Consider the following scenario: A WinXP client is connected to a LAN and has a DHCP assigned, private ip: 192.168.1.10. He "dials" a VPN connection to another corporate that has it's private subnet as 192.168.1.0 . So the WinXP client won't be able to reach the remote subnet because his routing table says that 192.168.1.0 is his LAN! What can be done in this situation?
Change the network numbers, or add network masks.
eg: if the local network has addresses in the range 192.168.1.1-100 and the remote has 192.168.1.130-250, then you can set the routes to something like
network 192.168.1.0 netmask 255.255.255.128 -> LAN (eg: eth0 for Linux) network 192.168.1.128 netmask 255.255.255.128 -> Remote (eg: ppp0)
I think the *normal* way in this situations would be that the RAS client to be able to access the remote subnet and not the one he is into, as long as the VPN connection is active.
That's only a side effect of having the routing table clobbered by the VPN.
Is this one of the reasons 3rd party VPN Clients are available?
No - what's happening for you is purely a side effect of having private networks on the same private range. Pick a different range (eg: 192.168.2.0) for one of the networks.
Alex