The death of RAM. (Intel and Micron introduce the future of high-speed storage)

Table of Contents

The death of RAM. (Intel and Micron introduce the future of high-speed storage) .As a huge jump in the race for “Mega Storage” Intel and Micron have unveiled the biggest development in storage technology for decades. 3D XPoint is an innovative memory technology that promises transfer speeds 1,000 times faster than current SSDs. This is the largest performance leap in-memory technology since the advent of NAND flash in 1989. And Intel claims that 3D XPoint (pronounced “cross-point”) will not only be a thousand times faster than current technology, it will also offer 1,000 times the write endurance of today’s NAND-based drives.

Say goodbye to RAM

Traditional computer architectures rely on a combination of high-capacity mass storage, such as hard disks and SSDs, and much faster volatile RAM, for working with data at the highest possible speeds.
The new 3D XPoint technology is non-volatile – so data isn’t lost when you pull the power plug – and offers a storage density eight to ten times greater than that of DRAM. Although latency is very slightly higher than current DRAM, Intel suggests that faster speeds aren’t far off. The implications of this are huge: it means the same technology could serve as both storage and memory. In the future, once 3D Xpoint becomes commonplace, the concept of discrete system RAM may become a thing of the past.

What’s driving this enormous leap in memory? The first answer is Big Data. Really Big Data. Introducing 3D XPoint, Intel said that as the pace of data creation ramps up – from 4.4 zettabytes of digital data created in 2013 to an expected 44 zettabytes by 2020 – new technology is needed to ensure that information can be accessed quickly and stored easily and cheaply. 3D XPoint aims to address that need. The tech giant proposed a gamut of potential applications, from high-fidelity pattern recognition in the financial sector to enabling healthcare researchers to analyse large data sets in real-time. Current Big Data applications require data centres with huge and prohibitively expensive amounts of RAM, so 3D XPoint has the potential to drive down the cost of data analysis across a wide range of services.

But it’s not just about large-scale computing. Another driving factor is the need for affordable, high-speed storage technologies for the Internet of Things. The emerging network of tiny connected devices will need to store and transfer large quantities of data, potentially on city-wide scales. 3D XPoint could be the technology that makes that feasible.

How does XPoint Technology work?

It might sound highly unlikely, but Intel describes 3D XPoint technology as ‘simple’.
As you can see from the picture below, a 3D XPoint module has a mesh-type structure. Wires arranged in perpendicular orientations connect submicroscopic columns; unlike DRAM, which requires a transistor to be attached to each memory cell,3D XPoint uses ‘selectors’ to read or write memory locations depending on how much voltage is applied.

These grid-like layers can then be stacked, which is how Intel and Micron have managed to increase the storage density compared to DRAM. In the first prototypes of the technology, 128 billion memory cells each store a single bit of data; two stacked memory layers in combination can store 128GB per die. For enterprise applications, meanwhile, it’s the technology’s claims of increased endurance which might be really significant. Where a traditional SSD based on NAND flash might be rated for a write-cycle of 40GB per day for five years, a 3D XPoint-based SSD could potentially write 40TB a day. That’s a lot of data.

Intel and Micron are currently in the process of developing individual products based on 3D XPoint technology and will be rolling the first samples out to ‘select customers’ later this year. The technology will probably take several years to trickle down into the mainstream, however. Thinking of holding off for that 3D XPoint-based MacBook? It’s likely you’ll be waiting a little while yet.

Written and Submitted by: Suman Gautam