You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Can two Cores be configured for active-standby application? #1857
If the base station only supports configuring a single MME instead of multiple MMEs, can two Cores be deployed in active-standby mode? When one fails, can the other take over seamlessly to ensure uninterrupted service?
Ella Core can be deployed in a High Availability cluster. Every node is active and can receive radio connections. It uses the Raft consensus algorithm to replicate data accross all members. We recommend a cluster of 3 or 5 nodes in production.
If your radio can only connect to a single core (AMF / MME), you may need to put a load balancer between the radio and the cores to direct traffic to the different cores.
For more information read the following resources:
If we use a 3-node cluster, do all three nodes share a single IP address? Or does the base station need to support configuring three separate MME addresses?
Each node in an HA cluster has its own S1 IP address and a separate MME Code. All nodes share the same MME Group ID, so together they form a single MME pool.
There are two ways to connect a radio to the cluster:
Multiple MMEs on the base station (recommended): Configure the eNodeB with the S1 address of every node. Most radios support this.
SCTP/S1AP-aware load balancer: The radio connects to a single load balancer address, which forwards signaling to the cluster nodes. UE context is not shared between nodes, so the load balancer must keep each UE on the node that serves it, for example by tracking UE S1AP IDs and S-TMSI. We have not validated this setup and I'm not aware of an open-source option.
Activity
gruyaume commentedon Oct 24, 2026
Hi @ggsnpdsn ,
Ella Core can be deployed in a High Availability cluster. Every node is active and can receive radio connections. It uses the Raft consensus algorithm to replicate data accross all members. We recommend a cluster of 3 or 5 nodes in production.
If your radio can only connect to a single core (AMF / MME), you may need to put a load balancer between the radio and the cores to direct traffic to the different cores.
For more information read the following resources:
ggsnpdsn commentedon Oct 25, 2026
If we use a 3-node cluster, do all three nodes share a single IP address? Or does the base station need to support configuring three separate MME addresses?
gruyaume commentedon Oct 25, 2026
Each node in an HA cluster has its own S1 IP address and a separate MME Code. All nodes share the same MME Group ID, so together they form a single MME pool.
There are two ways to connect a radio to the cluster: