Advanced Computing in the Age of AI | Friday, March 29, 2024

In Bid for Major Carriers and Service Providers, Dell EMC Rack Scale Infrastructure Offers ‘Hyperscale Principles’ 

Calling it “hyperscale inspired,” Dell EMC today announced a rack scale infrastructure – including compute, storage, networking, power, cooling and open management – aimed at major carriers and service providers, mammoth companies whose IT requirements rank just below those of the major public cloud and web-scale companies (i.e., FANG*).

The company said the offering, called DSS 9000 and scalable up to 96 nodes, is aimed at companies such as Sprint and Telefόnica to “accelerate the shift to software-defined data centers and cloud computing initiatives.” Built around Open Compute Project (OCP) specifications, the DSS 9000 is a pre-integrated, configurable rack that uses new Intel Xeon Scalable processors.

What could be called the “sub-hyperscale” market, comprised of roughly a hundred or fewer large carriers/service providers, is one that’s increasingly adopting the IT strategies and needs of the FANG set: compute-intensive, large-scale workloads working off of immense and growing volumes of data.

While FANG companies continue to buy from Dell EMC and HPE, they also – as has been widely observed – have increasingly multi-sourced their vendor mix to include low-cost, stripped down, custom servers and racks from low-cost Asian providers. Meanwhile, as the workloads of “sub-hyperscalers” have grown so have their IT requirements to the point where they demand rack scale infrastructures that a company like Dell EMC, with experience building solutions at the high end of data center market, can deliver. To wit, the DSS 9000 was developed under the auspices of what used to be called the Data Center Solutions (DCS) unit, now called Extreme Scale Infrastructure (ESI).

“We’ve been working with what we call hyperscale customers, large public cloud guys, for more than 10 years, started at DCS, and we’ve continued to build upon that experience as we try to grow our business beyond those hyperscale customers,” Jyeh Gan, director, datacenter scalable solution, told EnterpriseTech, adding that the carrier/service provider market is growing three times faster than the core server market.

“We’ve been talking about it for about a year and a half with the old Intel architecture,” Gan said, “we’ve been shipping it to a handful of customers as we’ve moved to the new Intel Xeon Scalable processors, and we want to open it up to more customers now.”

Source: Dell EMC

Dell EMC said CenturyLink, Telefόnica and Sprint are among the first customers using the DSS 9000.

“Our cloud infrastructure offerings are designed to help customers maintain control, increase visibility and be more agile so they can run both cloud-native and traditional applications more efficiently,” said Steve Nolen, senior product manager, CenturyLink Technology Solutions. “We chose the DSS 9000 to be the backbone of our Private Node offering because of its flexible design and ability to give our customers a build-to-order private cloud solution. Our public cloud offerings are also offered on the DSS 9000, providing a consistent and reliable experience for customers who have both public and private cloud needs.”

Sprint is using the DSS 9000 as part of an effort to make its traditional mobility architectures more streamlined, efficient and scalable as it moves to a virtualized environment. The DSS 9000 was deployed to test its new open source network functions virtualization (NFV) and software-defined networking (SDN)-based mobile core reference solution.

Telefόnica, meanwhile has chosen the DSS 9000 to support future networks, designed to be more automated and software-based. “(It’s) one of the industry’s most ambitious virtualization projects,” the company said in a prepared statement, “…designed to meet the stringent requirements of the network environment, including carrier-grade, performance and operational capabilities – while being as open as possible. We became interested in the DSS 9000 because it is built on open principles and aligns with our mission to make the project as future-proof as possible.”

Gan said the DSS 9000 combines both an at-scale infrastructure along with pre-integration intended to ease implementation.

“It was a product designed for the cloud (services companies), now we’re bringing it to more people than that, they can take the product and roll it right in, it’s pre-integrated, they can plug it in and away they go,” Gan said. “They can manage it the same way as everything else in their data center.”

Other DSS 9000 features include:

  • Provisioning, deployment and management conducted at the rack level, designed to accelerate delivery of new services.
  • Management of large-scale operations via a single interface based on the Distributed Management Task Force (DMTF) Redfish specification, designed for interoperability with heterogeneous systems.
  • Disaggregation of hardware resources to support the move toward a future of composable infrastructure with Redfish and Intel Rack Scale Design (RSD).
  • Adoption of open principles, including recognition of the DSS 9000 by the Open Compute Project as OCP-INSPIRED and designed with an open networking approach to support NFV and SDN initiatives.
  • Flexible configuration deployment with various rack heights accommodating full, half and third-width compute and storage sleds as well as other standard server, storage and networking options.

“Carriers and service providers need next-generation infrastructure in order to balance the need to innovate at the speed of business while spending less,” said James Mouton, senior vice president, Extreme Scale Infrastructure, Dell EMC. “With the DSS 9000, we are leveraging our 10 years of experience working with industry titans to make it easier to purchase, optimize, deploy, manage, scale and service infrastructure at scale.”

* Facebook, AWS/Azure (Microsoft), Netflix, Google.

EnterpriseAI