Primer Series: Intel® Optane™ Technology
In this article
Do you ever attend meetings where the conversation turns to an unfamiliar technology? Suddenly, acronyms are being thrown around, and you have no idea what they mean. Confused, you scan the room and notice that everyone else is nodding as if they know exactly what's being discussed.
We've all been there before. To help our valued customers make sense of it all, the engineers at WWT offer a continuing series of 'primer' articles to provide a basic understanding of emerging products and technologies. This article covers the fundamentals of Intel® Optane™ technology, a new tier in the memory and storage hierarchy to prevent bottlenecks cost-effectively — now and in the future, as data volumes and compute capabilities grow.
To understand what makes Intel Optane technology a game-changer in the modern data center, you should be aware of what's already upending the world of processors: an explosion in the compute capability built into a range of devices from edge to data center.
CPUs, FPGAs and other accelerators and NICs — anything with processing capability is seeing a dramatic rise in its number of cores. Compounding this issue is an increasing need to keep all that data as close to the processors as possible. Those burgeoning core counts are needed to process the immense volumes of streaming data anticipated soon, but the data needs to be "close" to the cores. DRAM density and NAND storage are not keeping up with that need.
But core count is only part of the picture: data centers also need a proportional growth in memory and storage performance to keep pace with the rising volume, speed and complexity of workloads — and therein lies the challenge. A significant performance gap between these technologies still exists. Real-time data exposes these gaps, and systems can't answer to those demands. DRAM density growth rate is decreasing, which will create significant memory bottlenecks and limit system performance.
Recognizing the need for a purpose-built storage and memory solution, Intel and WWT are actively promoting a path to the data center and edge coupling, along with modernization that introduces new tiers in the memory-storage hierarchy to improve overall system performance and reliability. Intel Optane technology is a game-changing technology purpose-built to address these needs in the modern data center.
Intel Optane technology bridges critical gaps in the storage and memory hierarchy, delivering persistent memory, large memory pools, fast caching and rapid storage acceleration. This revolutionary innovation enables organizations to optimize, store and move larger, more complicated data sets.
By fitting between the layers of memory and storage, Intel Optane technology keeps data closer to the CPU to eliminate bottlenecks before they happen and ensure that CPUs operate at peak performance and efficiency to enable faster analytics, timely insights and other benefits.
As core counts rise, the fabrics that tie systems together can become saturated. Keeping data closer to CPUs, FPGAs and other acceleration devices helps mitigate that effect to support faster performance.
Combined with the Intel® Xeon® Scalable processor platform, Intel Optane technology helps increase overall platform performance in the most dynamic environments by maintaining an optimum memory-to-core ratio even as core counts proliferate and accelerating larger, slower backend storage devices. This new technology — comprising Intel® Optane™ persistent memory and Intel® Optane™ SSDs — provides a cost-effective way to prepare data centers, edge and cloud applications to accommodate future generations of data processing demands.
An added benefit: Intel Optane SSDs work with all PCIe 3.0 NVMe compatible slots and Intel Optane persistent memory works with all Gold and Platinum 2nd Generation Intel® Xeon® Scalable processors.
Memory and storage exist on a continuum that can typically be measured by cost and latency, which usually have an inverse relationship: higher cost, less latency and vice versa. Intel Optane technology bridges the gap between capacity and performance on the memory/storage continuum. Whether this underlying media packaging is memory modules or SSDs, Intel Optane technology possesses several characteristics that make it versatile, reliable and desirable. Users can employ different configuration modes to achieve a range of benefits, including:
- Massively expanded total system memory capacities or lower total cost of ownership (TCO) at current system memory capacities.
- High virtual machine (VM) density, enabling more VMs per server and total server count consolidation.
- High application performance partially due to less need for disk access.
- New tiering possibilities from responsive storage and memory.
- Persistent memory for remarkably fast system recovery.
As mentioned earlier, Intel Optane technology consists of a pair of complementary solutions.
Intel Optane persistent memory is a highly significant advancement in-memory technology, cost-effectively combining large capacity and support for data persistence. It combines the most useful qualities of robust performance, excellent memory capacity and rapid-access storage to fuel business innovation.
Operating flexibly in Memory Mode or App Direct Mode (or both), Intel Optane persistent memory enables denser, more consolidated virtualized server environments, extending memory beyond DRAM limits. Intel Optane persistent memory offers up to 512 GB per DIMM, versus DRAM's peak of 128 GB per DIMM. And it can operate in a persistent mode, storing data even when powered down with enabled applications. Although DRAM is faster, Intel Optane persistent memory delivers a lower cost per GB; this, combined with the higher capacity and reliability, results in a superior TCO.
Another key feature: Intel Optane technology promotes memory security through automatic hardware-level encryption.
This memory solution requires a system powered by the 2nd Generation Intel Xeon Scalable processor — ideal for heavy read applications where extreme endurance and low latency can contribute to a sound return on investment. These include in-memory database, virtualization and high performance computing (HPC), as well as AI, big data and analytics.
The Intel Optane Solid State Drive (SSD) is a block storage device that can improve overall performance while helping eliminate storage bottlenecks to allow larger data sets. These SSDs can accelerate applications, reduce transaction costs for latency-sensitive workloads and improve overall data center TCO. Even though its capacities might be smaller than those offered by NAND SSDs, the Intel Optane SSD can boost overall system performance when accelerating, caching and/or tiering data.
It does this by offering high throughput, low latency, extreme endurance and consistent quality of service (QoS). These characteristics make Intel Optane SSDs well-suited for handling even the most demanding workloads in hyper-converged infrastructures, high performance computing and large-scale database environments.
Think of the difference between the two Intel Optane technology ingredients this way: when you need to improve your memory footprint, consider Intel Optane persistent memory. If storage is the bottleneck, go with the Intel Optane SSD. For the best of both worlds, the combination of Intel Optane persistent memory and SSDs can break through the storage bottleneck and offer a great memory footprint, which in turn increases CPU utilization and overall system performance.
Nowhere is the ability of these two technologies working together to break memory and storage bottlenecks more evident than in the virtual desktop infrastructure (VDI). In a VDI environment, this "better together" solution leads to more VMs per node, larger memory VMs and more consistent VM performance due to the low latency, great QoS and high performance of Intel Optane technology. Also, the Intel Optane SSD reduces the risk of wearing out a cache drive in a virtualized environment due to the very high endurance it offers.
The introduction of Intel Optane persistent memory and Intel Optane SSDs to a VDI environment enabled up to 87 percent more VDI users, reducing the per-user cost by 16 percent. This efficacy is due to the increased memory capacities offered by Intel Optane persistent memory and the performance consistency of the Intel Optane SSD.
In particular, using Intel Optane SSDs as a cache layer in a VDI environment delivered consistent response time and user experience. This performance improvement is noticeable, especially during a "boot storm" at the beginning of the day, as employees all arrive and boot up almost simultaneously. In contrast, a NAND SSD might be unable to deliver consistent performance, bogging down productivity at the start of the workday.
When added to the virtual desktop infrastructure, Intel Optane technology massively expands total system memory capacities, enabling either more VMs per server or a consolidated total server count.
This has been a brief introduction to a pair of memory and storage technologies that are enabling the explosive growth in computing capabilities that will be needed to extract insight from tomorrow's immense data volumes. Check out more resources below to continue your dive into the Intel Optane technology story: