AI at the edge: simplifying infrastructure with Cisco and Canonical

Cisco and Canonical have published a Cisco Validated Design (CVD) guide for deploying AI inference workloads at the edge. The solution pairs Cisco's Unified Edge hardware (UCS XE9305 chassis with NVIDIA L4 GPUs) with Canonical's open source software stack including Ubuntu Pro, LXD, Canonical Kubernetes, MicroCloud, and Charmed Operators for AI/ML toolkits like Kubeflow and MLflow. Key capabilities include Zero-Touch Provisioning via Cisco Intersight, support for mixed VM and container workloads, automated CVE patching, and a 15-year security maintenance lifecycle — all aimed at reducing the operational complexity of managing thousands of geographically dispersed edge sites.