Replacing a failed SBC with a virtual one

An SBC rarely gets attention. It joins two phone systems, smooths over what they disagree about, and nobody notices it — until it fails. Then everybody notices it at once.

That is what happened at a customer site. A beroNet SBC sat as a hardware appliance between a Unify OpenScape Business and a 3CX system. When it died, the two systems could no longer reach each other. Sourcing a replacement unit, configuring it and installing it would have taken days.

What we did instead

The customer runs a VMware cluster. The software on the appliance is the same software that also builds for x86 — it had simply never been packaged for it. So we built it for x86 that afternoon, installed it on an Ubuntu VM in the existing cluster, and started it as a system service. The VM took over the IP address of the failed appliance. As far as both phone systems were concerned, nothing had changed.

The decisive find was on the old device

What made the swap fast was not the software. It was something far more mundane: the configuration backup of the failed appliance. It let us carry the dial plan across byte for byte, instead of reconstructing it from packet captures and guessing at it piece by piece.

That is the real lesson here, and it holds regardless of the hardware: a current configuration backup decides whether an outage costs an afternoon or a week. If you do not have one, you repay it in hours when it matters.

What it does — and what is still open

The licence carries 32 transparent sessions and is tied to the machine identity of the VM. The system has been in production since 6 August; it was never meant as a stopgap.

Being straight about it means saying what this case does not show. The media path under load has not been measured, hypervisors other than VMware have not been tried, and we built the system ourselves — not the customer, and not an integrator working from a manual. That is exactly what we are working on now: a packaged version that installs without us.

If you are in the same position

A failed SBC is no reason to wait for replacement hardware when a virtualisation platform is already there. Get in touch with what you have and what it needs to talk to — and have the configuration backup of your current device ready.