Until fairly recently, infrastructure plans followed a familiar pattern: add processing capacity, increase storage and make the network faster. Investment kept pace as businesses generated more data and came to rely on a broader range of applications and digital services. Now, another factor is shaping those decisions: power.
Those demands have not eased. Data heavy applications, high performance computing and some uses of AI all require substantial computing resource. Yet the technology is only part of the equation. The cost and availability of power, together with sustainability commitments, can affect where new infrastructure is practical. Responsibility for power remains with facilities teams, but its impact now needs to be considered as part of wider IT infrastructure planning.
This changes the questions organisations need to ask. Capacity alone is a poor basis for a long term plan. IT leaders also need to know whether today’s environment can cope with the workloads they expect tomorrow, and at what cost. Performance and resilience matter, but so do affordability and room to grow, as Wayne Kiphart, CEO, CloudFirst explains…
From Traditional to Hybrid
For many organisations, this means looking beyond the continual expansion of infrastructure hosted on site. Systems kept on site remain essential for some critical applications, but they are no longer the default for every workload. Many businesses are therefore using a mix of on site systems and public or private cloud services, placing each workload where it makes the most sense.
Power is one of several factors in those decisions. Analytics and other resource intensive applications can place greater demands on infrastructure than traditional enterprise workloads. In some cases, plans to expand an existing environment may be limited by power availability, cooling capacity or running costs. Rather than adding more hardware by default, organisations may need to ask whether some workloads would be better suited to the cloud.
To Cloud or Not to Cloud?
Cloud is not always the answer, though. Every workload has its own performance, security, compliance and data sovereignty requirements. Critical production systems, applications that depend on very low latency and regulated data may still be better suited to infrastructure kept on site. There is no standard formula. The answer will differ according to the workload and the needs of the business.
Good monitoring can expose unused capacity, bottlenecks and avoidable costs. Combined with sensible automation, it gives IT teams a firmer basis for deciding what to tune, replace or move. It can also help them spot weaknesses before they lead to disruption.
A full migration is not the only route to improvement. Newer servers, storage and virtualisation platforms may give an organisation more capacity without adding the same level of overhead. They can also improve utilisation and reduce day to day complexity.
Keeping Services Running
When a business relies on digital services, the infrastructure behind them cannot be treated as an afterthought. It has to cope on an ordinary working day, during a sharp increase in demand and when something fails. Plans for backup, recovery and workload movement need to be tested against those situations, not left on paper.
These requirements cannot be addressed through isolated technology decisions. A useful starting point is to trace the dependencies around an important service: the applications it calls, the data it uses, the network it depends on and the securitycontrols around it. That view often reveals risks that are easy to miss when each system is considered separately. Energy use and running costs belong in that assessment too.
Choosing What Runs Where
Making sense of this more complex picture requires organisations to consider each workload in context. The answer is not simply to deploy more hardware or move everything to the cloud. It is to decide where applications should run, how the infrastructure needs to change and how practical constraints, including power, should inform future investment.
Different organisations will reach different answers, and even workloads within the same business may need different treatment. A heavily regulated system may need to stay on site, while a service with unpredictable demand may benefit from the cloud. Available capacity, resilience requirements and cost will shape the choice.
This is why power now belongs in the infrastructure conversation. It will rarely be the only factor, and it may not be the deciding one. But if it is considered too late, an otherwise sound plan can become expensive or difficult to expand. Bringing it into the discussion early gives organisations a more realistic view of what their infrastructure can support.
Image credit: https://unsplash.com/photos/man-in-pink-button-up-shirt-sitting-beside-woman-in-blue-and-black-shirt-lLrZy195sIU
