If the world wasn’t convinced that people’s business and personal lives were inextricably linked to and through the internet before, then the pandemic underlined that in no uncertain terms. But after a decade or more of overall improvements in the efficiency of critical infrastructure (especially among hyperscalers), it’s hard not to wonder whether or not the current course and speed of growth can continue to deliver similar improvements in the future.

Getting Future-Ready

There are three immutable truths that all mission critical data center managers must confront if they are going to be able to efficiently serve the future data demands of their organizations.

1. Data is growing Exponentially, Data Centers are not.

Numerous reports all tell a similar story. The volume of data being created, as well as the number of devices that want to access it, is growing exponentially. International Data Corp. (IDC) predicts a 61% compound annual growth rate of generated data being accessed by an additional 1 billion people (bringing the total to 6 billion) by 2025. This equals numbers that are difficult to fathom — 175 ZB, or 175 trillion GB, if that helps. That is a lot of data, and there seems to be no end in sight for the growth of it.

Advances in storage media technology pack more storage into the same rack space, of course, but it’s not enough. Data center managers have been grappling with these trends for years, but the curve appears to be getting steeper and more challenging as time goes on.

2. The industry is Hungry and Hot.

With the growth of data comes a growing need for processing — cooling it all is still quite a big ask. Racks generate heat — lots of heat. What’s more, average data centers consume around 30% of their total power in cooling alone. The best-performing data centers tend to be public cloud vendors and hyperscalers that have the resources to pursue exotic solutions that simply aren’t accessible to standard enterprises. It’s probably also fair to assume that, in the rush to the cloud, investment in enterprise data center infrastructure probably stalled for many businesses. 

3. Data is Shifting to the Edge.

While data is still growing in the core, the shift toward edge computing makes the first couple of points an even greater issue and will make the edges of the network more mission critical than today. Moving data to the edge means moving to locations that are near in proximity to where the end users operate. This includes cities and other densely populated areas where space is at a premium and optimizing efficiency is limited. While this is a problem everywhere, its particularly problematic on densely populated island nations where the challenge of growth is all too present. 

Data Returns to the Enterprise 

There’s an increasing trend of shifting data out of the public cloud and back on-premises. There are a few different reasons for this. Many wish to apply greater security controls over more sensitive data by keeping it “local.” It also solves latency issues for regularly accessed data and enhances performance. Finally, it comes down to getting a better handle on the egress costs being charged by the cloud vendors.

The big public cloud vendors are jumping on these issues and trying to stem the tide with solutions, like AWS Outposts, but does that really solve the issue of data security? And what about the challenge of still being locked into a proprietary architecture?

How to Respond

Abounding hurdles continue to pop up as data centers continue on their path to sustainability. While there is no single approach to overcome these challenges, appliance design and manufacturing can potentially make a great impact. 

Today, it seems that many data centers — especially those moving to software-defined architectures — have decided almost implicitly that hardware doesn’t matter. But, throwing processing power at a problem to solve performance issues is inefficient. And inefficiency drives up power, cooling, and rack space requirements — commodities that are in increasingly short supply. 

Task-specific appliances are built from the ground up and designed to perform one software-defined function. These high-performance, highly efficient designs reduce both power and coolings requirements. The knock-on effect is the ability to pack more petabytes of storage into each rack and stay within the power and cooling budgets available. This will free up square footage in data centers, and it can make edge deployments possible in environments that may have otherwise been problematic.

Based on the trends, the current trajectory of data centers does not appear to be sustainable. It’s time as an industry to take a step back and rethink the approach to appliance design.