When the developers at Lopez Food needed more compute power to compile software in their agile development environment, the infrastructure team chose to use the computing power available in their Intel NUC plant floor workstations rather than purchase compute cycles in Microsoft Azure or Amazon Web services.

Lopez Foods is a prepared meat processor based in Oklahoma City, with plants in Oklahoma, Iowa, and North Carolina. They process millions of pounds of meat each year for customers such as McDonald’s and Burger King. A highly automated operation, Lopez Foods uses computers to track intricate details about their production process. They track everything from the raw materials used in preparing the products, the equipment used in production, the workers on the line, the maintenance of the plant, and where each package is routed through the plant and each person and machine that is involved in preparing the product. This requires computing power throughout the plant, often houses in sealed enclosures to protect the meat from contamination.

Intel NUC workstations

As the existing PC’s on the plant floor needed replacement, the team looked at a variety of options, including Android devices. Eventually, they settled on the Intel NUC stations running an x86 processor with Linux.

During their research and testing for these new plant floor workstations, the IT staff needed to take into account the environment and usage requirements at each station. These workstations were typically single-task devices that were dedicated to a particular process. Users on the plant floor only needed access to specific web-based applications running on internal servers. In addition, many of the workstations were located in harsh environmental conditions that required installation in an air-tight enclosure. The devices needed to run without cooling fans.

The NUC made a perfect station for their needs.

Lopez Foods IT staff decided to run each NUC on the Linux operating system, thereby saving Windows licensing costs. The devices boot to the network and then pull an update configuration file from an internal GIT repository. This config file will delineate the operating system and software packages to be installed, and the functionality of each device. If a workstation needs to be reconfigured for another purpose, the IT staff makes a change to the config file, sends a command to reboot the NUC, and it downloads the latest config file, reboots again and then is ready for its new role. This made maintaining and updating the NUC stations much easier, and provided the functionality needed for the plant production staff.

Kubernetes Clusters

As the Lopez Foods programming team increased their software development cycles, they needed additional processing power for software builds. The infrastructure staff realized the plant floor NUC workstations were available and had extra CPU compute cycles that could be utilized for these software compiles without affecting the plant floor workers. Already an extensive DOCKER and VMware user, the team decided to configure the NUC’s to run Kubernetes in the background. To orchestrate their Kubernetes clusters, they chose Rancher as their orchestration tool.  “Rancher makes it easy for us to utilize Kubernetes” said Matt Flaming, IT Infrastructure Manager for Lopez Foods. “As workstations are brought online, they join the Kubernetes cluster and then Rancher controls which node is used for the build workloads. The users don’t even notice the background CPU usage, and our developers have faster build times.”

End Result – Faster, agile development without additional cloud usage fees

The end result is that Lopez Food developers can now commit their code into their GIT repository and the build is done automatically and seamlessly in the background, using idle CPU cycles that have already been paid for and that would otherwise go unused. It allows Lopez Foods developers to shorten their software build times while avoiding expensive cloud usage fees, all without impacting the productivity of plant floor workers. Money is saved and developer productivity is improved.