The problem: Leatherman Tool Group Inc., a manufacturer of multifunction tools and knives based in Portland, Ore., updated its 20-year-old, DOS-based ERP system with Microsoft Dynamics AX, increasing the complexity of the system from one physical server to nine virtual machines. The upgrade to Microsoft Dynamics AX tied systems together that were previously siloed, enabling the company to better track products across the warehouse — from assembly to packaging to shipping. But it also created a new problem: slow data center performance, according to Dameon Kirchherfer, database and systems administrator at Leatherman.
The strategy: Because the new ERP system is dealing with an uptick in data volume, data movement and data complexity, performance issues were not unexpected, Kirchherfer said. In advance of the implementation, Leatherman purchased new hardware and back-end storage to mitigate those issues, but the data center performance issues persisted anyway. And the added complexity of the system made it difficult to pin down where the bottlenecks were occurring. The network and CPUs, for example, appeared to be functioning smoothly. “So we started our search into how we could speed up our data center,” he said.
The results: Leatherman turned to PernixData. Initially, the software helped reduce data latency by caching “hot traffic,” or traffic that needed to be moved the most frequently, closer to processors and RAM, according to Kirchherfer. And, it turned out, data latency was a symptom of another problem: The CPUs, which had been running at 50% utilization before the PernixData installation, couldn’t handle the speed at which the system needed to run. “So we upgraded our CPUs,” Kirchherfer said, providing Leatherman with improved performance speeds. Today, Leatherman uses PernixData’s infrastructure analytics product Architect to keep tabs on data center performance.