Skip to main content

How to implement fallback and rollback?

SGP.32 has two safety nets that help a device get back online if a profile stops working. They work differently, and using both is the safest setup.

Written by Jakub Kubinski

We strongly recommend that you set a fallback profile and use rollback for any planned profile switch, especially across a whole fleet. These two mechanisms can ensure that your devices stay connected and maintain smooth operation. Fallback and rollback can guarantee that a device that cannot connect to the new network can then recovers on its own, without a human intervention needed.

Fallback

You can mark one profile on an eUICC as the fallback profile. If the device detects a permanent loss of connectivity, it switches to that profile.

  • Only one profile can be the fallback at a time.

  • The profile must be allowed to act as a fallback.

  • How the device decides connectivity is "permanently lost" depends on the device implementation. If you use an IPAd, you control that logic in your firmware.

Rollback

Rollback is tied to a single enable operation. If the newly enabled profile can't attach to the network, the chip automatically goes back to the previous profile.

Fallback

Rollback

Triggered by

The device detecting permanent loss of connectivity

A new profile failing to attach after an enable

Switches to

The profile marked as fallback

The previously active profile

Set up

In advance, per eUICC

As part of the enable operation


What if a device has no connectivity at all?

It can still fall back or roll back on its own. New orders from the eIM can't reach it, though, so they wait until the device next checks in.

Did this answer your question?