Linux and Hana/Persistent Memory in the Data Center
It is in the nature of things that SAP end users tend to focus less on the infrastructure technologies used or on system operation. Although they benefit time and again from the constant further developments inherent there.
In particular, the computer or chip/processor technologies used play quite a weighty role in SAP usage. This has been the case in the past and continues to be the case.
With the Intel Optane DC Persistent Memory technology for SAP Hana workloads, a new chapter is now being opened in the area of memory. It is based on NVDIMM technology, which, in simple terms, makes it possible to move large amounts of data permanently and persistently closer to the processor.
At the same time, latency times are minimized when retrieving data from the main memory. NVDIMM stands for Non-Volatile Inline Memory Module; for a main memory where contents do not volatilize even in a powerless state.
SAP Hana as an in-memory database and SAP Hana-based applications such as S/4 derive particular benefits from this. You could say: Persistent Memory is made for Hana. SAP itself considers the possibility of providing Persistent Memory for Hana to be nothing more, but also nothing less than a "milestone".
SAP, Intel and Suse hand in hand
Suse is the first Linux distributor and SAP open source companion to fully support Intel Optane DC Persistent Memory in conjunction with Hana with its SLES for SAP Applications operating system platform since mid-January.
Whereby the abbreviation DC means Data Center and thus Persistent Memory is designed for use in data centers. This once again underlines the close and deep partnering between SAP, Intel and Suse in mission-critical SAP deployment.
For SAP customers, persistent memory technology enables new or optimized applications for data access and data storage compared to traditional technology (using SSDs). In the end, this means nothing other than that the business can derive even more monetary benefits from the use of Hana.
One aspect, for example, is the use of cost-effective in-memory database solutions with large data volumes. Another is the flexible definition of "warm" and "hot" data. For example, in SAP analytics or SAP Big Data applications.
Advantages also with cloud use
Persistent Memory also allows in-memory load times to be significantly reduced or recovery times to be minimized. In realistic tests, for example, the startup/load time for a six TB SAP Hana database with Persistent Memory was drastically reduced compared to a traditional system (with SSD). Namely, from 50 minutes to four minutes.
In addition, persistent memory enables higher memory capacities (greater than three TB per CPU) and thus lower total cost of ownership (TCO).
Both SAP Hana on-premise customers and SAP Hana public cloud service providers, such as Microsoft with Azure, Amazon with AWS or Google with the Google Cloud Platform, as well as the SAP Cloud Platform (SCP), benefit from the manifold advantages of Persistent Memory.
Intel Optane DC Persistent Memory in conjunction with Hana is supported by the latest Intel Xeon processors. Hana as of version 2.3 and Suse SLES for SAP application-side as of version 12, Service Pack 4.