X.25?

From: Frank McConnell <fmc_at_reanimators.org>
Date: Sun Jan 10 12:48:36 1999

"Max Eskin" <kurtkilgor_at_bigfoot.com> wrote:
> Wired (no, I don't usually read it) reports that hackers were intending to
> disable Iraqui computer networks, which were supposedly accessible by
> dialup, and using a 'vintage' protocol called X.25. Could someone tell me
> what sort of protocol this is and what machines it's likely to involve?

There used to be (probably still are) X.25 public data networks.
"Public" means "not private" -- effectively, the network owner sells
connections and bandwidth to paying external customers, not just to
users within the owner's company. You as a customer would pay your
money, and the network owner would deal with details like telco so
that you effectively get RS-232 synchronous serial port(s) that want
to see and hand out X.25 packets at your various locations, and X.25
network addresses corresponding to those ports. Or maybe you would
pay for the telco stuff and an access charge to the network provider,
and get a port on the network provider's X.25 packet switch.

This might be of use to you if you had a need for computer/computer or
computer/terminal communication between distant locations. Host
computers (minicomputers and larger) often had X.25 network interfaces
available as an option, and software that supported
computer-to-computer communications and/or terminal-to-computer
communications over it. Terminals didn't, but you could get devices
called PADs (Port Access Devices? I can't remember) that had
connections for terminals, an X.25 connection for the network, and
could either act something like a multiplexer, running all the
terminal sessions over a permanent virtual circuit (PVC) to a host's
address, or like a terminal server, using switched virtual circuits
(SVCs) to connect to any host address on the network.

Or you might just have a host and a casual need for connectivity from
terminals in remote locations, e.g. you've got some database and you
sell access to it. So rather than installing a modem rack, you get a
connection to Telenet and tell your customers that they need to dial
in to their local Telenet access number (which is really a
Telenet-owned modem pool with something-like-PADs behind it) with a
terminal and modem and connect to such-and-such address.

Either way, the X.25 public data network solution could be cheaper
than installing point-to-point telco data lines between your hosts and
terminals, and the X.25 network provider could offer internal
redundant routing that might be prohibitively expensive for you to
implement on your own.

Now, that said, X.25 networks do not need to be public. If you have
the wherewithal, you can build your very own X.25 network out of the
same sorts of hardware that the public data network providers use.
I guess the Iraqi government could have one or several.

And I remember a data center that, in the late 1980s, got fed up with
its computer-to-computer communications being done entirely of
point-to-point links between computers (with a protocol that wasn't
smart enough to forward messages -- if you wanted to go from host A to
host C, and the only connection was through B, you had to log on to B
from A then C from B) and bought a small Dynapac X.25 switch to sit
between the computers.

X.25 can also be used to transport other protocols' datagrams,
e.g. IP. In fact the US Defense Data Network was IP-over-X.25 and may
still be for all I know. I think this sort of thing is either
incompatible with or distinct from using X.25 to carry host/terminal
traffic (in much the same way that e.g. rlogin and telnet are distinct
-- they both run over TCP but use different protocols, and different
ports so TCP can multiplex them), so you can't just dial into a PAD
and telnet to a host doing IP over X.25, unless the PAD also does telnet
and IP over X.25.

-Frank McConnell
Received on Sun Jan 10 1999 - 12:48:36 GMT