How to DFU Nrf52840DK boards over a "serial" topology network?

[Chris Baudia]

Hi!

We'd like to understand the most reliable way for DFU OTA our Nrf52840DK boards over a "serial" topology network?

We are using nRF5 SDK 15.2.0 and our configuration is as follow:

Gateway <--> Node1 <--> Node2 <--> Node3

Nodes are connected in multi-link point to point as follow:

Gateway (peripheral) <---> [Node 1 (central) -- Node 1 (peripheral)] <---> [Node 2 (central) -- Node 2 (peripheral)]

Ideally, the gateway would download the new firmware and pass it over the nodes that will run a background DFU process. Downloading the new firmware from the gateway is under control but how to "automatically" update all the nodes in the network?

Thanks in advance!

-Chris

[Joakim Jakobsen]

Hi.

Apologize for the long delay in response.

Do you still need assistance on this topic?

Br,
Joakim

[Chris Baudia]

Hi Joakim,

Thanks for replying. Yes, we still need some assistance on this. Can you help?

Thanks!

-Chris

[Joakim Jakobsen]

I will do my best to help you solve this.

What do you mean by "automatically" update all the nodes in the network? Do you want do update all the nodes in one operation? Or do you need some other "automation" for the DFU?

BR,
Joakim

[Joakim Jakobsen]

See: https://devzone.nordicsemi.com/f/nordic-q-a/71592/multiple-peripheral-device-dfu-ota/294433#294433

[Chris Baudia]

Thanks Joakim, much appreciated.

We are designing a system that will pull information from our server in order to know whether there is a new firmware version ready to be installed. If this is the case, the gateway will download the new firmware and each node will run a background DFU. By "automatic", I meant without any physical intervention on the network (not even using a mobile app).

Our concern is that we need a reliable DFU procedure with rollback feature in case of something goes wrong. We were thinking starting to DFU the last node then notify the upstream node when done etc etc until we reach the gateway.

I was hoping that we could use an existing procedure or library for that.