One of the big benefits of the outsourced, cloud-based infrastructure model is speed. There are several facets to this; performance clearly being among the most important. Another is speed of provisioning, which remains a major selling point for cloud, given that IT buyers can spin up new resources according to need and in near real time.
Take a manager responsible for buying infrastructure for a large enterprise, for example. When considering various providers, the criteria they apply might typically address factors such as price, compute capacity, storage, bandwidth, and geographic location. This is very helpful for narrowing down the choice, but it doesn’t show how easily the organization can deploy and operate the chosen environment.
At the same time, infrastructure procurement processes have become so quick and convenient that they are almost as frictionless as buying something from Amazon, and that’s where problems can start.
Although it can feel like it, infrastructure is not a one-off, fire-and-forget choice; the operational impact persists throughout its lifecycle in production environments. A decision that appears straightforward at the start can quite easily become more consequential or expensive once migration or day-to-day management begins.
For example, migration might require more internal effort than initially expected, or the organization may discover it lacks some or all of the skills or support capabilities to manage the environment as intended.
So, how can these issues be addressed? It is important to plan for the possibility that migration may require more engineering time and coordination than expected. Many IT teams will be familiar with the need to reconfigure existing applications or data before they can run in the new environment, among other common scenarios.
Taking it one step further, when a new or upgraded service is more successful than anticipated, costs can easily grow with usage levels and capacity. Addressing this issue may require additional scaling, which can itself introduce additional management or security complexity.
Very few of these issues are insurmountable, but they can take longer to resolve if responsibilities between the customer and provider have not been clearly established.
Understand your operational requirements
Instead, buyers need to establish who is responsible for what. This should certainly cover day-to-day management, incident response, support escalation, and the process for adding capacity, but every organization will have its own version of this. When these points are on the table and agreed, providers should be able to explain the practical effect that scaling will have on the customer’s internal team.
Another often overlooked consideration is that infrastructure services ultimately rely on physical data center capacity, power, network connectivity, and hardware supply chains. These capabilities have always been important but have jumped up the agenda since AI and other data-intensive workloads increased demand for compute capacity, and at a pace few anticipated.
Today, buyers also need to understand whether the capacity they expect to use will be available where they need it, and on the timescale they require. Identifying these potential constraints early on allows the organization to account for them in its deployment plan rather than discovering them once a project is underway and, potentially, it’s too late to get what they need.
Clearly, some infrastructure decisions are more important than others, but in certain cases, they can affect engineering, security, compliance, finance, and the business teams, all of whom rely on the service in question.
Each group may also introduce requirements that should be fully considered before contracts are signed. If not, problems can arise when a need is identified after a provider has been selected or a deployment plan has been agreed. Involving the relevant stakeholders early also creates a clearer link between the infrastructure decision and the business outcome it is intended to support.
It may be tempting to keep some stakeholders on the periphery and decide for them, but having a say is not the same as making the final decision. The difference is that early input allows potential issues to be identified while options remain open, rather than giving every stakeholder some measure of control over the selection process, which is also not ideal.
Ask the right questions before committing
When looking at potential infrastructure providers, buyers should be able to define, in simple terms, the problem the new environment is intended to solve. At that point, it’s time for buyers to ask some serious questions.
For example, what will success look like once the service has been deployed and is supporting live workloads? How will the environment change when demand increases, including the likely effect on costs and internal management effort? Which activities will still require manual intervention from the customer’s team? What support is available during an incident, who owns each stage of the response, and how will escalations be handled?
Getting satisfactory answers is one thing, but it should get buyers to a point where they can properly test the assumptions within their cost model against the capacity they expect to use over time.
We’ve featured the best cloud storage.
This article was produced as part ofTechRadar Pro Perspectives, our channel to feature the best and brightest minds in the technology industry today.
The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here:https://www.techradar.com/pro/perspectives-how-to-submit
