No Title

Tue, 30 Nov 1999 00:11:00 EST

Yes, it does need a central server that basically acts like a database server. The server is not much more difficult to implement than the protocol itself. You can experiment by going to

http://firstforge.wpi.edu/sf/projects/network_tables_2_0

and downloading the TableViewer. If you don't type in a hostname into that app it will switch to server mode. Its a good one to test any NT implementation.

I don't believe that FIRST has any restrictions on usage esp. on a research project but you should check around firstforge to see what licensing restrictions are present.

For a swarm based system, I'd probably look at something a bit more distributed .. but that gets a lot more complex. The theory is that one or more nodes could go down but the information state would be preserved not unlike a RAID system.

The NT system is built around the robot, if the robot goes down, then it really doesn't matter if the state is preserved ... cause you are out of the competition! Not the case in a swarm system.

Are there not distributed state machines that already exist for this kind of system?

STeven.

Network Tables

Tue, 30 Nov 1999 00:11:00 EST

I am intrigued by the Network Tables module. It may be applicable to my project.

I am creating a swarm that needs to share information amongst all the units. From what I can tell with the NT module it would do this but needs some form of server. If this is correct, how difficult would it be to implement a server, assuming FIRST does not have restriction to prohibit this?

RoboRealm Forum

No Title

Network Tables