3D XPoint: A Guide To The Future Of Storage-Class Memory

The IMFT Sideshow

Business makes for strange bedfellows, and the IMFT relationship is no exception. Both companies have their strengths: Micron brings the manufacturing prowess, Intel brings the engineering talent; Intel is the glitzy showman, while Micron (the Boys in Boise) is known for its plodding, reserved approach.

The differences shine through in how each company is talking about 3D XPoint publicly. Micron has been much more conservative, and Intel has been bullish. To Micron's credit, it didn't come forward with any technological demonstrations until there was actually something of gravity to show, and the company has been frank in our conversations, saying that it should have waited to introduce 3D XPoint. Meanwhile, Intel was out being, well, Intel, with its full court PR press. 

Micron and Intel cooperate on 3D XPoint development, but the two will compete in the open market. However, contractual obligations limit Micron to selling storage products for the first "few" years, while Intel can sell 3D XPoint for both storage and memory applications. The arrangement makes some sense given that the memory side of the equation will require platform-level support, which is Intel's forte with its control of the chipset and IMC. Micron can hit the ground running with plug-and-play, block-based NVMe storage, which will likely move more units in the early days. Micron already has its own NVDIMM initiative well underway, which probably serves as its future platform for 3D XPoint derivatives, so it will enter the memory segment later, when the technology has matured.

Micron will focus on the high-margin enterprise market, because the company doesn't believe profit will spring from the client market due to the high initial price point. However, it will benefit from the scale that Intel builds on the client side. Intel doesn't sell standalone NAND chips, so the company will only provide finished Optane products for both client and data center markets.

Micron has an entirely different go-to-market strategy for its storage devices. The company refuses to name the companies it works with but it's not difficult for an industry watcher to identify them. Micron will sell its own drives, but "a partner" will co-develop the drive and offer the same product, down to the hardware and firmware, under a different name. Micron and Seagate already have a strategic alliance that finds the companies selling the same products with different branding, which is an easily identified, unique arrangement. Seagate will sell its own branded 3D XPoint products.

The second partner isn't as easy to ferret out. Micron indicates that it would sell the products to a company that "uses FPGAs on its own custom modules" and has "reinvented itself several times." Micron will sell this company 3D XPoint packages, but will not manufacture the finished modules. Micron also indicates that it already has an existing NAND partnership with this company that powers the partners' current modules. Micron sells IBM its Fortis Flash NAND directly for custom FPGA-powered modules for the FlashSystem series, which we have covered extensively. The apparent partnership with IBM makes good technical sense, because it will require intense engineering talent to extract the underlying performance advantages (which IBM has in spades). Micron indicated "the partner" would use 3D XPoint across a broad section of its portfolio.

Micron's willingness to sell the modules as standalone chips is an important concept. Micron and Intel stated during the 3D XPoint launch that they would not license the technology to others. This statement is likely only confined to the actual manufacturing of the media. Neither company had acknowledged that it would sell the discreet components (i.e., standalone 3D XPoint chips) that would enable a third-party ecosystem, much as we see with the current SSD market.

Industry sources indicate that SMI, which is the dominant third-party SSD vendor, is developing the SSD controllers for Intel's Optane SSDs. Intel already has a strong partnership with SMI, and as we noted, it would take intense developmental efforts to engineer 3D XPoint SSDs. The complexity will put 3D XPoint SSDs out of reach of smaller players such as Kingston, Corsair, and Patriot. However, SMI may be in a position to enable the smaller players with its controllers, much as it already does with NAND-based SSDs. A third-party ecosystem would likely emerge in a few years, as Micron will undoubtedly focus on selling to key high-margin enterprise partners until it has excess capacity.

The IMFT relationship isn't all crimson and roses; Intel recently built its Dalian "Fab 86" in China to produce both 3D NAND and 3D XPoint, which marks its return to memory manufacturing (sans NOR) for the first time in 31 years. The development changes the dynamic of the IMFT partnership, which many predict will end when it comes up for renewal. Interestingly, industry sources indicate that Intel is licensing some aspects of Micron's manufacturing technology for the Dalian excursion. 

Intel's CEO stated that he expects 3D XPoint volume in the DIMM segment. Both companies stated that they don't expect 3D XPoint to cannibalize the DRAM market, which would expose the DRAM-heavy Micron to losses in one of its most important segments. 3D XPoint may not kill DRAM, but it clearly would replace some of it. Micron's CEO has commented that he expects 3D XPoint production to be close to its DRAM volume in two years. There are signs of Micron's preparation for the inevitable. Micron has its as-of-yet unfinished Inotera acquisition in the cards, which should provide greater flexibility to convert some of its DRAM production output to 3D XPoint in the near future.


MORE: Best Enterprise SSDs


MORE: How We Test Enterprise SSDs

Paul Alcorn
Managing Editor: News and Emerging Tech

Paul Alcorn is the Managing Editor: News and Emerging Tech for Tom's Hardware US. He also writes news and reviews on CPUs, storage, and enterprise hardware.

  • coolitic
    the 3 months for data center and 1 year for consumer nand is an old statistic, and even then it's supposed to apply to drives that have surpassed their endurance rating.
    Reply
  • PaulAlcorn
    double post

    Reply
  • PaulAlcorn
    18916331 said:
    the 3 months for data center and 1 year for consumer nand is an old statistic, and even then it's supposed to apply to drives that have surpassed their endurance rating.

    Yes, that is data retention after the endurance rating is expired, and it is also contingent upon the temperature that the SSD was used at, and the temp during the power-off storage window (40C enterprise, 30C Client). These are the basic rules by which retention is measured (the definition of SSD data retention, as it were), but admittedly, most readers will not know the nitty gritty details.

    However, I was unaware that JEDEC specification for data retention has changed, do you have a source for the new JEDEC specification?

    Reply
  • stairmand
    Replacing RAM with a permanent storage would simply revolutionise computing. No more loading an OS, no more booting, no loading data, instant searches of your entire PC for any type of data, no paging. Could easily be the biggest advance in 30 years.
    Reply
  • InvalidError
    18917236 said:
    Replacing RAM with a permanent storage would simply revolutionise computing. No more loading an OS, no more booting, no loading data
    You don't need X-point to do that: since Windows 95 and ATX, you can simply put your PC in Standby. I haven't had to reboot my PC more often than every couple of months for updates in ~20 years.
    Reply
  • Kewlx25
    18918642 said:
    18917236 said:
    Replacing RAM with a permanent storage would simply revolutionise computing. No more loading an OS, no more booting, no loading data
    You don't need X-point to do that: since Windows 95 and ATX, you can simply put your PC in Standby. I haven't had to reboot my PC more often than every couple of months for updates in ~20 years.

    Remove your harddrive and let me know how that goes. The notion of "loading" is a concept of reading from your HD into your memory and initializing a program. So goodbye to all forms of "loading".
    Reply
  • hannibal
    The Main thing with this technology is that we can not afford it, untill Many years has passesd from the time it comes to market. But, yes, interesting product that can change Many things.
    Reply
  • _TheD0ct0r_
    10 years later... still unavailable/costs 10k
    Reply
  • TerryLaze
    18922543 said:
    10 years later... still unavailable/costs 10k
    Sure you won't be able to afford a 3Tb+ drive in even 10 years,but a 128/256Gb one just for windows and a few games will be affordable if expensive even in a couple of years.
    Reply
  • zodiacfml
    I dont understand the need to make it work as DRAM replaement. It doesnt have to. A system might only need a small amount RAm then a large 3D xpoint pool.

    The bottleneck is thr interface. There is no faster interface available except DIMM. We use the DIMM interface but make it appear as storage to the OS. Simple.

    It will require a new chipset and board though where Intel has the control. We should see two DIMM groups next to each other, they differ mechanically but the same pin count.
    Reply