In 1921, as copper deposits in the mine became increasingly limited, Takeuchi shifted his focus to manufacturing and formed Komatsu Ltd. (link resides outside of ibm.com). It goes without saying that technology has evolved dramatically since then. But, since the beginning, Komatsu has been committed to implementing the latest technology to improve operations and empower its team. Today this commitment extends to its public cloud strategy.

Like many other IT organizations, the infrastructure team at Komatsu historically relied on disparate monitoring tools, user complaints and manual intervention to address performance issues. They lacked the visibility they needed to identify instances of overprovisioning. A business impact use case could take the team hours to resolve. This kept the team in a cycle of fighting fires and prevented them from directing a more significant portion of their time to advancing strategic initiatives.

Moreover, the team had no way to reliably assess the impact of a resourcing decision before it was taken. All they could do was monitor the impact after they allocated resources. As the team shifted their focus to the public cloud and began migrating all their on-premises workloads to Microsoft Azure, they knew they could not sustain this manual approach. They needed to be able to proactively prevent performance issues, reduce waste and minimize time spent on addressing user complaints. This is when they turned to the IBM® Turbonomic® hybrid cloud cost optimization solution.

Once they implemented Turbonomic, the team was finally able to break this cycle of fighting fires. “IBM Turbonomic provides a proactive approach to avoiding performance degradation. It has allowed us to detect issues before they have actually become issues,” explains Matthew Koozer, Infrastructure Architect at Komatsu. Armed with Turbonomic’s full-stack visibility, the team at Komatsu is now able to quickly assess potential performance risk and identify the optimal way to reallocate resources to eliminate that risk without compromising performance in another layer of their technology stack.

Full-stack visibility was just the beginning in Komatsu’s journey to continuously ensure performance while reducing cost. The next step was to implement Turbonomic’s AI-powered resourcing recommendations. The team began with manual execution of those recommendations. They quickly learned that Turbonomic’s recommendations were not short-term fixes that enhanced just one layer of their tech stack. “Turbonomic shows us how a decision is being driven based from a storage perspective—the IOPS being driven to a specific storage solution within our public cloud offering—but its decisions also considers DTU utilization from a database perspective, and whether the issue is related to memory or CPU consumption,” says Koozer. Because they considered Komatsu’s full tech stack, these resourcing recommendations empowered the team to improve performance across their entire environment. Soon, the team was ready to explore automation.

Presently, Koozer and his team have a combination of automation and manual execution of resourcing recommendations. They primarily rely on Turbonomic’s automated resourcing actions in Azure to optimize servers, storage and databases. Turbonomic provides recommendations on where and how to reallocate resources in real time. Then, the team has predetermined time blocks where those actions are executed with minimal impact to the business. After seeing early success with automating resourcing actions, the team now relies on Turbonomic to execute resourcing actions automatically for a very large portion of the organization, without the IT team being involved at all.

Once they had a comprehensive view of their environment, Koozer and his team identified numerous workloads that were overprovisioned and developed a clear plan to right-size them. Historically, application owners resisted resizing efforts, but with Turbonomic, Koozer and his team had the data they needed to show application owners that overprovisioned workloads could be right-sized effectively. “Shifting to a configuration with less memory or less CPU helped us significantly reduce waste across our Azure environment without sacrificing performance,” explains Koozer. This has allowed the team to simultaneously minimize cost and improve performance. Furthermore, the team can now see consumption utilization over a period of time and reevaluate resource allocations on an ongoing basis. Reserved instances have played an important role in this shift.

Koozer and his team have implemented reserved instances with the help of Turbonomic. These have allowed the team, on average, to save 33% or more on their server run rates. In some cases, they have been able to decommission servers. The team has also reduced the amount of time they spend reacting to user complaints and resolving performance issues. In fact, they have brought their user complaints down to about 10–12 per year. Additionally, they have reduced their overall cloud spend. Their cumulative savings achieved, at the time of this publication, exceed USD 650,000.

For over a century, Komatsu has been a manufacturing leader committed to quality and reliability. Technological innovation and employee development have been major pillars of this mission. In recent years, Komatsu’s IT organization has expanded its adoption of the public cloud, and automation has played an important role in that journey. It has helped Komatsu reduce waste from a financial and environmental perspective, and it has helped Komatsu free up its IT organization to focus on strategy and innovation.

Headquartered in Tokyo, Japan, and founded in 1921, Komatsu (link resides outside of ibm.com) provides essential equipment, technologies and services for the construction, mining, forest, energy and manufacturing industries. Komatsu serves clients across the globe and states that its purpose is to “create value through manufacturing and technology innovation to empower a sustainable future where people, businesses, and our planet thrive together.”