New memory offers huge capacities and persistence, but fits in a DDR4 slot.
Ever since Intel and Micron announced 3D XPoint memory in 2015, the world has been waiting for the companies to use it to build memory sticks.
3D XPoint blends the properties of flash storage and DRAM memory. Like flash, it’s persistent, retaining its value even when systems are powered down, and it’s dense, with about 10 times the density of DRAM. Like DRAM, it supports low latency random access. Intel also claimed that its write endurance is substantially better than that of flash.
This combination of features created the prospect of memory sticks that look like DIMMs and appear to the system as if they’re DDR4 RAM but with much greater capacities and with persistence: data written to "RAM" is retained permanently. Memory with these properties is exciting for a wide range of applications—for example, databases that no longer need to concern themselves with flushing data back to disk—and might one day provoke significant changes in the way operating systems and software are designed.
But while persistent memory was perhaps the most interesting application of 3D XPoint, the first products to hit the market were simply storage drives using "Optane" as their branding. There was a series of drives for enterprise customers and some consumer-oriented M.2 sticks designed to be paired with a spinning disk to produce a high-speed hybrid. While 3D XPoint did offer some benefits over flash SSDs—in particular, the latency of the drives is significantly lower than that of comparable flash units, and the I/O performance is sustained even under heavy mixed read/write workloads—this wasn’t quite the revolution that we were hoping for.
No longer. Today, Intel announced Intel Optane DC Persistent Memory. This is a series of DDR4 memory sticks (with capacities of 128GB, 256GB, and 512GB) that use 3D XPoint instead of traditional DRAM cells. Result? The latency is a bit worse than real DDR 4, but the sticks are persistent. Although they use the standard DDR4 form factor, they’ll only be supported on Intel’s next-generation Xeon platform.
Intel is pitching the new memory as a way to greatly increase the amount of memory available to processors and eliminating the latency that normally occurs when moving data from memory to persistent storage. This is valuable to a range of database-like and caching workloads. The persistence means that freshly booted servers no longer need to load terabytes of data into memory—the data is, in effect, already there. Because persistent memory has such big implications for software developers, Intel will also have a scheme that gives developers (under NDA) remote access to machines using Optane Persistent Memory so that they can develop and test software that takes advantage of its persistent capabilities.
Beyond this basic information, there’s still a lot we don’t know about Optane DC Persistent Memory: performance, endurance, power consumption, system/processor compatibility—all remain unknown at this point. Intel is also vague on the product’s availability: wide availability is going to happen some time in 2019, but selected customers will be able to get their hands on it this year.