jim@tellab2.UUCP (Jim Gast) (08/02/84)
On July 19, I asked net.unix-wizards and net.lan if anyone knew of an X.25 implementation for 4.2BSD on a VAX. ihnp4!gjm (G. J. Murakami) said: I have heard that the System V BX.25 implementation is a bit messy, but have you considered a port to 4.2 BSD? In my experience, porting efforts from BSD to USG haven't been too bad. [Ed. NOTE: BX.25 is not precisely X.25, but it is very close] teltone!ira said: Simpact Associates, Inc. (5520 Ruffin Road; San Diego, CA 92123; (619) 565-1865, ext 246) have a programmable communications processor on a UNIBUS board with X.25 available. DEC is writing a Ultrix (4.2) driver for them. [Ed. NOTE: It will be available soon, hardware cost approx $5,000 and software cost approx $5,000.] utcsrgv!dave (Dave Sherman) said: Buy a PAD. Life will be a lot easier, and you won't weigh down your VAX with all that X.25 work. (We have a Micom X.25 PAD which works fine.) [Ed. NOTE: Some of the X.25 boards actually claim to be **LESS** overhead than a DH! The DH is DMA out, but one interrupt per character incoming. The Cableshare (see below), for example, is DMA out and in, and the dedicated processor handles all of the layer 2 and layer 3 overhead.] randvax!guyton (Jim Guyton) said: Yes, the csnet folks are running X.25. They're running 4.2bsd and use the ACC IF-11/X.25 board modified with new firmware chips from someone other than ACC (I think, Interactive Corp of Santa Monica). hplabs!fowler (Greg Fowler) said: Interactive Systems in Santa Monica, CA, is selling a UNIBUS board (Incard) that allows connection to an X.25 system. This card is the underlying hardware used for CSnet's X.25net. This card isn't at all inexpensive though. I have also heard that a university in Canada has done a software implementation of X.25 but haven't been able to track down any details. [Ed. NOTE: INcard Board alone approx $12,000, 1 link, up to 56Kb. Add software cost to that.] obrien@csnet-sh.arpa (Mike O'Brien) said: CSNET is the Computer Science Research Network, a not-for-profit net that links up computer science researchers in both academia and industry. It's a logical net, using whatever transport mechanisms are handy. Currently, these include dial-up telephones (PhoneNet), Arpanet, and Telenet (X.25Net). To support the latter, we have a software package that runs under 4.2 BSD which connects the TCP/IP stuff in the OS to an X.25 net through something that looks like a very hairy device driver. It opens virtual circuits to other hosts (or gateways) and shoves IP packets at them. Works just great. The hardware currently supported is the Interactive Systems INcard, which is a reworked UMC-Z80. We have no notion how fast the INcard really is, since we only use it on 9600 baud DAFs to Telenet. The UMC-Z80 (which is really manufactured by ACC in Santa Barbara -- Interactive only redid the microcode) is a workhorse product, though, which has recently resurfaced as an Ethernet interface, so it can't be all that slow (barring microcode botches). That's about it for info on the INcard. Our X.25 software, sort of as a freebie, can also bypass the TCP/IP stuff and do incoming and outgoing PAD. If you send me your U. S. Snail address, I'll send you a package of info on CSNET. Your company might be interested, particularly since an X.25Net connection gives you complete Internet access. [Ed. NOTE: CSnet folks were very helpful, but it costs $$$ to join CSnet and they currently will not sell to non-members.] PEREIRA@SU-CSLI.ARPA (Fernando Pereira) said: I know of an X.25 implementation in Britain that runs or is about to run under 4.2. Ask ...!vax135!ukc!edcaad!edee!keith. tymix!feldman (Steve Feldman) said: Over a year ago, I spoke to Gerald Neufeld at the University of British Columbia, who was working on an X.25 for 4.2. I haven't spoken to him since, so I don't know its status. His phone number was (604)228-6129. I don't have a net address for him. And, of course, lots of people asked for a summary of what I found. I also got a few leads from other sources: Cableshare, Inc., 20 Enterprise Dr., London, ONT. N6A 4L6, Canada, (519) 686-2900, has an LSI-11 based product that connects to a DR11W in the VAX. The version with an 11/23 in it is 5 links, 28.8 Kb total, for $11,400 including all hardware and software. 11/03 version $9,900. 11/73 version $12,525. They do not have a Berkeley 4.2 driver yet, but this is my favorite anyway. Winterhalter and Assoc, 313 N. First St., Ann Arbor, MI 48103, (313) 662-2002, sells X.25 boxes for many micro-computers, including MS-DOS and several UNIX look-alikes, but does not have 4.2BSD. They sell a inexpensive 1 line PAD, though, for PCs. Jim Gast ihnp4!tellab1!tellab2!jim
julian@deepthot.UUCP (Julian Davies) (08/08/84)
And there is a package available from York University in the UK which uses a falcon or lsi-11 front-end-processor, and integrates pad lines and X.25 VC circuits into the unix environment via a set of special files. Versions for various unix releases. The fep code needs a bit of work on its line initialization hand-shaking to suit datapac; I don't know what would happen with Tymnet or telenet for example. deepthot!julian