Over 2000 Nodes Monitoring Via Wi-Fi Mesh Network.

pasan · 2016-05-15 11:11:16

Hi, There....

This is for the Rex Professionals here.

I want to monitor ON/OFF status of over 2000 machines.

1) If I fix one RPI+REX per machine with static IPs, would it be possible to monitor them all using another RPI+REX via huge Wi-Fi mesh network ?

2) Does RPI+REX capable of handling that much of quantity ?

3) If this is possible which REX package should I buy. (I may need data logging in future)

4) Is there someone who undertake custom projects like this ?

Have a pleasant day...

scoobsalamander · 2016-05-15 15:01:39

I'm not a REX pro but I was wondering how you want to check those nodes? By pinging and see if they react?
If so, you could use a Python (or other language) script which pings all your nodes one by one and stores the result in a file. This file can be read by REXcontrols.
(Or you store the results in a MySQL dB and then you already have your logging)

BTW : I think pinging 2000 nodes will take some time...

It all depends on what you want to do with it....is it that you want REXcontrol for a GUI or do you need some control logic afterwards?

Last edited by scoobsalamander (2016-05-15 15:03:05)

pasan · 2016-05-15 16:04:31

Hey thank you very much for the reply...

My first approach was using few master microcontrollers and link all the nodes via High Power RF Modems and as you said let the microcontroller masters ping all of them one by one. The reason to use few microcontrollers is reduce the polling time.
My end product is pretty simple, I just want to know the current status of the factory floor. Which means need to show all the data in a huge LCD TV.

This is not time critical so even if the polling take couple of minutes thats kind of okay...

I think this is totally doable using REX but I need some advice from who have good experience with REX.

scoobsalamander · 2016-05-15 18:26:43

Sorry, some more questions...

'High Power RF Modems'?? Are those 'nodes' not on the same (physical) network?

How did you planned to get the info visualized? Just a list with all nodes which are not responding or graphical pages with a network overview with visual feedback of the status of every node (by changing the color for example).

Do you need any kind of alarming if one or more nodes are offline?

pasan · 2016-05-15 18:39:41

Hi,

When I use Radio Modules I can define my own simple protocols. I simply can give give them IDs 0001 to 2000. Then I could POLL them one by one. So since I need only few bits of data I could POLL them pretty fast. But still it will take at least 3 to 5 minutes i guess.

Visualizing part is just basic let say I display 0001 to 2000 numbers on a 55" LCD screen. And if a machine goes down number will be RED. If its alive number would be GREEN.

Thing is its implementation part is pretty big but the end outcome seems to be ultra simple.

jaroslav_sobota · 2016-05-16 15:47:16

Hi Pasan,
supposing all the nodes are in the same subnet, I would use one RPi as a "server". It would listen to UDP datagrams from individual nodes. You can do that using the REXLANG block. Each node would send its ID in the UDP datagram e.g. every 20s (again using the REXLANG block). A node from which no datagram arrives for 60s is considered OFF. That way you have information about all the machines in one device (the "server") and you can e.g. display the IDs of the faulty machines.

A starter licence is sufficient: https://www.rexcontrols.com/product/rex … spberry-pi

Does that sound like a solution for your project?

REXYGEN Community Forum

Announcement

REXYGEN forum has been moved. This one is closed.
Head over to the new REXYGEN Community Forum at https://forum.rexygen.com.

Looking forward to meeting you there!

#1 2016-05-15 11:11:16