Private 4G/5G HA (High-Availability)

High Availability in on-prem private networks

Traditional HA for MNO environments are considered as 24/7, `mission critical` , such that any failures must be handled automatically. These solutions are very expensive and require specialist consulting in order to implement and support. Due to the scale, cost and complexity it is unsuited to the typical sXGP uses.

What is sXGP/private mobile network mission critical?

As long as the private mobile environment is needed to provide a service, whether this is one hour a week or 24/7, then HA is a fundamental requirement, it handles unexpected outages/failures. For example, JIT have deployed our HA solution in our live private 4G environment for sXGP in Shibuya and our HA environment has enabled our private 4G network to continue to provide service in the following real-world failure scenarios;

Power outage caused by building maintenance have seen successful auto restart when power is recovered
Recent earthquakes dislodged LAN cables from an active server resulting in a successful failover to a hot standby server
Application core dump (due to USB read failure) on active server resulting in successful failover to a hot standby server
Multiple single points of failure scenario
Planned upgrades (HW & S/W) during business hours by utilizing the HA environment and maintaining service at the same time.

There are other traditional HA options available such as using freeware (Corosync & pacemaker & keepalived), but an IT consulting team is required to implement. Other options could be third party licensed solution such as HP Service Guard or Red Hat Cluster Manager or other licensed s/w which will again need consulting and none of these case easily scale down to suit private mobile networks.

What if there is no HA for the private network?

Service interruption for multiple scenarios that will impact support costs and reputation of Integrator/Solution Provider as well as day to day business and productivity of the private mobile operator. How to handle these planned & unplanned failures and determine the impact of any to business. For example:

Power failures
Application failures
Network failures
Server Hardware upgrades
OS upgrades
Application upgrades
Other planned / un-planned maintenance activities

Next steps

At JIT, we have created a fit for purpose HA application manager to provide high availability for private 4G/5G server application software. When deploying this solution it is important to consider to degree of availability required and consider how to handle Disaster recovery. For example, there can be a loss of service due to any one of a variety of causes. The private operator needs to consider the level of redundancy required. For example

As a minimum - 2 servers with separate power supplies for each server in the HA environment
As a next level – each server in a separate room in different locations in a building -> and add battery backup
At a top level – servers in separate buildings and/or geographic locations