve performance, optimize memory usage and increase RTO through efficient backup and recovery. We will explore the use of SAP HANA on All Flash storage for the following use cases:
A. Improved performance for data load into virtualized HANA.
B. Optimized write back Performance during regular operations.
C. Streamlined back and recovery for HANA data
D. Dynamic Tiering to reduce memory footprint with warm data on Flash