Folks - here is my situation:

We are building a network of distributed seismographs (number ~ 100) that each 
have wifi.  The instruments are deployed in a 1 km radius around a central base 
station.  We know that we can get good connectivity out to a few hundred 
meters, but not to the edge of the array.  

The data rates are of two kinds: 
Parameters: sending out recording parameters (sampling rate changes, etc) that 
is a few tens of bytes, and "rare" - a few times a day.  But must be propagated 
through the network reliably (with acknowledgement).

Seismic data: The 2nd type of data is receiving seismic data back from the 
seismograph to the central site- this is higher volume, but does not need to be 
reliable (all data are on the seismograph for later pickup).  ~100kB/s/device.

Right now we use a set of access points distributed around the array that have 
a 2.4GHz radio for talking to the seismographs and 5GHz radio as a backhaul.

I would like to explore using a pseudo-mesh-network approach that uses the 
Zyre/ZRE protocol.  

In particular, I want to know if some ad-hoc way to RELAY messages through the 
devices that are (directly) visible to the base station to other 
(invisible-to-the-base) devices.  

I am at the point where my fake array (modeled on ztester in Zyre) sends data 
to the base (and receives parameter updates).  In order to propagate out from 
the "visible" array I am at a crossroads:

- client nodes V1, V2 enter and join and form a set that includes the base 
station B (B, V1, V2)
- client nodes I1, I2, enter and join and form a set that include (V1, I1, I2)
 
at this point V1 is connected to (V1, B, V2, I1, I2).
I use brute force to distribute parameters: B sends out parameter updates to 
everybody it sees (V1, V2).  V1 sends THAT parameter update to everybody it 
sees (except those it has gotten THAT from - V2, I1, I2).  
As a consequence, V2 gets updates from B and V1, so there is some waste, but 
OK, this is rare.

What I am mulling is how to send seismic data from I1 back to B.   (and then, 
because fleas have fleas, from II1 to I1 to V1 to B, ad infinitum).

Thoughts?

This is a fairly common problem out there that lots of folks are grappling 
with, and I suppose zigbee or some such would be appropriate, but we don't have 
those in the seismographs.

Sincerely,
Sridhar


_______________________________________________
zeromq-dev mailing list
[email protected]
http://lists.zeromq.org/mailman/listinfo/zeromq-dev

Reply via email to