Multiple tedge installation on one device to connect more then one Cloud Providers

Forum
, ,
1

Moin, moin,

I’m new to tedge and cumulocity and I need help with an unusual problem. I want to send data from one device to two destination addresses (cloud providers). One solution would be to send the data to, for example, a Kafka instance and implement two Kafka topic subscribers to distribute the data. However, I have neither Kafka nor the ability to implement the subscribers at the cloud providers. So my question is: Can I install two instances of tedge on a single device and connect them to different cloud providers?
Any help or advice is very welcome.

Mischa

2

Moin!

I’m pretty sure you don’t need two thin-edge instances for your use case. It depends of the target system/cloud you want to send data to of course.

As thin-edge leverages MQTT as internal message bus, you can have multiple mappers that map data to cloud provider 1 and cloud provider 2 at the same time with just a single instance of thin-edge (but extended with custom mappers).

The dependency to your target system is important: If they use a MQTT Broker you can easily bridge it and your data mappers publishing to your local + bridged brokers. If no MQTT Broker is provided by your cloud providers but any kind of API/other Message Bus you need to implement that integration.

Tedge-flows are pretty handy for your use case: Extensible mapper and user-provided Flows | Thin-edge

You can find some examples here: tedge-flows-examples/flows at main · thin-edge/tedge-flows-examples · GitHub

They make it easy and straighforward implementing your mappings / integration.

1 Like
3

And just to extend what @Stefan_Witschel said.

The upcoming thin-edge.io 1.8.x release (due in the next few weeks) will also allow users to connect to other cloud MQTT brokers, this would enable such things like HiveMQ, EMQX, Event DataGrid (with MQTT interface enabled), etc.

Then using custom mappers with flows means you’ll be able to write a fully fledged mapper to get your data to the cloud (and push from the cloud).

4

Thanks for that.
I’ll have a look into the custom mapper capabilities.
It would be great if that is sufficient to solve my problem.

Rerads

Mischa

1 Like
5

Thank you for the hint to that upcoming feature. May be it makes my life easier.

Regards

Mischa