NVMe over Fabric (NVMeF) is an advanced storage technology that, unfortunately, works optimally in small doses. Or it did until now, according to James Meeker (pictured, left), director of enterprise solutions at Micron Technology Inc., and Eric Endebrock (pictured, right), Micron’s vice president of storage marketing.

Meeker and Endebrock discussed NVMeF and the future of storage with Dave Vellante (@dvellante) and David Floyer (@dfloyer), co-hosts of theCUBE, SiliconANGLE Media’s mobile live streaming studio, at One World Trade Center during the Micron Summit this week in New York City. (*Disclosure below.) 

Spreading NVMeF across multiple servers has often defeated the purpose of the high-speed, low-latency storage. “Latency typically will get lost across your fabric, and applications won’t be able to take advantage of it,” Meeker explained.

This is the main hurdle that Micron’s new SolidScale platform architecture seeks to overcome with the NVMe flash storage innovation. “SolidScale is an architecture that we’re delivering that allows you to share NVMe low-latency benefits, high-capacity across all of your servers,” Meeker said.

The actual path that data must travel from storage to the application and the intervening tasks can inhibit flash’s considerable potential, Endebrock added.

“Every time I want to write one byte onto an SSD [Solid State Drive], by the time I get up to an application level through all the different things — file systems, the application itself, even the firmware on the SSD — you’ll find that I’m probably writing hundreds, if not thousands, of bytes of data all the way down at the very bottom of that stack,” he said.

Breaking through this and enabling a more direct access layer for workloads is the next frontier for storage and the basic idea of products like SolidScale, Endebrock stated.

Big data analytics and Internet of Things applications are particularly dependent on super-speed retrieval and serving of data and connecting these processes to transactions, Floyer and Endebrock agreed. Micron’s customers have also echoed this. In fact, SolidScale’s development was largely informed by customers who wanted their data processes to sync with transactions very quickly, Endebrock said.

To be clear, though, SolidScale converges components for greater plug-and play capacity. Micron is not suddenly claiming to be an all-in-one solutions provider for customers. “They know their apps best. I don’t want to pretend that Micron’s coming in and advising them on how to architect and build their full solutions; that’s not the case,” Endebrock added.

What Micron offers is a wealth of knowledge on flash and NVMeF based on extensive research on how to use it, as well as how not to use it, Endebrock explained. “A lot of our value is making sure you avoid those common pitfalls so you don’t actually suboptimize it. If you’re dealing with 200 microseconds, one screw-up turns it into 1,000 — and anybody can get 1,000 microseconds out of a system,” he said.

IoT and data analytics in 200 microseconds

Micron has touted those 200 microseconds in the press for SolidScale. TheCUBE co-host Dave Vellante asked Endebrock to explain why 200 microseconds is a number customers should care about.

Data’s end-to-end route between CPUs locally is about 195 microseconds with an NVMe bus, Endebrock explained. “For a five-microsecond penalty, you can unleash that flash from being direct and dedicated to an individual server and get all the benefits of that shared, accelerated storage pool,” he said. Such benefits include superior management and lower cost of ownership, he stated.

In some cases, the CPU may even by bypassed. Managing I/O [Input/Output] on servers consumes up to 22 percent of CPU cycles in some cases, Meeker said. “By bypassing this or offloading your I/O requests, it frees up your CPU, so you get a better TCO [Total Cost of Ownership],” he stated, adding that this can increase the density of servers in a rack.

Flash’s time to shine

Flash storage could have solved problems like this in the past, but it has typically been tacked onto existing architecture too late, Meeker stated.

“We’re going to get it right this time by starting grounds-up by looking at flash as the medium to solve data center and storage problems,” he said.

The SolidScale array might be a good fit for companies with granular knowledge of their applications who want full control of them, rather than Fortune 1000 companies with tens of thousands of apps, Endebrock added.

“A lot of what we’re solving for are companies who are running 10 applications — they’re more focused apps, they’re bigger, they’re scale-out,” he said.

Watch the complete video interview below, and be sure to check out more of SiliconANGLE’s and theCUBE’s independent editorial coverage of Micron Summit 2017. (*Disclosure: TheCUBE is a paid media partner at the Micron Summit. The conference sponsor, Micron, does not have editorial oversight of content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE