Intel and Micron just announced 3D XPoint, a new memory technology that blends the speed of DRAM with the storage capabilities of Flash. Given that the technology won’t be shipping for a bit, you’ll have time to cycle out your current tech devices. But much like Flash memory made nearly everything that used a hard drive obsolete, this effectively makes many things that use DRAM or Flash obsolete.
And much like Flash did, this technology will move where performance in storage has high value and can support a higher cost. Think real-time analytics, financial services, defense (weapons and analytics), real-time translation and AI.
In a way, this should follow a path similar to Flash as it moves through the market, and it may render HP’s somewhat similar Memristor technology obsolete before it even ships.
Let’s talk about that.
Intel and Micron represent that this new memory technology is 1,000 times faster than Flash, which is currently used in high-performance systems. That is a massive increase in performance and the companies indicated that this comes without a significant increase in energy consumption, which keeps densities up and heat down. Given that Flash arrays are used in high-end storage systems and continue to be driven by industries that are willing to pay almost anything for performance increases, this should dramatically change the storage landscape.
EMC, which has a very tight and close relationship with Intel, should be the most obvious beneficiary at the high end. HP, which has a competing technology in the Memristor, is the most likely to be damaged, with others falling between the two extremes.
Buyers of technology in this class buy strategically and will likely begin gravitating to the firm that they think will have the greatest advantage when the products hit market, and the market pivots to optimize around them. So even though actual products are months out, this announcement may force some really hard decisions for all of the major vendors as they realize that whichever company is last to market with this technology may lose entire market segments to competitors that are more aggressive.
As with any technology, there will be a learning curve; however, vendors won’t be able to simply swap out Flash for 3D XPoint. With this level of performance increase, optimizing the resulting hardware so it doesn’t badly bottleneck will take time.
Personal Devices to Self-Driving Cars and Robotics
Just as handheld devices moved from magnetic drives to Flash over a decade ago, 3D XPoint promises to do the same thing. The result should enable entirely new classes of apps that run locally, like real-time translators, or far more intelligent (and off-line capable) versions of Siri and Cortana. But fast, in-memory processing is critical to this new wave of ever-more intelligent systems. Obvious applications include autonomous robotics and self-driving cars. Both require a massive amount of in-memory processing to operate in the real world and the performance increase promised by this new technology should make these systems for more capable far sooner than otherwise would have been the case.
Initial implementations are likely too far down the development path to make this switch easily, but by the end of the decade we should see a sharp increase in overall performance and accuracy. The next generation should aggressively move to this technology, and given that it is the second generation that is expected to mainstream, it is timed ideally for the market spike.
Wrapping Up: Memory over Processors
We often focus so much on processor performance that we forget that systems often bottleneck in other areas, and memory performance is more important because that is where the bottlenecks often reside. Upping that performance 1,000 times promises a massive change in anything currently limited by the performance of Flash memory. This technology is going to do great things for firms closely aligned with Intel and Micron and really ugly things for companies that are trying to bring to market a competing technology.
As with most technology changes of this magnitude, we can’t estimate how big the overall impact will be or all of the things this will enable, but the word “massive” is likely inadequate when it comes to just how big an impact it will have. This is a once-in-a-decade kind of event, and given the rate of change we are seeing, likely not the only one. It is pretty clear that the end of this decade isn’t going to look anything like the beginning.
Rob Enderle is President and Principal Analyst of the Enderle Group, a forward-looking emerging technology advisory firm. With over 30 years’ experience in emerging technologies, he has provided regional and global companies with guidance in how to better target customer needs; create new business opportunities; anticipate technology changes; select vendors and products; and present their products in the best possible light. Rob covers the technology industry broadly. Before founding the Enderle Group, Rob was the Senior Research Fellow for Forrester Research and the Giga Information Group, and held senior positions at IBM and ROLM. Follow Rob on Twitter @enderle, on Facebook and on Google+