For many applications running in today’s data centers, performance is impeded by I/O bottlenecks that slow the movement of data to and from the processor. While IT leaders understand the problem well, the solution has been elusive. Ideally, they would put more data into memory to close the data-to-processor gap, but that hasn’t been a viable solution because memory has historically been small, volatile, and expensive.
Today, this story is about to change in dramatic ways, thanks to a new generation of Intel DIMMS coming next year from Intel that will make memory big, affordable, and persistent. This new generation of Intel persistent memory, based on the groundbreaking 3D XPoint™ media developed by Intel and Micron, will enable data center operators and application developers to overcome a historical barrier to higher application capacity and performance.
Intel persistent memory will allow users to improve system performance dramatically by putting more data closer to the processor on nonvolatile media, and do it in an affordable manner. This will truly be a game-changer when it comes to the way applications and systems are designed. It’s a complete reversal of 50 years of the small, volatile, and expensive memory IT orthodoxy.
Intel persistent memory is in the spotlight this week at the SAP Sapphire conference in Orlando, where Intel’s Lisa Davis, Vice President of IT Transformation, presented the first public demo of the technology. The demo shows SAP’s HANA in-memory data analytics platform working with Intel persistent memory. It’s fitting that an SAP application should be showcased in the public debut of the new DIMMs because Intel and SAP have partnered on a persistent memory vision for many years, representing Intel’s commitment to delivering industry innovation and building upon the industry leading solution performance we continue to deliver to end customers today.
Big, affordable, persistent memory will be a game-changer for users of SAP HANA and other in-memory databases. More data in-memory equates to better, faster insights and more business velocity. And because data remains in-memory through power cycles, restart times are a fraction of the loading times from disk, for more HANA service uptime.
Of course, it’s not just in-memory databases that will reap the benefit of Intel persistent memory when it becomes available next year. The new DIMMs enable countless use cases across data center segments, from mass-scale virtualization and private clouds in the enterprise to, cloud hosting, and search offerings in public cloud environments. The new technology will also help accelerate the performance and capacity of high performance computing (HPC) applications and software-defined storage. The list of potential use cases goes on and on.
So when can you get it? Intel expects to ship its persistent memory solution in 2018 with a processor refresh of the Intel® Xeon® Scalable family platform, code-named Cascade Lake. And if you’re a developer who wants to get your application ready for Intel persistent memory, check out: http://pmem.io/ and https://youtu.be/p2aXZ_bmVV8 for an Under the Hood look at the Non-Volatile Memory Libraries.