# Connecting one computer to two different logical networks

My company has two sets of robots (24 computers in all) with controllers run
under QNX4 that are connected via 100BaseT. We have two physically separate
networks that each facilitate communications between twelve different
computers. Due to network traffic considerations we want to keep the 24
computers split into two physically separate networks. However, we want to
resides on each network but maintains the physical and logical separation of
networks. Currently node numbers 1 through 12 are setup on each network.
I.E. A computer on each network is configured as node 1, a computer on each
network is configured as node 2, etc.

The current applications are dependent upon their current node number
assignments to coordinate their various activities. (Not the way I might
have set this up, but I inherited the applications) Because of this, I am
reticent to change the node number assignments to eliminate the duplication
of node numbers. I would like to be able to communicate with specific
computers based upon the logical network number combined with the node
number.

I understand how to setup the network card driver and the netmap to specify
the different logical network numbers. However, what is not clear to me is
how to get the new computer to communicate with for example, node 3 on
logical network 1 at one moment and then to communicate with node 3 on
logical network 2 at the next moment, etc.

Is what I want to do possible. If so, does anyone have any suggestions as to
how I can accomplish this?

David

“David Ellis” <dpellis@home.com> wrote in message
news:95mf2q\$f4a\$1@inn.qnx.com

My company has two sets of robots (24 computers in all) with controllers
run
under QNX4 that are connected via 100BaseT. We have two physically
separate
networks that each facilitate communications between twelve different
computers. Due to network traffic considerations we want to keep the 24
computers split into two physically separate networks. However, we want to
resides on each network but maintains the physical and logical separation
of
networks. Currently node numbers 1 through 12 are setup on each network.
I.E. A computer on each network is configured as node 1, a computer on
each
network is configured as node 2, etc.

The current applications are dependent upon their current node number
assignments to coordinate their various activities. (Not the way I might
have set this up, but I inherited the applications) Because of this, I am
reticent to change the node number assignments to eliminate the
duplication
of node numbers. I would like to be able to communicate with specific
computers based upon the logical network number combined with the node
number.

I understand how to setup the network card driver and the netmap to
specify
the different logical network numbers. However, what is not clear to me is
how to get the new computer to communicate with for example, node 3 on
logical network 1 at one moment and then to communicate with node 3 on
logical network 2 at the next moment, etc.

Is what I want to do possible. If so, does anyone have any suggestions as
to
how I can accomplish this?

To my knowledge this is impossible. All nodes in a QNX (FLEET) network
must be unique. It’s the node id that identifies machine and not the node
id/lan id
combinaison.

Withough changing node id, I see one wait out of this, but it’s not very
realistic
nor efficient.

Setup one group of node as lan 1 and the other as lan 2. Then setup up
a node 13 on lan 1 and a node 14 on lan 2. Then setup up a node 15 that
is connected to the two lans. Node 15 netmap should only be allowed
to see node 13 and node 14. That way node 1-12 lan 1 can write on node 13
HD and node 1-12 lan 2 can write on node 14 HD. At any time Node 15
can access data on both nodes 13 and 14.

This solution is not really good, cause Node 15 cannot directly access
data or do IPC on any of the node 1-12. It as to go through files
store on node 13 and node 14, or through a program running
on node 13 and 14.

None of the node 1-12 should know about node 15.

You might also consider using TCP/IP.

At any rate all these solutions are so ugly that the best one is probably
to “fix” the system to not use fixed node id.

better at reducing traffic then splitting the system in two lans.

David

David Ellis <dpellis@home.com> wrote:

< has two sets of nodes 1-12, wants a node “13” that can communicate
with both sets of nodes 1-12>

I would like to be able to communicate with specific
computers based upon the logical network number combined with the node
number.

This can’t be done using QNX native networking. Well… a really ugly
solution would be for node “13” to have two completely different netmaps
available… and to use one netmap to communicate to nodes 1-12(prime)
and a different netmap to communicate with nodes 1-12(secundus), but that
means both sets of 1-12 are going to see node 13 appear and vanish from
there network periodically. It could be made to work… but it would be
ugly. And, as noted, you couldn’t communicate with both lans at the same
time.

Another, probably better, solution would be to use TCP/IP.

## -David

QNX Training Services
dagibbs@qnx.com