The Way we Deploy Racks Needs to Change

The Way we Deploy Racks Needs to Change

Today, virtually everything is available on demand. Whether it’s your favorite meal, beverage, or feature film, you can get what you want when you want it.

What does this have to do with IT infrastructure? Well, what if there were more opportunities to have network capacity delivered to your door? Whether that door was in the desert or in Manhattan, you could access capacity to meet compute, and storage needs as they arise.

IT planning pitfalls

For as long as IT departments have existed, budgeting and planning have been largely based on a best, educated guess. Leaders consider current systems, the company’s projected growth, and the available technology. From there, they factor in a smattering of inputs and evolving priorities from various business units to produce a budget and plan for future needs. The resulting infrastructure in most organizations is a mesh of interconnected nodes, each node representing various functions within the organization. Nodes are added as needed through capex projects with long planning horizons.

The problem with this strategy is how quickly these functions change and grow in number. This model isn’t set up for the rapidly changing world we live in today. Tools and budgets are fixed and firmly set.

At the start of the year,?Gartner?published a report by?David Cappuccio?titled “The Evolution of IT Infrastructure — Designing for the Unknown.” In the article, David says “IT infrastructure strategies are evolving from those based on specific location or technology stacks to those flexible enough to respond to rapid change. New infrastructure strategies must support the idea that IT doesn’t know an enterprise’s future requirements but must be able to quickly deploy applications and workloads wherever they are needed."

Cappucio goes on by saying that to account for this, IT leaders must rethink infrastructure delivery to account for the many unknowns, including some significant things like how much capacity is needed, where and for how long. Building a flexible network creates an opportunity to contain cost and deploy capacity where and when it’s needed without significant capital expenditure.

IT on demand

One-way businesses are responding to this concept by adapting more flexible storage and server systems that can be scaled up or down to keep pace with ever-changing levels and locations of demand. Historically, users bought excess capacity upfront to ensure they could meet future IT service demands. This concept was inefficient because it often meant paying for services or capacity you didn’t use. Wasted capital.

A better model is capacity on demand, when and where you need it. This allows the business to upgrade systems without a massive capital outlay.

Compass Quantum recently launched White Space as a Service (WSaaS) to add incremental IT capacity, either to expand an existing facility or extend operations everywhere from the network's edge to the downtown metro. Quantum combines a modular data center and a comprehensive package of services, along with the personnel to handle installation and on-going operation for each 100kW data center deployment.

The key benefit of capacity-on-demand type offerings is how quickly and cost-effectively they can be available to meet enterprise demands. In the case of Compass Quantum, incremental local, remote, or edge capacity can be added in as few as four months without the business needing to modify operational procedures or add the additional headcount normally associated with network expansion. The service’s lease structure also negates the need for significant capital expenditure to facilitate rapid deployment.

IT in the future

The days of predefined placement of workloads in specific data centers are waning. Future workload placement strategies will be based on business requirements and value rather than traditional locations. Workloads could be needed in specific regions to reduce latency or at specific sites for a temporary period of time. All of this requires flexibility and the need to move fast and be flexible.

If the pandemic taught us anything, it’s never to underestimate the force of a curve ball and be prepared to pivot fast. There are lessons we can take forward from that experience as we look to the future of enterprise computing and storage.

Michael Lesniak

Industrial Water Exec | Net Positive Water Catalyst ?? | Water Author | Trusted Advisor | Board Director

1 年

Tony Grayson there will likely be a movement toward deconstructing racks someday into hyperscale components. Racks are incredibly underutilized at any one time depending on workload type. By deconstructing into separate buildings on hyperscale campus’, incredible efficiency of scale will be unlocked to handle every workload with every component running at full capacity. Not sure if you you could do that at edge within a single building? This will likely be a massive industry shift, but still some things to be sorted to make it work.

回复
Taylor Monnig

Chief Technology Officer at Cleanspark

1 年

Racks should be horizontal and submerged in fluid. Open bath immersion is the best way.?

Greg Lythe

Senior EMEA Enterprise Solutions Consultant @AHK Assisting companies reimagine the modern datacenter

1 年

Excellent post, very insightful and thought provoking, thank you.

Billy Callahan

Senior Director, Product Management at Accelsius | Retired Submarine Officer | TS/SCI

1 年

I like the Compass Quantum model. It’d be really ‘cool’ (pun intended) if there were a 2PLIC option!

要查看或添加评论,请登录

Tony Grayson的更多文章

社区洞察

其他会员也浏览了