ScaleOps on AWS: Kubernetes automation that builds competitive advantage

Learn how ScaleOps with Amazon Web Services (AWS) helped Adobe accelerate AI development and eliminate waste while reducing time spent on resource management and ensuring consistent performance.

Overview

Adobe runs over a million CPUs and thousands of GPUs across
hundreds of Amazon Elastic Kubernetes Service (Amazon EKS) clusters. As the company scaled traditional and AI workloads across many business units, idle CPU and GPU waste became the real cost problem: developer teams added buffer to each request for CPU and memory, and allocated entire GPUs to guarantee performance. Existing fractional-GPU options couldn’t flex enough to close that and in-house CPU and memory right-sizing approaches weren’t dynamic enough to handle production use cases.