Servers & Systems: The Right Compute
1752297 Members
4488 Online
108786 Solutions
New Article
ComputeExperts

How to gain faster database load and restore for SAP HANA with HPE Superdome Flex

Learn how to achieve faster load and restore times for SAP HANA with HPE Superdome Flex Server using HPE Persistent Memory kit featuring Intel® Optane™ DC persistent memory.

HPE Superdome Flex-SAP HANA-Persistent Memory.png

Traditionally SAP HANA uses volatile memory and synchronizes it with the persistence level on disk using mechanism like save points. One disadvantage to this approach: During an SAP HANA restart all data must be reloaded from disk into memory and it can take a few hours until all data is back in memory. This can impact business-critical data which is processed and analyzed at very high speed.

Imagine your systems are utilized as disaster recovery or an all-DRAM system. If a server needs to be restarted, you must wait for data to load from storage before accessing SAP HANA—which can be over an hour for large databases. Time lost for data to load can take hours. IT will need more time to bring the up the system and users will experience drop is application speed.

The major concerns boil down to this: The time taken by SAP HANA database to restart and load data back into memory. To address this, we would like to introduce HPE Superdome Flex Servers with SAP HANA in-memory database with next-generation HPE Persistent Memory kit featuring Intel® Optane™ DC persistent memory.

HPE Persistent Memory and HPE Superdome Flex Server solution details

With the combination of HPE Persistent Memory and HPE Superdome Flex Server solution, not only the SAP HANA restart times can shrink to minutes but also bring up of the applications will be reduced from hours to minutes. This reduction of downtime will increase the performance when compared to an all-DRAM system.

As shown in the Figure 1, HPE Persistent memory will be used for the main storage of column store table that is typically the most important factor for data space consumption in SAP HANA environments. Other aspects like delta storage, caches, intermediate result sets or row store remain in Dynamic RAM (DRAM).

In other words, SAP HANA knows which data structures benefit most from persistent memory. SAP HANA automatically detects persistent memory hardware and adjusts itself by placing these data structures on persistent memory, while all others remain in DRAM.

Figure 1: Contents in HPE Persistent Memory and DRAMFigure 1: Contents in HPE Persistent Memory and DRAMHPE Persistent Memory Module

Persistent memory possesses properties of both dynamic random-access memory (DRAM) and storage, but it doesn’t replace DRAM. As shown in Figure 1, it can be viewed as a tier between DRAM and flash storage. What really makes HPE Persistent Memory technology stand out is that it is fast enough for the processor to access directly without stopping for the block I/O required for traditional storage.

Figure 2: HPE Persistent Memory and remaining memory types by tierFigure 2: HPE Persistent Memory and remaining memory types by tier

HPE Persistent Memory modules are available in following sizes with this solution and supports maximum of 1.5TB per processor socket persistent memory capacity:

  • HPE 128 GB 2666 Persistent Memory Kit featuring Intel® Optane™ DC persistent memory
  • HPE 256 GB 2666 Persistent Memory Kit featuring Intel® Optane™ DC persistent memory

Figure 3: HPE Persistent Memory ModuleFigure 3: HPE Persistent Memory Module

What is DRAM to HPE Persistent Memory Ratio?

HPE Persistent Memory modules will be installed with DRAM DIMMs in the same system. HPE Persistent Memory modules do not function without any DRAM DIMMs installed. Different sized HPE Persistent Memory modules and DRAM DIMMs can be used together as long as supported ratios are maintained.

For example for a 1:1 ratio on each 2-slot channel, a single 128GB DDR4 DIMM is matched with a single 128GB HPE Persistent Memory module. For 1:2 ratio on each 2-slot channel, a single 128GB DDR4 DIMM is matched with a single 256GB HPE Persistent Memory module.

SAP HANA currently supports various capacity ratios which include:

  • 1:1 ratio
  • 1:2 ratio
  • 1:4 ratio
  • 1:8 ratio

Key performance benefits, starting with accelerated start-up times

In conclusion, the key performance benefit is in greatly accelerated start-up times. This ensures that data can be quickly reloaded into memory and any system downtime can be minimized. The solution offers the additional higher memory capacities per server which allows to run larger databases. It also provides you an exceptional advantage of increased CPU scalability with HPE Superdome Flex Servers, ranging from 4 sockets to 12 sockets.

The need of replacing a system for outgrowing capacity can be avoided, meaning HPE Superdome Flex Servers can start small and scale up as the needs increase.

Other benefits offered include faster insights with more data managed in-memory, greater efficiency and increased business agility.

For more details on solution configurations and supported CPU and Memory sizes, please refer to the Solution Architecture Document at HPE Solution Architecture for SAP HANA on HPE Superdome Flex solutions with D3710 Direct Attached Storage using Cascade Lake ProcessorsAlso check out the HPE Superdome Flex Quick Specs.

For more details, contact your Hewlett Packard Enterprise representative.


Meet Compute Experts blogger Badri Prasad DS, Software Specialist, HPE Solution Engineering team

Badri Prasad-HPE.pngBadri is mainly focused on creating solutions and certifications for enterprise use cases around SAP and/or SAP HANA technologies, besides being a technology evangelist, he is also an automobiles enthusiast and loves to travel around the world and capture the beauty of nature in the lens.

 

Compute Experts
Hewlett Packard Enterprise

twitter.com/hpe_compute
linkedin.com/showcase/hpe-servers-and-systems/
hpe.com/servers

 

 

 

 

About the Author

ComputeExperts

Our team of Hewlett Packard Enterprise server experts helps you to dive deep into relevant infrastructure topics.