Industry pundits confirm that cloud computing has not only gone mainstream, but also that nearly all organizations are following the trend. For example, according to recent research, more than 90% of all organizations used public cloud services in 2020, and spent more than $50 billion in the process. This spending is expected to go in only one direction: up.
One of the biggest reasons why companies are making this move is to save money. They intend to lower their operational costs by shifting part or all of their IT infrastructure and applications to the cloud. The strategy must make good business sense and have a clear and compelling ROI, because many CFOs and other top executives have signed off on it.
Wasting Money to Save Money?
But there is an odd variable in the cloud spend equation, one that’s jarringly out of sync with the overall goal of lowering costs. It’s called ‘cloud waste,’ a term that refers to the wasteful spending that occurs under most (if not all) cloud service arrangements. Generally, this waste stems from over-provisioned infrastructure that ends up being unused or under-used.
It’s not a small problem. Industry analysts estimate that roughly 30% of cloud spend is wasted. In 2020, that added up to $17.6 billion, according to the experts at DevOps.com and research from ParkMyCloud. Other market watchers, such as Accenture, put the rate of waste closer to 40%.
One aspect of cloud computing that is getting more recent public attention is the negative effect on the environment. Although most people don’t realize it, the cloud industry is one of the biggest environmental challenges and one of the most overlooked contributors to climate change. Within a matter of years, data centers became one of the largest annual energy consumers and emitters of carbon pollution. Data centers currently represent roughly 3% of the world’s total energy consumption and they emit nearly 100 million metric tons of CO2 into the atmosphere each year.
It’s Time to Curb Cloud Waste
Cutting wasteful cloud spending is obviously a good idea. So is doing the right thing by the environment. But organizations of all kinds now rely heavily on their cloud infrastructures, so making changes can be complicated and risky.
It doesn’t have to be that way, however. This 10-step process allows organizations to lower their cloud waste while maintaining their operational efficiency and agility and reducing their carbon footprint. It’s a straightforward approach that should be applicable for many organizations. If we all strive to reduce our cloud waste, we’ll see better business results and big, positive environmental impacts.
Define the Problem
Before attempting to address any operational issue, leaders and their teams need to understand exactly what the issue is. To do that, they need to define it, including its general nature, its negative impacts, its component elements and their likely sources.
In the definition process, teams should focus on the primary problem of wasteful spending, of course, but also be sure to include the resulting secondary problems excessive cloud use is causing to the environment. The following simple definition is a good place to start.
Cloud waste refers to purchases of cloud resources that go unused or underused. The resulting wasteful spending increases operating costs without adding value, and precludes using those same funds in other areas of the business. Furthermore, cloud waste, whether generated in private or public cloud environments, needlessly contributes to the negative impacts that data centers have on the environment (higher energy consumption, more CO2 released into the atmosphere and more).
Educate, Communicate and Equip
Before moving ahead with any specific actions, it’s helpful to first raise the issue with staff at appropriate team meetings, in corporate communications and in other venues. Defining the problem naturally leads to identifying likely causes and contributing factors. That involves pointing out where cloud waste typically happens in an organization’s processes and workflows, and by which departments, groups and roles.
Rather than framing this as a negative that needs to be rooted out and eliminated, frame it as a positive initiative, one aimed at curbing costs without impacting operational effectiveness. The idea here is to educate staff about an ongoing problem with both financial and environmental ramifications and to enlist their support in addressing it. A big part of this is equipping developers with tools that provide them greater visibility into what it takes to run their applications at the levels required. Developers also need better controls so they can make resource adjustments quickly and efficiently or deal with potential downtime while changes are being made. With better tools and systems that create better visibility, developers can make smarter resource decisions.
Build a Cloud Efficiency Culture
When developers, DevOps teams and others are on the hook to make sure a critical application always works as needed, the nearly infinite resources available in the cloud are awfully tempting. Start building a culture based on the assumption that using ever-increasing levels of cloud resources isn’t the answer. In fact, it’s lazy, and financially and environmentally irresponsible.
When your team is turning on a cloud resource, encourage them to use only the resources needed to get the job done – not double or triple that amount. Discourage activities like instance hoarding; don’t set the default to on-demand services. Make sure that your team or your cloud provider shuts resources off when not in use.
Another important step for companies is making sure that it’s okay for developers to step back from their day-to-day duties and research and experiment with new and better ways of doing things. These things usually take time and thoughtful consideration — such as investigating the feasibility of moving from AMD64 to ARM chips for greater power efficiency. Write this ‘quality time‘ prerogative into job descriptions and performance review processes. And when good ideas surface, don’t be afraid to make the change.
In short, set efficient cloud resourcing as the goal. Don’t solely rely on a top-down mandate, but instead, also cultivate and encourage the cultural shifts that will contribute to this becoming a bottom-up, grassroots push in your organization. Challenge your team to make changes and make this new approach a point of pride with them.
Conduct an Assessment
Next, determine the present state of your organization’s cloud waste problem; where it’s happening and to what degree. To do that, examine all IT operations and identify any systems, processes, or workflows that rely in any way on cloud-based elements.
Once there’s a comprehensive list of cloud-dependent systems and processes, determine the contractual arrangements covering those resources. Establish a priority order, determining which cloud-consuming systems and processes are most critical to the business, and which ones are more tangential.
Another helpful tip is to get a sense of how much cloud waste your peer companies are generating and what is happening across your industry segment. If this cannot be obtained through market data or analyst reports, anecdotal information from industry contacts can at least give you a general lay of the land.
Next, gauge the amount of cloud-based resources an application, system or process is currently provisioned versus how much it actually requires. Subsequently, you should determine which systems and processes are using (and being billed for using) the less-expensive “reserved instances” versus the more expensive on-demand variety. And last, but far from least important, compare the anticipated need versus real usage.
While there will be room for debate in these resourcing discussions, the first step is to identify obvious outliers – apps, systems and processes being supported by way more cloud resources than they will ever reasonably need. It’s critical to have this list in hand before moving to the next steps.
Set Realistic Goals and Communicate Them
Cloud migrations are complex. They require a fair amount of time both for planning and execution. Ramping up cloud usage and becoming enmeshed in cloud waste situations also took some time. So, it will also take time to pare back cloud waste. Plus, there are risks – operational and reputational – that companies face when altering their cloud service arrangements. Handling those risk exposures thoughtfully and carefully is a must.
It pays for organizations to be just as careful when taking reduction actions as they were during their initial cloud migrations. It’s smart to set modest goals, such as an initial 10% reduction. It’s also a good idea to pick a reasonable timeframe – six months, for example.
Then, as is the case with any corporate IT initiative, the overall project, along with its goals and timeframe, needs to be communicated to all who will be directly and indirectly involved. Regularly communicating updates on progress toward the goal is also a best practice.
Initiate Actions and Get Specific
With the necessary assessments and planning complete, it’s time to make cuts to cloud services. Systems and processes that are clearly over-provisioned (either unused or underused) should be the first targets for reduction.
Next, make sure that static workloads are provisioned appropriately. Likewise, with dynamic workloads, develop the most accurate predictions of their fluctuations in demand for resources that you can, and provision accordingly.
Next, target those messy, harder-to-predict items that exist in every organization. This includes resources your team(s) thought they needed but actually didn’t, as well as artifacts left over from previous IT efforts. Also in this category are things that were once important but aren’t any longer, but are still running or being paid for. Top offenders here can include things like suboptimal container settings, old images, orphaned volumes and widowed instances. All of these items should be cleaned up, pronto.
If cloud waste is happening in the organization’s private cloud, the change can be precise and done without any outside involvement. If it’s a change to a public cloud arrangement, then contractual items, such as provisioning types and mounts, and service level agreements need to be changed or updated with the cloud services provider.
Once this house cleaning has been completed, useless and vestigial items will be gone. You’ll also have an accurate estimate of your actual capacity requirements (types and amounts). Then, whether it’s in your own private cloud operation or with your cloud services provider, you can take further action to close the gap between present provisioning and what’s actually required.
Stay Focused and Be Responsive and Efficient
Organizations change, so their cloud computing requirements do, too. Stay focused on these changes; not just new needs, but old arrangements that are no longer relevant. Make sure your team is just as responsive in ushering in the new as they are with getting rid of the old and outdated.
Ensure that your service provider – and/or the tools your team uses – guarantees your organization is able to spin up services smoothly and cost-effectively when needed. And make sure the inverse is true; make sure that you can spin resources down rapidly and efficiently after the need for them subsides. Leverage container orchestration tools and technologies, such as auto-scaling, to ensure that the required resources get automatically allocated and released as dictated by workload needs.
Track and Measure Progress
As with any other organizational initiative, efforts to reduce cloud waste will only work if teams set measurable goals, track them closely and try to make progress toward them every day. This may sound simple, but teams should make specific goals, ones that are narrow enough for team members to focus on while also keeping the larger team and organizational goals in mind. The goals need to be measurable, and include clear definitions of what will be tracked and what progress will look like. The goals should also be realistic. In other words, they should be attainable within a given time period. With cloud waste reduction, teams could set specific percentage-reduction targets for the year broken out by quarters. Lastly, it is helpful to align a specific program, like cloud waste reduction, with a broader corporate initiative or initiatives (such as a green or operational excellence program).
Don’t Lose Sight of the Movement Ecosystem
As mentioned above, the data centers from which cloud services emanate are enormous consumers of energy and, as a result, release lots of CO2 into the atmosphere, contributing to climate change.
The fact that data centers have massive carbon footprints used to be a little-known fact. But that is changing. Market researchers and industry analysts are getting the word out and increasing mainstream awareness of this problem. Various mitigation actions are initiated by leading cloud service providers, as well as by a growing roster of companies and individuals.
Get involved with these efforts and lend your organization’s and your team’s voices to these efforts.
Recognize and Reward Success
This is an extension of step three, yes, but it is also worth calling out separately. The impact that cloud service efficiency can have on an organization’s operations and its bottom-line results can be significant. The environmental impact, while perhaps less tangible in the short term, is nonetheless important to the industry as a whole, and can have a big impact in the long term.
Team members who focus on cloud service efficiency goals and take the actions needed to achieve them and reduce cloud waste, should be recognized and rewarded. Not doing so is a lost opportunity for organizations to advance their efforts to not only do well, but to also do good.
From corporations to nonprofits to government agencies, our collective appetite for cloud computing resources isn’t likely to diminish anytime soon.
As market data makes clear, the cloud waste problem – the wasteful spending – is also predicted to continue its rapid rise, and with it, the associated financial and environmental issues.
We can change this trajectory. We can reduce the amount of money we’re wasting, and the amount of pollution pumped into our environment as a direct result of data centers. All it takes is a shift in focus, and a commitment from all of us to embrace a new way to consume cloud resources.