Increasing memory capacity for critical enterprise workloads in the Microsoft Azure cloud can help enterprises manage the high data volumes needed for a range of applications including artificial intelligence (AI) and machine learning-based applications, as well as analyzing growing data from edge devices. But with the size limitations of DRAM-based memory, how can Azure engineers and solution architects build solutions that help customers streamline and accelerate workload performance while achieving a lower total cost of ownership (TCO)?
For SAP customers migrating from on-premises to a cloud-based instance on Microsoft Azure, Intel® Optane™ persistent memory (PMem) is ready and able to help address these challenges. Together, Intel, SAP and Microsoft have developed new SAP HANA Large Instances with Intel Optane PMem and the performance benefits and power of 3rd generation Intel® Xeon® Scalable processors. However, the potential of Intel Optane PMem extends beyond SAP workloads. It can be built into a host of new cloud solutions, and there are three sets of opportunities it creates for Azure engineers developing new solutions for customers.
Once you’ve read this article, if there are solutions you’re developing that could benefit from Intel Optane PMem, get in touch with your Intel account team to arrange a demo.
1. Lower TCO with Higher Memory Density
Optimizing the Microsoft Azure cloud to provide a lower TCO for customers on SAP workloads requires a co-engineered solution that delivers flexibility. SAP HANA Large Instances with Intel Optane PMem dramatically increase memory density compared to DRAM, which is fast but expensive, and has low density compared to other storage technologies.
As a result, with new SAP HANA instances the smaller DRAM-only node at the primary site can replicate the data into a larger Intel Optane PMem node, offered in 1:2 and 1:4 ratios. The data is then preloaded in persistent memory. In the event of a primary site failover, the higher-density Intel Optane PMem node can act as the primary node. This means siloed instances for quality assurance and disaster recovery are no longer needed, saving Azure customers money. As the data is also preloaded into the Intel Optane PMem node, data does not need to be loaded from disks and downtime is also reduced – further improving recovery times1.
To quantify the potential of these innovations, testing was conducted using two S224 HANA Large Instances. The first was a DRAM-only system running 6 TB of memory, while the second was a system with 9 TB of memory consisting of 3 TB DRAM and 6 TB of Intel Optane PMem in a 1:2 ratio. The data reload time on the Intel Optane PMem-enabled instance was 22 times faster as compared to the reload time on the DRAM-only system. The load time on the DRAM system post reboot is around 44 minutes, versus 2 minutes on the Intel Optane PMem node1.
2. Improve Latency
A key consideration for cloud architects and engineers building solutions for customers is the strain that growing datasets and higher-velocity data are placing on traditional data-deployment models. With increased memory density and near-DRAM speeds, Intel Optane PMem makes low-latency access to larger data volumes a reality for innovative Internet of Things (IoT), AI and machine learning-based use cases2. This means that less data needs to be stored elsewhere, such as in flash-based and spinning disk-based storage – both of these storage solutions have slower response times.
3. Simplify Complex Architectures
With more memory available, smaller DRAM-only SAP HANA nodes can be consolidated. Transactional and analytics workloads can also be combined onto single nodes. By reducing deployment complexity with Intel Optane PMem, organizations have realized 39 percent cost savings per database terabyte2. Both of these benefits simplify SAP HANA architectures while further reducing TCO. As databases grow, servers that support Intel Optane PMem can be upgraded with higher-capacity memory modules to accommodate growing data volumes. Data-tier management can also be simplified, helping to lower infrastructure costs.
In addition, SAP HANA backup, recovery and system replication operate in the same way with Intel Optane PMem. Backups save all persistence volume data without accessing persistent memory. This means data can then be restored to the persistence volume. Similarly, system replication can be used in a mixed environment with systems configured with or without persistent memory.
Contact your Intel account team to arrange a demo for your engineering team, and to discover more about how the solutions you’re building could benefit from Intel® Optane™ technology on Microsoft Azure and Intel.
Or find out more about how Microsoft Azure and Intel are collaborating to deliver a high-performance and flexible cloud to meet all customer needs.
Intel® Optane™ persistent memory