X.25 Pad
Derek Vair
derekv at dvlmarv.UUCP
Sat Apr 22 00:46:44 AEST 1989
In article <151 at oldcolo.UUCP> dave at oldcolo.UUCP (Dave Hughes) writes:
>What is the price differential between doing X.25 by
>means of a X.25 external PAD, and by internal software?
>Just a ballpark please.
>
We have a product, Netcom II, that is an X.25 communications controller for
SCO Xenix and UNIX on the AT-bus. If you want specific information on it,
please contact me directly.
On the economics of PAD versus integrated X.25 controllers, the answers
depend on your operating environment - essentially, how many users you're
supporting, at what speeds, and how far you are from the nearest packet
network node.
In terms of capital investment, here's a brief analysis of costs:
Our product costs $1950.00. It is a coprocessor board that plugs
into the PC, and allows the machine to connect directly to an one or two
X.25 links. Each link can carry up to 32 simultaneous terminal
connections. If you've got 64 users, this looks very good - about
$30 per user. For one user, it's very bad - $1950.00 per user.
Reality is probably somewhere in the middle - say, 8-16 users per
link, or $60 - $120 per user.
Outboard PADs are available at a variety of prices, depending on the
number of asynchronous connections and link speeds they support.
You can probably get outboard PADs cheaper than Netcom II, but then
you've got to buy some kind of asynchronous communications
controller board to provide a path for each of the terminals
connected to remote PADs to connect to your host via the outboard
PAD sitting beside your host. For instance, if you buy an 8 port
PAD, you will require at least an 8 port async card and need to run
cables from the PAD to the card. While individually each costs less
than the integrated solution, the combined cost of asynchronous
controller, ourboard PAD, and cables will probably exceed the Netcom
II price.
For very small numbers of remote users (1-3), you can use an
asynchronous PAD service from the X.25 common carrier - they can run
an async line from their siwitch to your site for each terminal you
want to support. Operating cost will depend on geography and usage
- but all you'll have to buy is the async ports and cables. The
carrier should do the rest.
If you're talking to a public packet network, your link speed will
be limited to 19200 or 9600 bits per second, and pretty well all the
PADs on the marketplace support those rates. If you're using a
private network, and can make use of 38.4 or 56 Kbit/second X.25
links, you will have to select the PAD carefully, because a lot of
them are limited to 9.6 or 19.2 Kbits/second.
Cost of operation will be the same, unless you get your PAD from a
public network vendor. Some public networks (Datapac, here in
Canada, for example) charge a significant premium for PAD usage over
the normal traffic charges. If you've got a lot of traffic, network
usage is a significant cost factor.
Convenience may also be an issue:
With a PAD connecting your machine to the network, you'll have some
problems with terminal control. For example, character echoing.
Either terminal traffic from remote users will be echoed by your 386
(expensive in traffic terms, and also very poor perceived response
time, as each character transits the packet network, gets processed
by Xenix line discipline, and transits back to the terminal), or
your users set their PADs up to do echoing and turn echo off in your
Xenix application or shell (in which case, whenever they run an
application that puts the terminal into raw mode - vi, for example -
they have to manually change their PAD's echo control parameter to
stop echoing the characters received).
On the other hand an integrated product should control the remote
PAD's parameters automatically according to the termio modes of the
"tty" that is attached to the virtual circuit.
===============================================================================
Derek Vair "I don't need a disclaimer - I'm
The Software Group Limited a principal"
#201, 4701 Steeles Avenue West
Weston, Ontario
phone: 416-747-9490
fax: 416-747-1471
email: uunet!dptcdc!lsuc!dvlmarv!tsgfred
===============================================================================
More information about the Comp.unix.xenix
mailing list