// DEPLOYMENT MODELS

Run it where your network lives.

The same images and control plane deploy on bare metal, in your private cloud, at the far edge, or across a hybrid of all three.

MODEL 01

Bare metal

Maximum performance and lowest latency. Data-plane instances run directly on certified servers and whitebox switches with DPDK or ASIC offload.

Line rateDPDK / SR-IOVWhitebox
MODEL 02

Private cloud

Deploy as containers on your Kubernetes. The control plane runs as a stateful set; data planes scale as a node pool alongside the rest of your platform.

KubernetesHelm / OperatorAutoscale
MODEL 03

Distributed edge

Push user planes to central offices and aggregation sites while a single control plane governs them all. Terminate sessions close to the subscriber.

Central officeAnycastLow latency
MODEL 04

Hybrid

Mix and match. Keep the control plane in a regional cloud, run bare-metal data planes for peak throughput, and burst into cloud capacity for the evening ramp.

BurstMulti-siteOne control plane
ONE ARTIFACT, EVERY TARGET

The same build runs everywhere.

No forked images, no "cloud edition." You validate once and promote the identical artifact from lab to edge to core.

Plan your rollout →
© 2026 Elastic BNG, Inc. — ebng.net
Privacy Terms Security Status