Adtran introduced Ensemble Cloudlet multi-node, a high-availability edge platform aimed at AI inference and agentic AI workloads that require low latency, resilience, and local data processing. The system creates a single logical cloud with distributed storage and automated failover starting with as few as two nodes, targeting enterprises, service providers, and industrial operators that need edge infrastructure without the operational burden of traditional clustered deployments.
The platform runs on Adtran’s Cloudlet OS and uses Ensemble EdgeView for orchestration, zero-touch deployment, and lifecycle management. Adtran said the software installs directly on bare-metal servers, automatically builds a shared storage pool, configures inter-node networking, and brings workloads online with automated failover and live migration. The company positions the offering as a common edge foundation for networking, security, IT, and emerging AI applications, including GPU-accelerated inference services.
The announcement reflects a broader push to make edge infrastructure more cloud-like and easier to manage at small footprints. By starting at two nodes and automating cluster formation, storage, failover, and management, Adtran is targeting a common pain point in edge deployments: the need to support resilient services in distributed locations without highly specialized on-site teams. The company also emphasized workload mobility and flexible core allocation to balance performance and efficiency across mixed edge applications.
• High-availability edge cloud starts with as few as two nodes
• Supports AI inference, agentic AI, networking, security, and IT workloads
• Uses distributed storage and automated failover to protect services
• Runs directly on bare-metal servers via Cloudlet OS
• Managed through Ensemble EdgeView with zero-touch deployment
• Includes live migration, flexible core allocation, and GPU-based acceleration
• Targets service providers, enterprises, and industrial edge environments
“Deploying high-availability edge infrastructure for AI has often demanded a lot of cloud wizardry – stitching together storage, networking and failover manually. With Ensemble Cloudlet multi-node, we’ve made the hard parts easy. The nodes build their own shared storage, form their own network paths and protect their own workloads the moment they come online. You plug them in, connect to the internet and watch the cluster assemble itself.” — Christoph Glingener, CTO of Adtran
