Overview
The Celona Edge acts as the heart of a Celona private mobile network, providing mobile services across the entire array of devices and applications connected via private cellular. It interconnects the Celona access points (APs) across the private local and wide area networks for control and user data plane processing.
Celona Edge supports the clustering of multiple Edge Nodes to form a highly available and redundant Edge Cluster. When devices connect to the Celona network, their connectivity and traffic are load-balanced across the Edge nodes in the cluster, and the cluster continues operating through the failure of an Edge node.
High Availability options
Celona offers two High Availability architectures for multi-node Edge Clusters. The architecture is selected as the Cluster Scheme (V1 or V2) during Edge Cluster creation.
| HA 1.0 (Cluster Scheme V1) | HA 2.0 (Cluster Scheme V2) |
Availability | EdgeOS 2411 and earlier | EdgeOS 2026.1 |
Minimum nodes for HA | 3 (required for Kubernetes quorum) | 2 |
Operating model | Leader/standby control plane (effectively active-passive); workers active on all nodes | Active-active — all nodes active, devices load-balanced across them; no standby or follower nodes |
AP ↔ Edge control plane | Single Edge Cluster VRRP IP address (VRRP ID) shared by the cluster | Per-node IP addresses; no cluster VRRP IP required |
Layer 2 adjacency between nodes | Required (VRRP) | Not required |
Recovery on node failure | APs reconnect in ~4–5 seconds; device reconnection typically completes within 30–120 seconds | Under 15 seconds |
Radio access | 4G LTE and 5G | 5G |
HA 1.0 (Cluster Scheme V1)
HA 1.0 is based on a multi-node Kubernetes cluster with a single active leader, standby leaders, and a Virtual Router Redundancy Protocol (VRRP) IP address presented to Celona APs. A minimum of three Edge nodes is required.
HA 2.0 (Cluster Scheme V2)
HA 2.0, introduced with EdgeOS 2026.1, removes the 3-node dependency with a 2-node active-active architecture, delivering faster failover, simpler operations, and alignment with enterprise HA designs. All nodes are active, devices are load-balanced across them, and no cluster VRRP IP or Layer 2 adjacency is required.
Read more: High Availability 2.0 Overview
Which option should I use?
New Edge Clusters on EdgeOS 2026.1: select Cluster Scheme
V2(HA 2.0).Existing HA 1.0 clusters: a three-node HA 1.0 cluster that upgrades to EdgeOS 2026.1 can continue operating with its existing node count; no site reconfiguration or settings changes are required. See the Site/Edge Migration Guide for details.