Hyperscalers are delivering unprecedented compute power and redefining how products and services are built and delivered to customers, but with the risk of uncontrolled spending.

Even worse, we know many enterprises that have shifted workloads to the cloud are seeing rising IT costs without the ROI to justify the move.

But the business imperative to access the power and agility offered by the cloud is only increasing. When it comes to their modernisation efforts, many enterprises have already picked much of the low-hanging fruit.

For them, enabling the next phase of digital transformation means modernising their mission-critical infrastructure. Given that these systems are often particularly resource-intensive, taking smarter approaches to design and management of these systems is essential to controlling costs and eliminating waste.

The wild west of the cloud

IDC research shows that almost nine in 10 APAC IT leaders confess to highly concerned about the growing technology investment – particularly in cloud – needed to stay competitive.

There are a few reasons why so many enterprises are seeing their cloud spend drift higher than the projected total cost of ownership.

A big one is of course the rise of shadow IT, but getting the balance right between locked-down central governance and the agility needed to keep pace with business today is a discussion for another time.

For many enterprises, a quicker win for reining in cloud costs is tackling waste.

AWS itself has estimated 35% of cloud spend is wasted. This manifests in a multitude of ways, including:

• Poor planning resulting in avoidable egress or transfer fees for moving data off the cloud or between regions
• Running queries at the wrong time
• Not putting applications in containers
• Not rightsizing virtual machines (VMs)
• Not turning off VMs used for testing and development when not in use
• Not eliminating inactive storage

Egress and transfer fees can be particularly dangerous, as they are charged in arrears and can be unpredictable, potentially leading to awkward conversations with the CFO when the bill arrives.

Famously, a few years ago it was leaked that, in 2018 at least, transfer fees accounted for a full 6.5% of Apple’s total AWS bill – more than $US50 million.

Meanwhile, at least one researcher found two in five VM instances are at least one size larger than needed, estimating that users are paying up to 75% more than they should be for their VMs.

Similarly, keeping VMs used for testing and development running outside of developer working hours can easily result in users paying more than double what is necessary.
None of the above is inevitable. But it is made much more likely by taking a lift-and-shift approach to cloud modernisation.

Knowing what your cloud is for

Parkinson’s Law states that work will expand to fill the available time, and there surely is an equivalent in IT – user demand will expand to consume the available resources.

Simply recreating on-prem architectures in a new cloud environment inevitably results in inefficiencies.

It’s why Kyndryl partnered with IDC to define the most productive approaches to mission-critical modernisation.

What is overwhelmingly clear is that effective digital transformation requires a much more nuanced approach than simply “cloud-first”.

We see a business outcome-driven modernisation as key to not only unleashing the full business potential digital transformation offers, but also minimising needless cloud spending – maximising ROI on these projects.

When we begin working with clients whose cloud migrations had been driven more by market hype than by clear-eyed organisational objectives, it is not unusual for us to be able to cut 40% from an enterprise’s cloud spend.

Doing this calls for two vital capabilities – visibility and automation.

Seeing and doing

You cannot manage what you cannot measure, so having visibility is crucial.

There are many aspects to this, from knowing who is using how many resources, through to being able to predict usage and cost spikes.

What is needed is a single pane of glass, a centralised location where an enterprise can get fine-grained, real-time detail in what resources are being consumed where, and what that means for the budget.

Knowing how you are consuming resources across your operation is the first step.

Just as crucially, enterprises need defined policies to minimise cloud waste, such as when large data queries are run, shutting down VMs after hours, rightsizing instances or eliminating inactive storage.
Understanding what efficient operations look like for your enterprise – and where that is not happening – allows those policies to be devised, implemented, monitored and optimised.

Given the size, speed and sprawl of operations though, the enforcement of these policies must be automated if they are to be as effective as needed.

Making the change

Hyperscalers are vital partners for enterprises looking to transform how they operate and how they serve their customers, but they are not incentivised to help those enterprises minimise their cloud costs.

And while many enterprises struggle to gain visibility into one cloud, running multi-cloud and hybrid operations multiply the complexity and difficulty.

A new way of working is needed. That new way is called Finops, driving sustainable and systemic change in the way we manage IT usage and spending. FinOps is designed to optimise cloud investments by bringing together financial, technical and business functions to drive financial accountability and create a cost-conscious culture for enterprises.

It’s how enterprises can create the discipline they need to use their newfound cloud powers to, instead of laying waste to their IT budget, change the world instead.