* Having a web interface running, accessible via Ethernet or BLE (for
on-site maintenance) and remote (e.g. via LTE CAT M1/NB1 for
administration) to allow configuring sensor setup (number and type) as
well as general configuration (measurement interval, time of data
upload, etc.)
A simple graphical web interface would ease the maintenance for the
technician as he would "only" need to change configurations rather than
changing source code and compiling (see point above). Since I am
working currently with OpenMediaVault, I really appreciate their web
interface concept, with the possibility to revert all currently made
changes.
Okay, that point is a bit tricky. At least tricky, if you want to have
this web interface directly on the Microcontroller. I would use CoAP for
this use-case. It makes the most sense. But there is one issue: Normal
browser, like Mozilla, Chrome etc. don't support CoAP by default. There
is Copper(https://github.com/mkovatsc/Copper), but as you can see it is
discontinued.
You can use an old browser, but I guess your point was to use a modern
browser, right? There are other options as well, of course. LibCoAP is
well supported. So, you can use it in a programming language of your
choice. Every major language has some kind of CoAP lib. So, you could
provide an CoAP interface you can interact with.
Writing a CLI for it, shouldn't be a big deal. From the remote point of
view: Just host a central management frontend which interacts with the
devices. Collecting configs, collecting data etc.
You probably also have to use edge device/border router/gateway in your
setup. So, you could also put an instance of this web interface there as
well.
So that you have a web interface on-site as well. But having a CLI
backup is probably also a good idea in case the web interface crashes.
You can also have it more decentralized and let the edge device collect
the data in some way.
There is room for different architecturial decisions.
*All these configurations would need to be stored locally, I imagine as
JSON on an EEPROM with some kind of file system. Is this how it is done
in general?
Not sure, if JSONs is the best idea here, since you have to parse the
JSON on the MCU. Using a filesystem in RIOT is quite simple. So, you can
store it in files.
When you call the API, you change the value in memory and on the file
system. When the device wakes up it loads the configuration from the
filesystem.
I would also like to hear the input from other people, since I didn't
had dynamic configuration yet.
*All configurations should also be synced with the remote server (e.g.
running ODM2)
Just provide an API for this at your devices. So that you can change the
configuration remotly. Of course, this should be done in a secured way.
Maybe it makes sense to use Web of Things (https://www.w3.org/WoT/) in
your use-case. Especially considering you want to collect all these
data, change the configuration etc.
There is the WoT Scripting API which makes it quite simple to interact
with the device and write a web frontend for it, once the WoT TD is
present. Node-WoT(https://github.com/eclipse/thingweb.node-wot) is quite
an advanced implementation of it.
I wrote a PoC implementation for RIOT
(https://github.com/w3c/wot-testing/tree/master/events/2020.09.Online/prototypes/RIOT-OS).
If that sounds interesting to you, just ping me.
I could imagine that you write some generic plugin/extension for ODM2
and support WoT TD on it. So that you can provide an interoperability
layer. Eventually you could also replace your device with commercial
available ones. They just have to support the WoT TD.
best regards
Philipp
_______________________________________________
users mailing list
users@riot-os.org
https://lists.riot-os.org/mailman/listinfo/users
