The Content Delivery Summit, hosted by StreamingMedia.com in New York City on May 7, provided a stage for chip maker Intel to discuss a storage technology that it developed several years ago with Micron and later branded as Optane. The pitch was a head-turner for a few reasons.

First, let’s underscore a point. Intel was talking memory and storage, not micro-processing or CPUs, with which is it more popularly associated. Second, maybe you already knew about the 3D XPoint (crosspoint) technology that Intel and Micron announced back in 2015. But if so, you would have read about it on The Register, Electronic Design, and ComputerWorld, not StreamingMedia.com. In other words, reporters and analysts have been discussing Optane more in the gaming and gadgeteer segments than among IT and business pros involved with content distribution networks (CDNs), the focus of this Summit.

Latency and memory design

So what was going on at the New York Hilton Midtown? One clue came from the prevalence of “edge” in the session topics. Take the keynote address, titled “CDN is Dead (As We Know It) – Long Live the Edge!” In this talk, Marcus Bergstrom, GM of Unified Delivery Network (UDN) at Ericsson, defined the edge as the computer enclosure, space or facility geographically positioned to be closer to the point of origin of data or user base. Then he got more technical.

“For me, I define the edge as latency,” Bergstrom said. “As a baseline, to deploy something at the edge, you need to be within 10-20 milliseconds (ms) latency runtime.”

Here’s the connection: Latency is relevant to memory design. In a talk titled “New Technologies to Scale CDN Platforms,” Ed Dylag, market development manager for Intel Network Platforms Group, discussed the evolution of memory technology, from DRAM to HDD to SSD to today’s 2D and 3D NAND flash memory.

Like 3D NAND, Optane is also SSD-based and uses stacking to increase capacity. But it goes beyond. “What does it buy you over NAND?” Dylag asked. “QoS – and what we mean, is better latency.”

How much better latency is another question. To answer it, you’d need to review tests at the microsecond (µs) level of Optane and competitors, such as Z-NAND, the tweaked flash technology from Samsung. To understand why Optane performs better, you’d then have to delve into its post-silicon transistor, 3D Xpoint technology, or what appears to outside experts as a chalcogenide alloy-based, phase-change memory (PCM) design.

Intel simply states that Optane combines 3D XPoint technology with Intel memory and storage controllers, Intel Interconnect IP and Intel software.

Post-Summit Optane news

Intel had more news to share about Optane. Two days after the Summit, a bare-metal provider of infrastructure automation, Packet, announced that is collaborating with Intel on an Optane SSD-based, datacenter fast storage and memory solution.

What lies ahead? According to Seeking Alpha analyst Stephen Breezy, we’re at the slow dawn of a bright Optane future. That may happen, but Optane won’t boost data center and CDN performance by itself. Intel has plenty of mind and market share in computer processing; now it needs to win over more partners like Packet that are focused on storage.