Intel Defines “Data-Centric” Leadership with Cascade Lake

By Charles King, Pund-IT, Inc.  April 10, 2019

The tech industry so reflexively worships shiny new startups that folks often forget or even disparage larger, well-established vendors. That’s unfortunate for any number of reasons but high on the list is that such attitudes miss a critically salient point: Self-reinvention, not stasis is the key to long term survival in technology.

PR-minded venture firms continually trumpet the “unique” innovations of whatever IPO-bound company stands to fill their pockets. In contrast, established firms develop and deliver more substantial innovations month after month, year after year, decade after decade. At least the successful ones do.

That point was in clear view during Intel’s recent launch of its 2nd generation Xeon Scalable processors (aka Cascade Lake) and other new ”data-centric” solutions. Let’s consider these Intel offerings, why they are important and how they will help keep Intel and its OEM customers riding high in data center markets.

The view from Cascade Lake

So, what exactly did Intel reveal during its launch?

  • General availability of its new 2nd-Generation Xeon Scalable processors (aka Cascade Lake), including over 50 workload-optimized solutions and dozens of custom processors
  • The new offerings include the Xeon Platinum 9200 processor which sports 56 cores and 12 memory channels. According to Intel the Xeon Platinum 9200 is designed to deliver socket-level performance and memory bandwidth required for high performance computing (HPC), artificial intelligence (AI) and high-density data center infrastructures. Both IBM Cloud and Google Cloud announced plans to deliver services based on Xeon Platinum 9200 silicon.
  • Among the custom silicon offerings are network-optimized processors built in collaboration with communications service providers (SPs). These solutions are designed to support more subscriber capacity and reduce bottlenecks in network function virtualized (NFV) infrastructures, including 5G-ready networks.
  • The Xeon D-1600 processor, a system on chip (SoC) designed for dense environments, including edge computing, security and storage applications.
  • New features built into the 2nd-Gen Xeon Scalable processors include integrated Deep Learning Boost (Intel DL Boost) for AI deep learning inferencing acceleration, and hardware-enhanced security features, including protections against side channel attacks, such as Spectre and Foreshadow
  • Next generation 10nm Agilex FPGAs that will support application-specific optimization and customization for edge computing, networking (5G/NFV) and data center applications
  • Support for new memory and storage technologies, including Optane DC persistent (storage-class) memory, Optane SSD DC D4800X (Dual Port) for mission-critical enterprise applications and SSD D5-P-4326 (QLC 3D NAND) for read-intensive cloud workloads.

Pricing for the new solutions was not revealed during the launch. Product availability details can be found at intel.com.

Why it matters

Intel’s launch of its 2nd-Generation Xeon Scalable processors and other data center solutions comes at an odd time for the company. Conventional wisdom often depicts Intel as an overly complacent behemoth harassed and bloodied by swifter, agiler (often far smaller) foes. Critics claim the company is reacting too slowly to marketplace shifts and is being surpassed by more innovative technologies and vendors.

In some cases, Intel didn’t do itself any favors. The company unexpectedly dismissed its CEO, Brian Krzanich, then took months to find a new chief executive within its own ranks (Robert Swan, who joined Intel in 2016 as its CFO and led the company on an interim basis after Krzanich departed).

But in other cases, the company was at the top of its game. For example, while discovery of the Spectre, Foreshadow and other side channel vulnerabilities could have been a public relations nightmare, Intel’s willingness to take responsibility and transparently detail its efforts to repair the issues kept the situation from blowing out of proportion.

Despite these challenges, Intel continued to deliver steady, often impressive financial performance. While the company was no investor’s idea of a high flyer, it also had less distance to fall, as well as considerable padding to help mitigate unplanned impacts. The value of Intel’s conservative approach became apparent when GPU-dependent NVIDIA and AMD saw their fortunes swoon and share prices plummet when crypto-currency markets unexpectedly tumbled.

So, what do the new 2nd gen Xeon Scalable processors and other solutions say about Intel and how it sees its customers and competitors? First, consider the sheer breadth of technologies involved. Along with the latest/greatest features and functions that you’d expect in a next gen Xeon announcement, the inclusion of new SoC, FPGA and Optane memory and storage solutions, as well as workload-specific technologies, like Intel DL Boost demonstrate how and how effectively Intel is spreading its data center bets.

In addition, the depth of available solutions is impressive. That’s apparent in the 50+ SKUs that feature 2nd gen Xeon Scalable silicon and is also highlighted by the 56 core/12 memory channel Xeon Platinum 9200. It’s interesting that both IBM Cloud and Google Cloud proactively announced plans to develop offerings based on the chips since both focus on the needs of enterprise cloud customers and are sticklers for top-end hardware performance. Their support suggests that Intel’s claims about the Xeon Platinum 9200’s HPC, AI and high-density capabilities are fully justified.

Finally, the flexibility and customizability of the new chips is worth considering. Intel’s focus on workload-optimized solutions is one example of this since it reflects OEMs’ (and their customers) increasing focus on integrated, optimized systems. In fact, the new network-optimized processors that resulted from Intel’s collaboration with communications service providers offer intriguing insights into how Intel can and likely will continue to add discrete new value to its silicon portfolio.

Final analysis

Overall, the 2nd-Generation Xeon Scalable processors and other new data center technologies demonstrate both Intel’s ability to deliver substantial new innovations and its lack of complacency. That’s sensible from a tactical standpoint since the company’s competitors aren’t standing idly by – for one, AMD’s launch of its new Epyc “Rome” data center chips is just weeks away. However, it’s also strategically important for Intel to show how it became a leader in data center solutions in the first place and why it deserves to remain there.

The tech industry’s passion for perky start-ups and shiny new objects isn’t likely to fade any time soon. However, it would be a mistake to assume that leading-edge imagination and innovation reside solely in smaller organizations. Intel’s new “data-centric” 2nd-Generation Xeon Scalable, Agelix and Optane solutions prove that oft times, the most innovative vendor you can find is the one you’re already working with.

© 2019 Pund-IT, Inc. All rights reserved.