Solid-state storage continues on a path toward greater speed and higher density, leading to an increasingly diverse set of solutions that, frankly, is starting to trend toward the unusual.
But overall, the goal appears to be expanding capacity while boosting I/O to allow the enterprise to support ever-expanding workloads on smaller and smaller hardware footprints.
At this week’s Flash Memory Summit, Intel caused a buzz with a preview of its new 1 PB SSD that it dubbed the Ruler because, well, that’s what it looks like. Although the company hasn’t released precise dimensions yet, the device is about the width of a single rack unit, which means it can provide enough storage as four units’ worth of 3.5-inch drives. And to whet the appetite even further, Intel says the drive will support both 3D NAND and the new Optane Memory system to provide rapid access to all that data.
Meanwhile, other vendors are looking to capitalize on the new Non-Volatile Memory Express (NVMe) interface and other technologies to boost SSD performance for the extreme workloads emerging in Big Data and the IoT. ITBE’s Mike Vizard reports that Seagate is out with the 2 TB Nytro 5000 that couples NVMe with the new M.2 form factor to deliver 67,000 IOPS for random writes, along with the 15 TB Nytro 3000 that can hit 2,200 MBps in sequential read using a 12 Gbps SAS interface. This is intended to provide a smooth upgrade path as users start to swap out the earlier SAS drives for the more advanced NVMe format.
Toshiba is on pretty much the same track with its new 64-layer 3D TLC NAND devices, says AnandTech’s Billy Tallis. The company showed off the 30 TB PM5 SAS device and the 15 TB CM5 NVMe drive, both of which feature a common controller that supports both interfaces. While NVMe is the faster format, Toshiba still sees a lot to like in SAS thanks to new technologies like the four-port MultiLink system that delivers sequential transfer speeds of up to 3,350 MBps. The NVMe drive offers twice that performance but has less capacity, giving the enterprise options when it comes to matching workloads with storage architectures.
For the future, we can expect form factors to grow ever smaller while I/O gets faster. Samsung took the wraps off a 1 Tb V-NAND chip that it will pitch to commercial SSD vendors in the coming months. The company says it can stack multiple 16 Tb dies on a single drive to give it a memory capacity of 2 TB. The company is also already looking past the new M.2 form factor with the new 16 TB NGSFF drive that measures barely 30mm x 110mm x 4.3 mm, essentially providing about four times the capacity of a typical M.2 device. Samsung has already delivered a server reference architecture that packs 36 drives into a 1U rack for a total capacity of 536 TB.
With all of these options about to enter the channel, enterprise storage executives will be tasked with a fair amount of research over the coming months as they define their next-generation architectures. While it seems clear that higher-density, higher-speed infrastructure is in order, there are enough nuances in the available offerings to craft highly customizable solutions for key workloads.
Defining those needs now and into both the near-term and longer-term future will be a key challenge as the enterprise makes the transition to a digital services business model.
Arthur Cole writes about infrastructure for IT Business Edge. Cole has been covering the high-tech media and computing industries for more than 20 years, having served as editor of TV Technology, Video Technology News, Internet News and Multimedia Weekly. His contributions have appeared in Communications Today and Enterprise Networking Planet and as web content for numerous high-tech clients like TwinStrata and Carpathia. Follow Art on Twitter @acole602.