Let's get one thing straight right out of the gate: there is no ‘cloud versus data center.’ It's a false dichotomy. The reality is that cloud and data centers are critical parts of the digital infrastructure. More than that, their relationship is one of symbiosis. Advances in data center technology (think high-speed connectivity, efficient power, and more) boost the effectiveness of the cloud, and vice versa. This synergy has the potential to help clients fuel limitless future product innovation and service optimization for their customers.
From the start, the cloud seemed like a simpler, lower-cost (more on that in a bit) alternative to managing on-prem infrastructure. But as data and compute demand has grown, so too has the realization that organizations need more flexibility than a ‘cloud-only’ strategy can provide. It's here, in particular, that colocation data centers still lead the way.
For one, cloud providers are, by far, some of the largest consumers of colocation data center services and solutions. That means if data center clients need to operate a portion of their workloads in the cloud (that is, through hybrid architectures), data centers have a unique opportunity to facilitate better-faster-cheaper methods for connecting clients, partners and customers to the cloud. This means that both data center and cloud clients have access to better solutions and services through the utilization of both.
But what does that look like in practice? How do data centers help their clients make the most of hybrid architectures?
There are a few factors at play—namely interconnectivity, agility, and accessibility.
The world is going through a non-stop data and compute binge. Take one single application, Netflix, for example. It's been estimated that each Netflix user streaming in HD used around 9.6GB of data per day. With roughly 70 million subscribers in the US and Canada, that's over 18 million TB of data used in a single month. Once again, that's just a single (albeit massive) application.
If an organization operates digitally at all, distributing workloads across infrastructures becomes a major priority. But that means shifting those ever-growing workloads from one part of the world to another is an ever-growing challenge.
But because the speed of light is actually a speed limit, data can only move so fast. One solution is decreasing the distance workloads and data need to travel. It's like driving to San Francisco—it's a much faster trip if you start from Sacramento rather than Reno.
If you want to reduce the time it takes your services to reach end-users, you want to place your workloads as close to them as possible. A given cloud providers' servers may be anywhere. There's no guarantee they're anywhere close to your users. Colocation data centers, on the other hand, will likely have locations in valuable markets as well as direct access to high-speed fiber.
Also, as already mentioned, cloud providers leverage colocation data centers extensively. So, if you're a data center client, chances are pretty good that your provider can offer a direct, private, high-speed connection to the cloud—bypassing the public Internet entirely. Your cloud-based workloads need only to travel to the next hall or cage, rather than the next city or state.
The synergy between cloud providers and data centers can also make customers more agile. As we've seen, data centers put customers closer to the cloud through direct connections and on-ramps. That proximity yields a cascade of benefits.
For one, customers' cloud-based applications can operate much more efficiently, reliably, and cost-effectively with the help of support and services provided by the data center. If cloud applications are running smoothly, it's easier to devote more time and resources to other mission-critical activities—namely innovation.
It's not impossible to experiment within the cloud environment, it's just that data centers are more well-suited to dynamic workloads necessary for trial-and-error and evolution.
The flexibility, modularity, and level of support provided by colocation data centers mean they are natural sandboxes for new ideas.
For example, say you're a large content delivery network, and you have a new video product that you want to test in-market before a complete roll-out. With a colocation data center, you have complete control over your resources. You can easily shift the necessary workloads to a location in close proximity to a single target test market.
And that brings us to the third idea. The cloud is an ideal, cost-effective solution for applications and workloads that don't require significant mobility.
On the other hand, if your applications need a high level of fluidity and flexibility, ingress and egress fees can very quickly reverse any cost-savings found within the cloud. Also, once again, the hardware containing your stuff could be anywhere within a given cloud provider's entire footprint. More than likely, you don't know where your data and workloads are located, let alone how to access them. As with the example use case above, at a data center, it's your footprint. Data centers can also provide you with easy on-ramps to your more static cloud assets.
That said, nothing you've read until now should be construed as reasons to only use cloud or only use data centers. Your takeaway should be the necessity for balance and diversity within your infrastructure. When they work together seamlessly, cloud and colocation can and will amplify each other’s strengths.
For example, if your applications are running mostly hands-off in the cloud, you have more freedom to take advantage of the flexibility data centers offer—ultimately providing you with the freedom to create better products and experiences for your customers.
That means the only choice in front of cloud and data center providers is how we can work together to build better solutions that empower our customers with solutions and services that fuels innovation now and in the future.