O v e r v i e w

This e-book outlines how this application infrastructure cycle unfolds. It gives examples of how innovation and the excitement surrounding new and better ways of doing things has routinely turned into excess and waste.

It also highlights ways that AI and machine learning can be used to break this problematic cycle and prevent the negative outcomes.


For even the casual observer of the adoption of new technology in business, it’s easy to see a pattern that begins whenever there is a major advance in application infrastructure.

IT capabilities, especially an IT team’s ability to deploy and leverage infrastructure quickly and efficiently, have long been a key to achieving competitive advantage. That’s why many organizations push to aggressively adopt new advances in infrastructure technologies.

But it’s that very energy around leveraging new infrastructure that consistently triggers a cycle that turns infrastructure velocity advantages into liabilities.

The cycle goes like this: IT teams embrace the new technology, achieve initial success, and then look to do more with it. What follows is deployment sprawl with its attendant waste and spiraling costs.

That’s why infrastructure breakthroughs haven’t always lived up to their hype or delivered on all their promises. They just become part of the next phase of an endless chase for operational improvements, with lasting success always remaining just out of reach.

What You'll Learn

This Breaking the Cycle e-book will help readers to understand:

  • The nature of this cycle, its distinct phases, and their operational repercussions
  • The reasons why it always follows application infrastructure innovations
  • Examples of this cycle from recent technology history
  • How the same cycle is presently unfolding in the containerization and Kubernetes space
  • How artificial intelligence (AI) and machine learning technologies can be used to break the cycle and prevent organizations from incurring its downsides

This e-book also covers how StormForge’s solution can effectively break this cycle when it comes to the adoption of Kubernetes and containers, enabling organizations to take full advantage of these innovations without getting derailed by cost and complexity challenges.

It’s time for you and your IT team to stop the endless chase and “grab the brass ring” of success with your application infrastructure.

How the Cycle Begins

As mentioned earlier, infrastructure velocity is a fairly reliable leading indicator of differentiation and competitive advantage. Software is of course a big part of this equation. However, applications cannot run on their own; they need to be supported and delivered by application infrastructure. That’s why forward-thinking enterprises make it part of their strategic direction to pursue greater infrastructure velocity. This push has fueled some of the greatest technological innovations in delivering infrastructure.

When an infrastructure breakthrough happens, it gives IT teams radically different ways to do things, especially new ways to meet their users’ application delivery needs.

As word spreads about the new infrastructure model, excitement among IT teams grows. Trials and proof of concept exercises turn into small, initial deployments in test or lab environments, and finally in production environments.

And “Voila!” The technology works as promised. It’s faster and easier to deploy than the legacy model, and it delivers the expected results. With the success of their initial deployments, IT team members, and even business leads, push for more deployments in other areas of their organizations.

Sprawl Happens

With deployments gaining momentum, getting larger, and spreading throughout an organization with more people and more departments using the new technology, sprawl becomes a reality. Sprawl is the tendency of easy-to-deploy resources being over-provisioned or used when they are not actually needed. It often results from administrators having too much to manage, or resources that are no longer needed not getting decommissioned. Sprawl results in waste of both budget dollars and person-hours, eroding the financial and operational benefits promised by the new infrastructure technology.

In addition to sprawl and waste, these quantum leaps in application infrastructure technology have another downside. Their deployment puts new strains on organizations. These pressure tests quickly uncover weak points in IT operations that serve as inspiration for the next cycle of innovations. Patching those holes and or closing those gaps become the focus, and the cycle continues, with most of the benefits still out of reach.

One Example – Virtual Machines

The following, fairly recent example is one that many readers have likely experienced firsthand.

While some forms of virtualization have existed for nearly 60 years, virtual machines did not make their debut in the enterprise until the early 2000s. In less than a decade, they revolutionized expectations for infrastructure delivery and management in businesses of all sizes and types.

By the late 2000s, rapid provisioning of virtual servers boosted the efficiency of infrastructure teams. By that time, true to our description of the phases of this cycle, a new set of phenomena was occurring. It was the dual problem of sprawl and waste. Virtualized servers were sprawling across organizations. They were being deployed everywhere, even where business needs didn’t make them a requirement. Plus, since they were so easy to deploy, over-provisioning was common. That’s the over-commitment of resources in a shared environment to ensure that critical applications do not suffer a loss in performance or uptime. In short, over-provisioning created significant amounts of waste.

In response to these new challenges, a wave of orchestration, monitoring, and analytics technologies popped up at the end of the 2000s.

Unfortunately for most organizations, at that point in time the only real safety net in this paradigm was that there was a physical limit to the size of their environments. At some point, even the largest infrastructures became “full.” That forced organizations and their teams to find ways to curb sprawl and reduce waste.

Cloud and Containers: New Innovations, but the Same Cycle

In terms of the overall impact and scope of change, the biggest infrastructure innovation to date has been cloud computing, along with deeper layers of abstraction like containerization. The public cloud takes the proverbial lid off of the enterprise data center, making it easy to access essentially unlimited compute power and resources.

As our innovation-waste cycle predicts, exponentially expanded access to readily deployed computing resources inevitably leads to sprawl and its close cousin, waste. These negative outcomes are clearly being borne out as enterprises continue to increase their usage of all cloud deployment models – public, private, and hybrid.

As far as infrastructures go, the cloud really is a different animal. Virtual server environments from a decade ago might have had thousands of virtual machines spread across a small number of geographic locations. Compare that to an organization that today is shifting to the public cloud and Kubernetes. That organization’s environment can quickly encompass even more virtual machines hosting tens of thousands of containers spread across several providers and/or geographic locations.

The Cycle Keeps Turning

Managing all these abstracted and remote resources effectively is certainly difficult. In fact, it is proving to be that next phase in the cycle in which new and different pressures exerted on an organization’s deployment framework and team exposes new, previously hidden or non-existent weak points.

Prioritizing velocity in a business environment that rewards software-driven operations is tempting. However, doing so without mature management capabilities, for example observability and resource optimization, is a tall order.

Many IT pros who are dealing with these challenges look to the Cloud Native Computing Foundation (CNCF) for clarity and direction. But a quick survey of the CNCF landscape reveals significant complexity in this area of technology – even just the additional software and tooling required to support this cloud native development. That combined with the technical sophistication of all the supporting capabilities and functionality involved drives deployment and management complexity even higher.

Whether they recognize this challenge or not, many organizations are plowing ahead with their transitions to the cloud and deployments of cloud-native applications. For now, they have chosen to go for the benefits and competitive advantages. They assume that sprawl, over-provisioning and rising costs are simply the price they must pay in order to compete.

Where We Came From and Where We’re Headed

For many organizations, their transition to the cloud is serving as an inflection point. Since sprawl, over-provisioning, and waste are such easy traps to fall into with the cloud, these problems are coming into sharp focus, with many organizations now saying that reducing cloud waste is a top priority.

It’s their way of saying that they want to break the cycle – or at least slow it down. To make that happen, a good place to start is to step back and get a broader perspective of the bigger picture. Consider where this market has been and where it is clearly headed.

We start with bare metal – those big, expensive pieces of hardware. Then, as we move towards virtualization, cloud adoption, containerization microservices, and eventually to serverless frameworks, we simplify in some ways, but at the same time the complexity of managing it all continues to increase.

As we continue to decompose applications into their modularized parts, staying on top of IT operations grows ever more complex. It quickly goes beyond human ability to understand where all the various important pieces are, how and when they are being used, and what their costs are. Then there’s the need to be able to spot and address issues before they blossom into outages or other user-impacting events. Doing that with speed and accuracy in cloud environments is proving – once again – to be one of those weak spots uncovered in one of this cycle’s new stress tests.

Instead of having history repeat itself, and kicking off a whole new round of exploratory innovation, many IT leaders have had it with sprawl and waste. They’re saying “How about this time, we get the application infrastructure right?”

So, as IT professionals, how do we simultaneously prioritize velocity, quality, performance, and cost control?

This Time It's Different

The good news is that innovation has once again come to the fore. It’s not innovation that will take us in yet another, entirely new direction. Instead, new technologies are focused on enabling enterprises to get maximum value from the infrastructure model they have just embraced. For many, that’s the cloud, with microservices, and containerized applications running on Kubernetes.

New operational models and maturing technologies, such as machine learning, can help to sort out the complexities of the cloud-native, containerized world. These tools are geared toward helping humans make better and faster decisions, and to make it easier for them to understand and manage both capital and operational costs.

There are also new frameworks, for example financial operations, or FinOps for short. It’s an operational model in which cross-functional stakeholders join forces utilizing both processes and tools to understand and manage operational expenses. FinOps provides an approach for getting visibility, optimizing, and operating within this new world.

Let’s take a deeper look at each of these advances next.

Machine Learning: An Application Optimization Crystal Ball?

Machine learning (ML) technologies have risen in popularity due to their general effectiveness in dealing with complexity.

Recent history has given us a preview of the future of machine learning-driven solutions. Hardware vendors in the storage market, for example, have been implementing ML to predict impending component failures and forecast resource constraints for a decade. For many organizations, these proverbial “storage infrastructure crystal balls” have fundamentally changed how they manage their storage requirements – now with automation and more speed and precision.

As organizations shift to using Kubernetes, it seems obvious that one way of pre-empting waste and controlling costs while ensuring availability and performance for applications is to implement ML-driven testing and optimization within software development and delivery models, such as CI/CD.

This approach creates an “application optimization crystal ball,” both recommending fixes to inefficiencies in already-deployed applications, as well as serving as a quality gate ensuring any new deployments meet the business’ requirements.

It is widely recognized that Kubernetes is rife with complexity and has plenty of observability challenges. Those challenges are compounded by an industry-wide shortage of engineers with relevant skills and experience. That makes it a perfect place to apply machine learning. In fact, deploying Kubernetes at scale without ML is a recipe for certain disaster.

As IT professionals, we must find an effective way to break the problematic innovate/overdo it/break it/innovate again cycle that we have been in for so long. Machine Learning is a key technology piece of the puzzle, now we just need a framework to bring it all together.

The FinOps Framework: Best Practices for Cloud Financial Management

Fortunately, there is such a framework. The FinOps Foundation, a part of the Linux Foundation alongside CNCF, is a community-led organization with the mission of promoting cloud financial management best practices. They’ve developed a best practice framework for the journey to cloud financial management maturity.

While machine learning gives you the tools, the FinOps lifecycle gives you the process and framework for applying those tools effectively and breaking the cycle of sprawl and waste. There are three phases of the FinOps lifecycle: Inform, Optimize, and Operate.

Inform is all about visibility. It’s about understanding where the cloud spend is going, empowering your engineers with the information they need to make smart business decisions, and understanding the tradeoffs between cost, performance, and the time and effort needed to achieve both.

Optimize is about finding the sweet spot among those tradeoffs. It’s about meeting service level objectives at the lowest possible cost with the least amount of effort. It’s about right-sizing for the most efficient use of resources, and also ensuring the decommissioning of resources that are no longer needed.

Operate is about continuous improvement and efficiently managing the resources you have in production. Building an efficiency-oriented culture, adopting best practices, and establishing governance procedures, such as a Cloud Center of Excellence team, are all key. Operate is all about building efficiency into your day-to-day processes, for example establishing Continuous Optimization as a regular part of your CI/CD workflow.

StormForge: Breaking the Cycle through ML-powered Optimization

The continual boom/break/re-innovate cycle described in this e-book creates multiple problems for enterprises and other organizations. And quite frankly, it’s exhausting for their IT teams. A study by technology company D2iQ found that 51% of developers and architects say building cloud-native applications makes them want to find a new job. It’s high time we all got off this merry-go-round, and StormForge offers an effective way to do that.

According to a study by D2iQ


of developers and architects say building cloud-native applications makes them want to find a new job.

StormForge solves this complex, multi-dimensional optimization predicament with a combination of scalable load testing-as-a-service and machine learning-driven rapid experimentation. It’s an approach that gives IT team members the ability to efficiently and proactively tune the parameters of applications running in Kubernetes to align an application’s performance and costs with the business’ needs.

The following provides a closer look at the StormForge solution’s combination of features and benefits.

The StormForge Solution

Enabling Apps to Shine in their K8s Containers

With the StormForge solution, IT teams can ‘get it right’ with containerized and microservices-based apps running in Kubernetes environments.  StormForge helps IT teams to identify and address application cost, performance, availability and stability issues before they become problems that impact users, customers or the organization.

StormForge delivers Automated Kubernetes resource efficiency at scale, employing machine learning to accelerate and maximize the competitive advantages organizations gain from their cloud-based deployments.

The StormForge solution is designed to help organizations realize the promise of Kubernetes and enable innovation by automating resource optimization. Leveraging leading edge data science techniques and advanced machine learning technologies, StormForge proactively ensures application performance and developer velocity while reducing cloud application costs by 50% or more.

StormForge uses patent-pending machine learning to automatically find the optimal configurations for applications before deployment. That saves customers time and money while ensuring application performance and resiliency, and allowing developers to focus on innovation, allowing their organizations to accelerate past their competitors.

StormForge’s offering is the only optimization solution that is purpose-built for apps running in Kubernetes environments. Providing unique visibility into cost/performance trade-offs, the solution helps customers to proactively ensure operational efficiency. That same visibility, combined with powerful automation, enables IT teams to make intelligent business decisions about trade-offs between cost and performance – without the time-consuming and error-prone trial-and-error processes that many teams struggle with today.

A Look Under the Hood

This first-of-its-kind solution enables intelligent business decision-making by:

  • Eliminating time-consuming manual tuning with ML-powered rapid experimentation
  • Ensuring performance and reliability while minimizing costs and improving developer velocity
  • Empowering developers to make smart resource decisions without affecting velocity
  • Tuning applications automatically for any metric available via existing telemetry

    The StormForge solution positions IT teams for success with efficient and proactive resource optimization through:

    • Predictive, multi-objective. ‘what-if’ analyses to understand application behavior before deployment
    • Integrated performance testing and application optimization in a single platform
    • Smooth, ‘shift left’ incorporation of continuous optimization into organizations’ CI/CD pipelines

      Because the solution is purpose-built for Kubernetes environments, it enables IT teams to:

      • Offload K8s and cloud-native complexity for production applications
      • Avoid vendor lock-in – the solution runs anywhere K8s does and in any CNCF-certified distribution
      • Incorporate intelligent scaling, with automated HPA tuning for more efficient scaling behaviors
      • Leverage the solution’s built-in intelligence to overcome the widespread K8s resource and skills gap


      The time is now to stop the repetitive cycle of infrastructure sprawl and waste described here.

      With the StormForge solution and FinOps best practices, IT teams and their organizations can do that. They can attain the full promise of cloud architectures, cloud-native apps, containerization and Kubernetes. They won’t have to wait for some future innovation to patch up the problems they’ve experienced with the K8s deployments.

      In summary, with the StormForge platform, IT teams can leverage intelligent automation to effectively identify and address application cost, performance, availability and stability issues before they become problems that impact users, customers or the organization.

      Explore StormForge

      Discover how StormForge can help automatically improve application performance and cost efficiency in cloud native environments. 

      See a DemoStart Trial