Hi,
Do you think that RIOT is production-ready? What is still missing according to you to run RIOT on an IOT device in production?
Cheers,
Hi,
apart from a few known bugs (and required API changes in the lower to mid-high levels) in the network stack we still have, I would give the (optimistic) answer that RIOT is production-ready. However, this evaluation only refers to a 6LoWPAN-CoAP + SAUL stack, as the full stack demos we run a few weeks ago were quite successful to my knowledge. Other popular features like e.g. MQTT(-SN) are still in development or not even touched upon currently. This is however my personal opinion and maybe does not reflect the opinion of the majority of the community. Also, some user-level APIs for CoAP (in the spirit of "I want to run a CoAP server on this port which publishes these actuators/sensors" in less then 10 lines) might be something that a client wants to see.
Cheers, Martine
Let's say that the production-ready RIOT could be announced for the next release (when network bugs will be solved) for 6LoWPAN-CoAP + SAUL stack.
Thanks for your answer Martine.
Hi Baptiste,
I think you need to specify for what type of use/product. Else it’s difficult to judge… I mean: if RIOT is supposed to “make coffee and donuts out of the box”, it’s maybe not production ready. For some other types of use, it may be production-ready.
Cheers,
Emmanuel
Hello,
when speaking about production ready… just wondering if somebody is running some long term tests? like at least 48 hrs of hard exercises.
wbr malo
Actually I’d be interested in that answer as well.
I am trying to run a mesh (using RPL) with a standard (dual serial) samr21-xpro border router. Everything works perfectly for about 15 minutes, then it dies. I don’t think I am running an unusual setup but so far I have already found an endianness bug, an xtimer bug and a ND router advertisement bug. All of them caused critical failures in the mesh.
I’m not complaining, I am aware of even worse bugs in TinyOS and Contiki that have not been fixed for ages, so RIOT is doing very well, but it does make me wonder. Are there people running meshes sending/receiving global traffic for long periods of time? If yes, what are you doing differently to me?
Hi Michael!
I am trying to run a mesh (using RPL) with a standard (dual serial) samr21-xpro border router. Everything works perfectly for about 15 minutes, then it dies. I don't think I am running an unusual setup but so far I have already found an endianness bug, an xtimer bug and a ND router advertisement bug. All of them caused critical failures in the mesh.
That's weird, we had recently demos at conferences and Embedded World with a very similar setup running for hours or even the full day. And I ran a RPL network for one week last November on the IoT-LAB without any problems (but without a 6LBR in this setup).
I'm not complaining, I am aware of even worse bugs in TinyOS and Contiki that have not been fixed for ages, so RIOT is doing very well, but it does make me wonder. Are there people running meshes sending/receiving global traffic for long periods of time? If yes, what are you doing differently to me?
Very good question indeed...
Cheers, Oleg
Hi all!
I’d like to answer the question about RIOT’s production ready status, but the only thing that I can say is that we are working hard to make this happen. As you may know RIOT is very young, with only 3 years old and mostly supported by “volunteers” but I’m too, as Michael said, convinced that RIOT is doing things very well.
This BR history is actually very interesting, as this configuration IS, indeed, very common, and should work as soon as possible. I think if we work together we can fix these bugs and have the mesh of our dreams. With this, we are one step closer to RIOT’s production ready status!
Cheers!
Francisco.