RE: Newbee question|
Tue, 26 Mar 2002 15:57:37 +0100|
I can answer some of the questions.
> > Just pick two from the list. For example,
> > 192.168.0.1 and 192.168.0.2.
> Can u explain more how it works? (if u have time)
There are IP addresses reserved for people to use as they wish for
their own private networks, for example CIPE or LANs. If you read RFC 1918
(use a search engine to find copies of this) you can see the philosophy
behind these IP addresses.
> As far as I understand, the tunnel made from two point
> using ip address and the udp ports. When it goes
> inside it would expect the internel ip (192.x.x.x) to
> reach the correct box, but in my case it nerver find
> out where it is.
A packet goes through your interfaces more than one time. In your
case, a program sends a TCP packet to 192.168.0.2. Your routing table says
to use cipcb0, so the packet goes to cipcb0, but instead of an ethernet
circuit board, a program receives it. The program encrypts the packet and
then sends it to the real udp IP address and port. The routing table sends
this new packet to another interface, perhaps an ethernet board, and then it
leaves your system. The opposite happens when packets come back.
> one more thing, should I compile IP tunneling in the
> kernel or it is not needed.
I don't know.
Best regards, Scott.