Empowering Hyperscale Operators with a Collaborative Approach
December 03, 2020
With 2020 winding down, and the demands for a remotely-connected life firmly taking root, the rapidly growing hyperscale segment has become a category—and necessity—all its own.
Hyperscale operations started to more formally take shape over the last decade, when companies in the e-commerce, cloud services and social media markets needed to better support massive numbers of customers and customer data while being able to rapidly scale their market footprint.
To accomplish this, these “hyperscalers” reviewed each of the components that make up their data centers to become truly agile, flexible and scalable. They requested chip makers and server manufacturers to simplify and add more processing power to equipment. They streamlined their supply chains with adoption of specialized integrators that provide racks stacked with IT equipment ready for commission and deployment. They completely redesigned their infrastructure to match the advanced, power-hungry equipment needed to keep pace with unprecedented demand, all while developing and implementing more cost effective and energy efficient ways to effectively cool and operate their data centers to reduce their impact on the environment.
In a sense, the emergence of hyperscale data centers has flipped the coin on the typical supplier/customer relationship. Instead of buying an off-the-shelf, standard solution dictated by the product manufacturer, hyperscalers to a large extent have driven and created their own specifications designed to fit their unique needs. And because of their sizeable presence, the manufacturers and integrators that serve these hyperscalers have followed suit, gradually embracing these new demands and adapting segments of their operations to support them.
But how can product manufacturers still add value to spec-centric hyperscale data centers and integrators? More importantly, how can hyperscale customers benefit from a collaborative approach rather than relying on a top-down model where specs are met, no questions asked?
Some manufacturers are also able to provide a complete, kitted solution, with preinstalled components to further streamline the process and allow hyperscale integrators to keep focus on their core services and IT equipment integration.
As Director of Hyperscale Business Development for Chatsworth Products (CPI), I’m responsible for creating and fostering relationships and programs with hyperscale clients and integrators focused on rolling out large-scale data centers. In my 25 years of experience in the data center market, I’ve witnessed the invaluable advantages that a close collaboration among hyperscale data centers, integrators and manufacturers can yield.
As the search for a rack manufacturer that is prepared to guarantee a steady, robust and responsive supply chain gets narrower, it is important to consider selecting partners that are not only capable of accommodating purpose-built designs, but those that also provide global availability and support. (Image: CPI)
Consider the following:
Off-the-shelf, Customized or Both?
Today, typical hyperscale infrastructure designs—particularly racks—either come from standardized specs based around the Open Compute Project, or from highly customized designs that only a select-few manufacturers are able to accommodate into their routine engineering, design, prototyping and manufacturing processes and operations.
An ideal middle ground, however, is one where both the manufacturer and the hyperscale operator collaborate to create a solution up front that not only meets specifications around equipment loading, cable management, power and cooling, but also addresses how it fits into the operator’s overall data center ecosystem. Delivery schedules, transportation challenges, packaging waste and other considerations that rack manufacturers can address are all critical elements in developing the optimal solution.
The global footprint, massive scale and tight project timelines of hyperscale buildouts demand rapid engineering, prototyping and customization capabilities, combined with constant communication and scalable production flexibility to meet demand and keep projects on schedule.
This brings us to our next consideration:
Discrete vs. Design-Build Manufacturing
While discrete manufacturers can manufacture a product according to any given specification, design-build manufacturers have made this collaborative concept-to-production process engrained in their DNA, enabling them to provide a deeper understanding of customer needs and pain points. Because of their close interaction with customers to create custom solutions, design-build manufacturers have the benefit of calling upon real-world project applications and lessons learned. This first-hand experience means many design-build manufacturers become experts in achieving or even improving upon the design intent, based on that experience or any project-specific practical limitations. Some manufacturers are also able to provide a complete, kitted solution, with preinstalled components to further streamline the process and allow hyperscale integrators to keep focus on their core services and IT equipment integration.
Additionally, because hyperscalers’ designs are ever-evolving, working with manufacturers that have perfected their design-build operations means they are better suited to be adaptable to any critical last-minute changes. This is often the most efficient and expedient path to meeting the scalability requirements that hyperscale data center operators crave.
Here in today’s the digital economy, scalability connotates global capability—our final, and equally important, consideration.
Here, There, Everywhere
As the search for a rack manufacturer that is prepared to guarantee a steady, robust and responsive supply chain gets narrower, it is important to consider selecting partners that are not only capable of accommodating purpose-built designs, but those that also provide global availability and support. To hyperscale data center operators, these attributes mean improved product availability, service and support within region and ultimately faster, more predictable and consistent new data center build-outs.
Selecting a rack manufacturer that’s proven to provide local, regional and global capabilities is the easiest, most reliable way for hyperscalers to fast-track their deployments. Easier said than done? Not anymore.
This article was originally published in the Data Center Frontier website.