Message-ID: Newsgroups: net.lan X-Path: utzoo!decvax!ucbvax!ihnss!houxi!npois!harpo!vax135!cornell!dean From: cornell!dean Date: Sun Mar 14 21:49:06 1982 Subject: Request for info on network design X-Google-Info: Converted from the original A-News header I am a member of a group working on software for a campus-wide network at Cornell University. We are trying to address two separate problems. First, an interim system that will allow us to transfer mail and files. Second, a longer term solution that would allow remote file access, virtual channels between processes on different machines, remote login, and possibly even remote paging. We would eventually see the network supporting word processors, workstations, file servers, print servers, mainframes, and public terminals. For the interim system, we are currently using Sytek hardware to provide public terminal access and virtual circuits. We need a protocol that supports file and mail transfer among the various mainframes (an IBM with VM/CMS, a DEC 20 running TOPS-20, a VAX running 4.1BSD, etc.) and between micros (Apple, Terak, etc.) and the mainframes. Currently, we are looking at Columbia University's KERMIT protocol, slightly extended to allow automatic machine/machine transfer. Does anyone have experience with either KERMIT or any similar system in this application? One absolute requirement is that it be implemented on several of the machines described above, and that it takes no more than one to two weeks to implement on a new machine. For the long term problem, we are interested in any powerful and flexible network "protocol" supporting the above features. It would be nice if it were completely specified, and even nicer if it had been implemented on some group of machines running a variety of different operating systems. Please reply directly to me. If there is sufficient interest, I will digest the replies and put them back out on the net. Dean Krafft, Research Associate Dept. of Computer Science Cornell University (uucp address: decvax!cornell!dean, harpo!cornell!dean) (ARPA address: dean.cornell@udel) Message-ID: Newsgroups: net.micro.pc X-Path: utzoo!decvax!decwrl!sun!megatest!fortune!hpda!hplabs!sri-unix!G.DACRUZ@ SU-SCORE.ARPA From: G.DACRUZ@SU-SCORE.ARPA Date: Mon May 2 03:57:13 1983 Subject: Kermit vs. PC Hard Disk X-Google-Info: Converted from the original B-News header Posted: Tue Apr 26 16:22:59 1983 Received: Mon May 2 03:57:13 1983 From: Frank da Cruz This message touches some of the recent hot topics on this list -- free file transfer programs, DOS 2.0, the XT, hard disks, etc. At Columbia University, we develop, maintain, and distribute a file transfer package called KERMIT. It's roughly equivalent to MODEM, except that it has a "nicer" user interface, more documentation, and runs on a wider variety of systems (I think), including the PC. A while back I sent a message to this list to the effect that the availability of PC KERMIT would be announced for the ARPANET/Internet community shortly. In fact, it's ready to go, and has been for some time; I'm just waiting for a suitable distribution point to be set up -- a rather large area is required, since there are dozens of programs for many micros, mainframes, operating systems. Here's the current list: TOPS-10, TOPS-20, IBM VM/CMS on 370-series machines, UNIX on any machine; Z80/8080/8085-based CP/M systems (about 10 different implementations); 8086/8088-based MS DOS systems (several implementations); Apple II; RT-11 with OMSI Pascal; others (like VAX/VMS) on the way. Sources, binaries, hex, documentation, etc, are all provided. Anyway, I hope to have a distribution area ready "soon", and as soon as I do, I'll announce it to this list. Meanwhile, we have heard from some Kermit-86 users who have got either the XT or the IBM hard disk expansion chassis for the PC that Kermit does not work for them. It still can do terminal emulation, but as soon as it starts to transfer a file (in either direction), it hangs. The effect seems to be tied to the presence of the IBM hard disk controller -- even when the disk is turned off and the floppies are selected the symptom persists. The effect is independent of DOS version -- KERMIT works just fine under either 1.1 or 2.0 so long as there is no hard disk. My suspicion is that the new controller is preventing communication port interrupts from getting to the program. I have also heard speculation about seeing different behavior depending upon what slot the serial communication card is plugged in to, but that may be a red herring. Can anyone shed any light on this? Clearly, it must be possible to do file transfer over the communication line when one has a hard disk; I suspect some of the commercial packages may already be doing it. We're groping in the dark, since we don't have an XT, a hard disk, or even 2.0. - Frank da Cruz, Columbia U.