On Mon, 12 Jul 2004 23:33:10 -0400, jon@dakota-truck.net wrote:
>possible. Using this sort of method, it would be possible for
>telemetry from trucks in absolutely impossible locations to reach
>DML HQ because it can essentially jump from truck to truck until
>it finds a line of sight connection back home. :-)
That was a bit more complex than I was thinking, but it would work
with enough technology. I'm thinking of a cheap (D-Link or Linksys)
box in every truck. Short-range could even work with just the
built-in antennas. (Otherwise, magnet-mount omnidirectional antennas
would work great) All of the newer boxes include "repeater" mode,
which will just pass along all of the traffic they receive from other
connected nodes. In this way, you just build one giant bridged
network. In a "real" application that wouldn't be as practical for
performance and security reasons, but for this applicatoin I think it
would be just right. Every client could see every other client,
regardless of how many "hops" were necessary to get there. Since
we're not really talking about that many nodes, it wouldn't be too big
of a problem.
This is a project that would be a lot of fun. At least in my opinion.
:-)
> Another thing to think about would be what sort of TTL this
>data might have. For example, say a truck goes around a corner
>and loses communication with everybody. It is still gathering
>telemetry, so does it store this data and attempt to send it out
>upon reconnection with the network, or does it throw it away? If
>it stores it, how long should it do so before the info is
>considered to be so stale that it is worthless?
I think the thing to do would be to have each client just cache the
data until it were reconnected to the network. If the clocks on all
of the equipment were synchronized, every piece of telemetry data
could be timestamped and then assembled in order, plotted on a map,
etc. even if it arrived back at HQ later on.
-Bill
This archive was generated by hypermail 2b29 : Sun Aug 01 2004 - 00:46:14 EDT