This post is older than 2 years and might not be relevant anymore
More Info: Consider searching for newer posts

Which protocol to use?

Hi everyone,

We are trying to choose which protocol use for connecting devices based on Nordic SoCs. We use devices in following way:

  1. about 300 - 500 devices (wearable - smart watches)
  2. industrial site, walls, concrete, metal etc.
  3. site area up to 1 km2
  4. each devices send data to gateway - 6 kbytes every 10-20 minutes and 190 kbytes 3 times a day
  5. protocol should support identification of sent data (from which device it arrived)

I know it is a lot to ask, but what would you suggest?

Regards Pavel

Parents
  • Ahoj Pavle,

    My two cents: Just based on your high-level requirements pretty much any technology should be in principle able to transport this low throughput from hundreds of devices. BLE can do it (when you will install set of "hubs" which will be connecting to sensors and gather data one by one), the same with proprietary technologies like ESB and Gazzel, ANT is doing this in a gym or in a peloton so why not here, fancy things like BT SIG mesh or some proprietary mesh or even networks based on IEEE 802.15.4, all should do the job after certain work on the topology and application layer. So far so good.

    However what you haven't set are other three main criteria which will break many candidates easily:

    • These hundreds of devices will be static or moving?
    • Should they build self-contained solution meaning just one or few border routers installed or you can afford network of static hubs covering comfortably whole site? I guess you understand that typical visibility on 2.4GHz low-power radio inside industrial indoors complex with lot of metal, water, thick walls, people and 2.4GHz "noise" (e.g. heavy Wifi/Zigbee/BT traffic) is 10-20m (and there will be corner cases where it can easily be less then 10m). So if you don't put any assumptions on devices and hubs you can easily end up in situation where none of these BLE-like technologies on nRF5x chips solve your problems.
    • What are power consumption criteria? Are there any devices powered permanently from the grid? Which devices are battery powered and how sensitive you are about batter exchange/recharging?

    Cheers Jan

Reply
  • Ahoj Pavle,

    My two cents: Just based on your high-level requirements pretty much any technology should be in principle able to transport this low throughput from hundreds of devices. BLE can do it (when you will install set of "hubs" which will be connecting to sensors and gather data one by one), the same with proprietary technologies like ESB and Gazzel, ANT is doing this in a gym or in a peloton so why not here, fancy things like BT SIG mesh or some proprietary mesh or even networks based on IEEE 802.15.4, all should do the job after certain work on the topology and application layer. So far so good.

    However what you haven't set are other three main criteria which will break many candidates easily:

    • These hundreds of devices will be static or moving?
    • Should they build self-contained solution meaning just one or few border routers installed or you can afford network of static hubs covering comfortably whole site? I guess you understand that typical visibility on 2.4GHz low-power radio inside industrial indoors complex with lot of metal, water, thick walls, people and 2.4GHz "noise" (e.g. heavy Wifi/Zigbee/BT traffic) is 10-20m (and there will be corner cases where it can easily be less then 10m). So if you don't put any assumptions on devices and hubs you can easily end up in situation where none of these BLE-like technologies on nRF5x chips solve your problems.
    • What are power consumption criteria? Are there any devices powered permanently from the grid? Which devices are battery powered and how sensitive you are about batter exchange/recharging?

    Cheers Jan

Children
Related