This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Custom IEEE 802.15.4 protocol, please review

Other Parts Discussed in Thread: SIMPLICITI, TIMAC, CC2591, CC2530

Hi, I am trying to design custom communication protocol, suited for my application and wish to discuss some details. Consider following project. There are multiple (1000 per 100m2) low power device that should every 1 minute ask some center device for data and go to sleep. Device could be moved in much larger area, assume 1km2. I would expect single zigbee network (coordinator with some router) could not handle such heavy traffic (aproximatelly 16 devices per second, plus overhead communication between routers and collisions). Also some simplier stacks like SimpliciTI will not handle many routers and end devices. My idea of protocol is close to GSM network
- there will be some number (like 10 or 16) of isolated "center" points, which will connected to server via ethernet.
- each center point use different set of channel and PANid
- each device contain list of all possible sets of channels and PANid which are used in network
- all center points send beacon, first assumption every 1-2 seconds
- when device does not detect beacons from recently used center points trys to search best center point with list and beacon rssi, and "switch" to use this center points
- device does not associate (association request/response) each time with new network, just wait for beacon and try to communicate (every minute)

My doubts are about device power requirements, does devices during searching consume much power? My expectation are device could work few days on 100mAh battery.

Or maybe for my project better solution is divide all 1000 devices for 10 (or more) independent zigbee network, and setup 10 sets of routers to cover whole area?

  • ele, 

    You have many requirements in your statements above and some of them are contradicting. So me me try to have boil the requirements down to some bullet points.

    • Power : 100mAh for 3-4 days of operation
    • Data  :  1 packet of < 64 bytes every 60 seconds
    • Number of nodes : 1000 or more
    • Distance between gateway and nodes : As much at 1000 meters.
    • What frequency of operation (what country will you deploy in?)
    Did I get that right? 
    With the low power requirements, you will be limited to some kind of a star network and then you need to have enough transmit power to get 1000 meters. So output power is important and therefore frequency of operation and country where this system will be deployed.
    Regards,
    /TA 
  • Hi, thank you for response

    Yes, you are right in your summary, only point I can comment is distance between gateway and nodes. I do not really need so large distance in my application, there are no contradiction to put gateway more dense, but this may generate more interference between gateway radios. I want to cover larger area like 100x100m2

    About frequency I am thinking about 2.4GHz.

  • Ele,

    I think you should consider looking at TIMAC, which is the lower layers of the Zigbee Stack. TIMAC can be used as a standalone product and it enables the generation of very large star networks. It has both beacon mode and non-beacon mode (wake up on timer event).

    The choose a CC2530+CC2591 combination as they will easily provide 100 meter of range.

    Regards,
    /TA 

  • And how TIMAC will solve my problems? IMHO it is just a wrapper for IEEE 802.15.4. I still need to define some kind of custom protocol, at least some set of "rules" how device will communicate with gateways, for example if there will be one gateway and "repeaters" or many gateways.

  • Ele,

    TIMAC does CSMA and basic star networking. If you have 1000 nodes per gateway I do not see a need for complicated network topology. First order payload calculations:

    • 60 bytes * 1000 nodes / 60 seconds = 1000 bytes per second * 8bits/bytes = 8kbps.
    this tells you that you have lots of spaces left over in the capabilities of a typical 802.15.4 radio.
    Regards,
    /TA 
  • I have feeling your calculation are wrong. Consider following effects:
    - some of 1000 devices sooner or later will try to communicate simulatously making collisions. On collision both devices will retry communication. I hope it will not generate snowball efect.
    - one communication between device and gateway consists at least of few messages:
    1. device: give me data
    2. gateway: ack
    3. gateway: data for device
    4. device: ack
    - if there will be some retransmitters, they will consume some bandwith for data retransmission/routing

  • Ele, 

    I recommend downloading Contiki and use Cooja (network simulator) to simulate your setup of 1000 nodes. Contiki also comes with 5-7 different MAC's and you can pretty easily try them all without having to build any large scale systems.

    http://www.contiki-os.org/start.html

    Regards,
    /TA