Organizations adopting a cloud-first strategy need to define their outcomes first - their “roadmap for success” - in order to understand long-term goals and to provide guidance throughout the process.

That’s according to Ryan Mallory, data center industry veteran and chief operating officer at Flexential.

“Having a virtualized aspect to your overall architecture is important, but it’s only important if you have defined outcomes. You have to create that roadmap for success. If you’re not defining what your success criteria is, both from a near- and long-term perspective, you’re not necessarily going to have anything to work towards to achieve success.”

Mallory was speaking DCD’s most recent Datacenter-as-a-service event. Any IT strategy based on cloud-first, he continued, will nevertheless have a number of basic components.

“You have your public and your private component. Your public component gives you access to the large infrastructure-as-a-service providers, such as Microsoft, AWS, Google, or some of the others that are out there. These kinds of companies can provide versatility for the compute and storage capabilities that you may need; they can give you flexibility for accessing new markets. And they can provide support for specific applications that are cloud-ready.

“The private cloud strategy is ever-important and I don't think it's going away. The private cloud gives you access to your private compute and storage infrastructure enabling you to keep sensitive data closer to your end users in a very controlled manner. This can also be data that is highly ‘chatty’, meaning that there's a high level of access to this data required and [therefore, you can prevent the] high egress charges that you may see in a public cloud environment.

“Now let's move to your as-a-service component. You can have your as-a-service capabilities inside a public cloud offering. In essence, what that means is you have a service that sits inside a public cloud capability. That's why it's so important to have this forward looking roadmap – to understand what those services are,” said Mallory.

“You can also have those sit inside your Edge-distribution model... you always want to have the appropriate storage component for your cold storage, or your deep storage for things that you just have to have for compliance,” said Mallory.

He continued: “So you can see how important it is to have a longer-term roadmap for workload optimization, as well as performance. This really matters unless you're accessing your cloud-first strategy via an interconnection platform that gives you both physical and virtual connectivity.

“This allows you to have dynamic access through the public Internet for anytime access, or through private, secure connectivity, physically through a cross-connect, or virtually through a virtual cross-connect to the cloud or as-a-service components.

“While these are critically important, it's also essential to understand the importance of ‘data gravity.' As you start to look at your workload optimization, more importantly, you need to start to evaluate the overall importance of the ecosystem associated with your roadmap for your cloud-first model. The data gravity quotient here is about the usability of data and access to that information to internal and external clients, and stakeholders so you can interact with that information and make appropriate decisions.

“Data gravity is becoming as important as your IT infrastructure strategy… The evolution of the ecosystem is what is going to ensure that your business and your IT stack is not only effective, but it is a revenue center, not just a cost center,” he said.

Living on the Edge

Obviously, decisions made pre-pandemic will have helped many organizations to cope over the past 12 months or so. For example, having a well-established company policy of issuing laptops to staff, rather than having them chained to desktop PCs in the office.

With the anticipated rise in more dispersed working environments – smaller offices and more home working – the need for more processing to be done at the Edge, and the introduction of IoT, corporate IT is set to get even more complicated.

Even something as supposedly straightforward as the Edge will carry with it new challenges, Mallory warned. He foresees a three-tier model emerging. “There’s the near Edge, the middle Edge, and the far Edge. The near Edge is companies and colocation providers that sit at the highly interconnected fiber connection points across the globe. You know what the value is for those near-Edge deployments.

“The middle Edge… is the locations that are the emerging population centers that are starting to become much more relevant due to the trend of the past 12 months. Then, lastly, there’s the far Edge. This approximates to cell towers and low latency applications. We’re still seeing the evolution of the far Edge with content delivery networks and the need, in the future, for autonomous vehicles,” said Mallory.

The far Edge will therefore require ultra-low latency, he added, with autonomous vehicles, by implication, requiring guaranteed uptime and low-latency.