Skip to main content

If you’re heading to the cloud, take your organisation with you

(Image credit: Image Credit: Melpomene / Shutterstock)

As the public cloud has seen widespread adoption, DevOps has become a standard approach for organisations seeking to take advantage.

This isn’t surprising – a well-run DevOps function is a remarkably effective way to deliver new business value on a continuous basis, and is a natural bedfellow for any cloud-based initiative. However, businesses can expect too much from it. DevOps alone isn’t sufficient to run a digital-first, cloud-based business, for the simple reason that an organisation is far more than its IT department.

Among businesses that have cloud experience, there’s an increasingly widespread understanding that to really take advantage, the entire organisation needs to be set up to support iterative, continual improvement. Where it isn’t set up like this (often as a result of moving into the cloud too fast), uncontrolled costs can undermine what the cloud should be delivering –  not just lower costs, but also flexibility, higher quality and the ability to develop new capabilities at pace.

Enter FinOps

In an attempt to solve the broad challenges that remain and bridge the gap between the ideal and the reality of digital business, a new concept is gaining ground: FinOps.

The FinOps Foundation summarises it well:

“FinOps is the operating model for the cloud. FinOps enables a shift — a combination of systems, best practices and culture — to increase an organisation’s ability to understand cloud costs and make trade-offs. In the same way that DevOps revolutionised development by breaking down silos and increasing agility, FinOps increases the business value of cloud by bringing together technology, business and finance professionals with a new set of processes.”

It describes a journey characterised by three distinct phases:

1.            Inform: A business-spanning exercise to enable the organisation to understand where its cloud spend actually goes, and to monitor those costs more effectively going forward. This involves establishing a standardised taxonomy across the organisation, so that costs can be broken down by function, technology and so on.

This remedies a common issue; an inability to see which business areas are driving which types of cost. Yet this knowledge is vital if you’re to make the right decisions about how to scale, or how to re-architect infrastructure to stabilise and control cost.

2.            Optimise: Suitably informed, the business can begin to tailor its cloud infrastructure for greater efficiency. Cloud cost structures such as by-the-minute and by-the-resource can be hugely beneficial compared to the block-like cost profile of an on-premise platform, but they’re complex. The upside is that there are often plenty of ways to lower costs once you really understand it all.

3.            Operate: The operate phase should be familiar to anyone working with iterative development cycles. It’s an ongoing feedback loop, as changes are observed in order to inform further optimisation over time.

As with DevOps, which is a cultural exercise as much as a technical one, FinOps demands a culture of ongoing refinement and improvement to analyse the changes made. The key difference is that this culture must span the whole organisation.

We’re all in IT together

At Architechs, we think the FinOps approach makes a lot of sense, because it makes it crystal clear that a successful cloud implementation requires a holistic approach. Responsibility must be balanced across the organisation – and not seen as ‘just something the techies do’ – if costs are not to spiral out of control.

If the FinOps concept makes sense to you too, then we’d like to suggest some practical steps your organisation can take to gain more control over your cloud costs. And in the true spirit of FinOps, you don’t need to involve your devs in all of them…

Pay less for what you buy

To demonstrate FinOps’ business-wide approach, note how the Inform stage empowers your finance team. Every organisation’s use of the cloud will differ, but once the business beyond the IT department understands the resources you’re using (and how much you’re paying for them), it’s time for them to ask providers to lower those costs.

Committing to certain amounts of usage, or signing up for longer periods of time can make a big difference to how much you pay – and is a straightforward way to remove cost quickly. Armed with knowledge and a certain amount of flexibility, procurement departments will find a receptive ear with most of the major cloud providers.

Buy only what you need

Perhaps it’s something to do with the allure of the cloud, with its futuristic promise of limitless resources, but many businesses overspend when it comes to their cloud infrastructure. Of course, there always needs to be some headroom, but there may be ways to reduce the amount of resource you buy without significant impact. For example, moving to virtual machines with fewer CPUs might have a negligible effect on your applications’ performance, but move you down a tier in your provider’s cost structure.

This activity would fall firmly under FinOps’ Optimise stage, and once again, it’s an initiative that shouldn’t keep your developers tied up; your ops team will have all the information required to make improvements like this.

Need less of what you buy

Don’t worry though, FinOps will keep your developers busy too. They should be put to work at the application level, looking for ways to use less resource and re-architecting (where necessary) to make it happen. In other words, their purpose is to provide more bang for your buck.

For example, they might examine how and when a database is used, and consider its impact on cost; reordering or combining processes to reduce database access could conceivably reduce the associated cost by a significant amount.

FinOps is a wake-up call

Many organisations expect a move to the cloud will bring benefits as a matter of course, but the devil is in the detail; without close control over the myriad factors that drive cost, unknown or unexpected issues can scupper an otherwise successful cloud migration. Consider, for example, a situation where an application consumes a particular resource in ways that hadn't been predicted. By the time this happens, it could be too late to reverse out of the cloud and/or re-architect the application.

It’s within this context that we see FinOps as being such a promising development. It’s an overdue wake-up call for any organisation that thinks the cloud is something that can simply be left to IT.

Jan Frydenbø-Bruvoll, MD, Architechs