Why Trust and Autonomy Matter for Cloud Optimization
In a tough global economic environment, optimization is increasingly critical, but at the same time, regulations concerning security and data privacy create trade-offs that need to be sensitively balanced.
To make things even more challenging, these are trade-offs that span functions and departments. Sure, you can pretend cloud optimization is only an operations or IT task, but with such an attitude you’re inevitably going to fail. Effective optimization requires a cross-functional mentality where operations, development and finance collectively acknowledge their interdependent goals and challenges.
Failing to adopt such an approach can cause optimization efforts to backfire: conflict created by these interdependencies slows decision-making and action, causing costs to spiral and innovation to become hampered and severely restricted.
The consequences of this can lead to distrust and disillusionment across an organization. That’s something few people — whether they’re accountants, software developers or solutions architects — will want to be a part of.
So, when it comes to optimization and implementing CloudOps and FinOps successfully, how is it possible to foster the trust required?
Bridging the Gap Between FinOps and DevOps
Trust is a product of organizational culture. It requires leadership. But it also requires an understanding of the role each team plays in the organization’s mission.
Granted, there’s a “healthy tension” between developers, operations engineers, and the business side of an organization, Shon Harris, developer relations manager for strategic alliances at Spot by NetApp, told The New Stack.
“The business has a mandate to make the best business decisions for their shareholders and stakeholders,” Harris said. “Dev has the mandate to push out the best product; Ops has the mandate to make it work.”
The reason cloud optimization is so important is that it ties together two impulses: first, the more conservative drive for efficiency, and second, the drive for innovation and invention. Without the right levels of alignment or a lack of tooling, these two things can quickly come into conflict with one another. But they are, in fact, absolutely dependent on one another.
When you get the balance right, Harris suggested, “you’re not solely focused on efficiently maintaining your operational workloads. You’re doing that but also consistently executing on how ‘we can go and build other cool stuff that people want to buy and really grow our business out in a way that we hadn’t thought of previously.’”
These two impulses aren’t going to disappear. However, the strategy — and the right tooling — can make them work in tandem. It starts with providing software developers with a solution that supports the performance they need from a specific application in the most efficient way, making it possible to “do more with less” at every level.
Perhaps it’s by bringing optimization and efficiency together that the industry can make the shift from seeing FinOps as a business exercise to seeing it “integrating optimization in a way that bridges the gap between Finance and IT,” as Harris described it.
In other words, it’s making FinOps something that is built into the way an organization views and understands the relationship between finance and technology.
“Until FinOps becomes business as usual, you’re still going to have a disconnect,” Harris said.
Building Trust through Shared Context
Having a common context for discussions can help foster that understanding — and, ultimately, that trust. One way to create that common context is by using the same tooling.
“It used to be that you had to use so many different tools and you had to give access to Dev, and you had to give access to the business,” Harris said.
However, with something like the Spot by NetApp portfolio of CloudOps tools, he suggested, the issue of fragmentation and context switching can be overcome: “The fact that we can do that right out of the box with one pane of glass sets us apart in a way that makes that natural tension a little bit less draining on the people who have to manage the relationships.”
Pulling different tools together inside a single suite creates a common context that makes it easier for the different functions that need to worry about cloud — whether that’s Ops, Dev, or finance — to interact with one another. This is a solid basis for trust; it ensures that the decisions made by different stakeholders can be more clearly communicated and explained.
If, for example, finance wants to know why they’re having to manage an unexpected $50,000 jump in their Amazon Web Services bill, Spot features tools that not only make it easier for Ops to actually go and have that conversation, but also do something about it — with Eco, for example, which is designed to help users get more their cloud investment by automating resource allocation and making reserve capacity more efficient.
These sorts of tools have the potential to be incredibly powerful, making it possible for teams that might previously have very different perspectives and ways of talking about things to work much more effectively. But more than that, it can also help unlock a collective sense of what optimization is for.
“When it comes to optimization, everybody in the organization … has to come together and understand what the long-term goals of that optimization are,” Harris said.
In theory, this is something that should be established upfront; in reality, this is difficult. Having a shared space — like a tool or set of dashboards — is one effective way of reaching this shared vision.
Building Team Confidence
With organizations beginning to ask teams to do more with less, optimization — of all kinds — is going to become a vital part of what technology teams (development and operations alike) have to do. But for that to be really effective, team autonomy also needs to be founded on confidence — you need to know that what you’re investing time, energy and money on makes sense from the perspective of the organization’s wider goals.
Fortunately, Spot can help here too. It gives teams the data they need to make decisions about automation, so they can prioritize according to what matters most from a strategic perspective.
“People aren’t really sure what’s going to be happening six, nine, 10 months down the road.” Harris says. “Making it easier for people to get that actionable data no matter what part of the business you’re in, so that you can go in and you can say, ‘Here’s what we’re doing right, here’s where we can optimize’ — that’s a big focus for us.”
One of the ways that Spot enables greater autonomy is with automation features. Its artificial intelligence capabilities, based on years of cloud performance data, take care of foundational infrastructure concerns so teams can attend to value-adding activities and initiatives.
“You don’t want to be thinking about the infrastructure,” Harris said. “You’re either going to over-provision or under-provision.
“We come in and say, ‘Let’s take that off your plate and let you go in and write the code that you’re good at, and we’ll tell you what you need to be running.’ We’ll tell you what instance works best based on your historical usage. We’ll tell you how to balance out your Kubernetes cluster based on the metrics of your application.”
Here, Harris is referring to Spot’s Ocean product, which helps teams better manage and optimize container workloads, but there are similar products across the Spot platform.
Eco, for instance, helps organizations manage cloud capacity, automatically selling or acquiring reserve capacity when needed, while Elastigroup allows teams to automatically scale workloads by maximizing spot instances and reserved capacity while maintaining reliability.
However, while these automation capabilities are essential to understanding Spot, what’s particularly significant is that it’s designed to allow users to respond to the specific needs of their organization. Organizational contexts, after all, are always unique — automation always needs to be done with that in mind.
This informs more than just the development of the product: it’s also built into the way the Spot team onboards new customers. The company, Harris says, will typically recommend that customers wait two weeks before programming automated tasks. This is largely to ensure that users are actually engaging with the specifics of their organizational context and they actually understand what the data they’re seeing means.
“We really want our customers to take a balanced approach to disseminating the information that they’re getting,” he said. “If you don’t know what you’re looking for or what you’re looking at, you’re going to have problems.”
Creating a Shared Language
Cloud optimization can’t be reduced to the purely technical or the strategic. It sits at the intersection of different teams, perspectives and skill sets. And although trends like DevOps have been insisting on the importance of mindset shifts and cross-collaboration for more than a decade, what’s particularly important — perhaps challenging — about FinOps and CloudOps is that it requires collaboration across teams that may sit some distance from software engineering.
Spot by Netapp has been developed with a clear sense of that very challenge; as business and technology functions become increasingly interdependent and as pressure increases on everyone to do more with less, finding common ground — a shared language — is going to be fundamental to future success.
Learn more about cloud optimization in this recent episode of The New Stack Makers podcast: