Intel announced the availability of its Rack Scale Design (also known as Intel® RSD, and formerly known as the Intel® Rack Scale Architecture). It is the first framework to use and to be based upon the Redfish™ industry standard from DMTF for modern and secure management of scalable platform hardware in the modern datacenter.
This might mean that the era of the software defined data center is officially here – as evidenced by many of the new software products in the market, that range from converged to hyper-converged infrastructure.
And hardware OEMs are shifting with the times, building new solutions to leverage these new technologies. Where the enterprise market in the past was more focused on the creation of separate components (server, storage, and networking) – the new trend is towards all of these components being combined into a single solution.
Traditional general purpose platforms – as well as the new hyperscale iterations – fall on different ends of the spectrum. And while they may share similar features, they’ve got different design points that address different requirements
Traditional hardware was designed around reliability and ease of maintenance; hyperscale hardware prioritizes cost, density and overall lower TCO.
Rarely do the ends of the spectrum meet – and the personal preference of the end user can vary wildly. For example: the enterprise customer typically uses VMware ESX, while the hyperscale customer leans towards what’s available from the open source community.
A newer enterprise might have totally different needs than those of a traditional data center customer. So being able to tap into an architecture that allows the customization of hardware and software helps us be more flexible. Intel’s new Rack Scale Design (formerly known as Rack Scale Architecture) is exactly that. It’s the next phase of the Intel Rack Scale Architecture, a modular approach that empowers customers to compose their IT hardware needs from varying pools of IT resources. Intel Rack Scale Design simplifies advanced technology to accelerate adoption of open, interoperable solutions of hyperscale data centers.
Intel Rack Scale Design (RSD) allows customers to purchase exactly the computational, storage and networking components they need. This delivers a configuration that suits – and they get to stack their “building blocks” to service appropriate workloads.
But how are the varying components bound together?
The mortar is software: it holds everything together, and customers can manage and orchestrate their hardware on demand. There are two parts of this equation: the first is the solution needed to manage the hardware; the second is the means by which workloads are orchestrated. There is an appetite in the market for this software; Converged Infrastructure has seen an increase in adoption in the market. Customers want to manage and orchestrate their hardware and software easily so they can meet the demands they are seeing in the business.
Intel Rack Scale Design will leverage leading edge, open source system management software (like SMASH or Redfish), to manage the all of the hardware building blocks – effectively leveling the playing field. This lets customers use multiple hardware vendors without compatibility issues.
At a higher level, Intel Rack Scale Design can act as an umbrella – helping to take the configuration a step further. It provides a single, common interface for customers to manage the workload, using a customer’s preferred orchestration platform. The Intel Rack Scale Design components, consisting of PODM, PSME, RMM, and VTS will provide a single common interface that will allow customers to manage their environment. In the end, this means a common platform managing physical hardware – and – the software layers.
In the future, as everything comes together, a customer will be able to purchase a Rack Scale Design solution that includes the so-called “building blocks,” of their choice, arranging and composing them in a way that fits their workload needs.
On our end – at the OEM level – we’re working with Intel to develop new capabilities around Rack Scale Design, leveraging our expertise in hyperscale. We’ve been assisting hyperscale customers in China – and other countries -- and we’re tackling new, unique customer segments that require a different approach to design. This new architecture of customized hardware and software, tailored to workload needs, is exactly what we’re being asked to deliver.