Date: Sat, 25 Jun 94 04:30:02 PDT From: Advanced Amateur Radio Networking Group Errors-To: TCP-Group-Errors@UCSD.Edu Reply-To: TCP-Group@UCSD.Edu Precedence: Bulk Subject: TCP-Group Digest V94 #129 To: tcp-group-digest TCP-Group Digest Sat, 25 Jun 94 Volume 94 : Issue 129 Today's Topics: [BTITMARS%ESOC.BITNET@vm.gmd.de: help ip tos] History and the Final TNC (2 msgs) IP-TNC, the beginings? (2 msgs) MOSAIC Server (2 msgs) Router Project Router Project (TNC from Hell) (3 msgs) Telnet Client for JNOS Send Replies or notes for publication to: . Subscription requests to . Problems you can't solve otherwise to brian@ucsd.edu. Archives of past issues of the TCP-Group Digest are available (by FTP only) from UCSD.Edu in directory "mailarchives". We trust that readers are intelligent enough to realize that all text herein consists of personal comments and does not represent the official policies or positions of any party. Your mileage may vary. So there. ---------------------------------------------------------------------- Date: Fri, 24 Jun 1994 10:25:43 -0700 From: Phil Karn Subject: [BTITMARS%ESOC.BITNET@vm.gmd.de: help ip tos] To: karn@unix.ka9q.ampr.org Barry followed up with a direct note explaining the behavior: there was a mixture of TCP packets from FTP data and control channels. and these packets (from a Linux box) had different TOS settings (as per the more recent IETF recommendations). The NOS box was only doing what it had been designed to do with these different TOS settings. Now suggestions on what NOS *should* do with those bits are of course always welcome -- I came up with the current scheme probably 8 years ago when I was originally designing the code and haven't changed it since. Phil ------------------------------ Date: Fri, 24 Jun 94 13:24:43 +0100 From: agodwin@acorn.co.uk (Adrian Godwin) Subject: History and the Final TNC To: tcp-group@UCSD.EDU > From: ssampson@sabea-oc.af.mil (Steve Sampson) > enjoyable. Finally you probably shouldn't judge the journey on what is > printed > in the tcp-group archives. Only a fraction of the discussions get in there, > while the subscribers talk amongst themselves. Sort of like contract > side meetings with only the major points being printed :-) Makes it even worse .. all that private talk and the same ideas are still put forward because we've done the frilly bits instead of the real jobs, like non-subnet routing and fixing the RF layer. Apologies to anyone who feels they've made a bigger contribution : I'm looking at my performance and generalising. In the meantime, the rest of the IP world has got closer to a wired model, making it necessary to limit our view of what's really possible to what's possible in an imperfect RF simulation of wired nets. > I find the word "horrified" pretty amusing; what did you expect to find given > the state of packet radio in most communities. Packet radio will always be > a side interest in amateur radio. It can't compete with CW W.A.S. awards or > weekend contests with robot cq machines. Packet means sharing, and that's a > hard road to follow in most first-world countries. OK, rewrite 'I was horrified' as 'I cringed' - I'm really not that naive :-). -adrian ------------------------------ Date: Fri, 24 Jun 1994 16:42:19 -0700 From: Phil Karn Subject: History and the Final TNC To: ssampson@sabea-oc.af.mil >> It's no wonder that some of those contributors have disappeared from >> the list >Most of them graduated and went to work... Or changed jobs and started doing this stuff for real... Phil ------------------------------ Date: Fri, 24 Jun 94 07:14:00 -0000 From: mikebw@bilow.bilow.uu.ids.net (Mike Bilow) Subject: IP-TNC, the beginings? To: tcp-group@UCSD.EDU Cc: jackb@mdd.comm.mot.com In a msg on , Jack Brindle writes: >>Phil Karn said: >>> What's wrong with PC clones? They're widely available, cheap >>> and give a pretty good bang for the buck. >> >>I must agree... JB> Gosh, the chorus of "I agrees" is quickly beginning to sound like JB> "Because it's always been done that way." There actually are JB> better and cheaper ways to do the task, plus some folks JB> (myself included) actually ENJOY designing hardware. The PC clone is not an especially attractive piece of hardware as a networking platform. For example, there are limited IRQs, and they are all edge triggered. The port addressing structure looks like someone went out of their way to make the whole 80x86 family emulate an 8080A. The bus not only runs at 8 MHz, but still requires wait states on I/O. I could go on, but the PC clone is the platform of choice for a lot of practical reasons. You could obviously design special purpose hardware to outperform a PC clone, perhaps by an order of magnitude, for similar manufacturing costs in similar quantity. However, there is no real hope of getting similar quantities manufactured, and your special purpose hardware will never turn up on the used market. Nevertheless, cost is still not the driving factor for use of the PC clone platform. The deciding issue is ease of software development. Most of us already have compilers for the 80x86 PC clone, and a fair amount of experience working with them. Almost every important development tool is available on the 80x86 platform somehow; GNU C++ has even been ported to MS-DOS (although it does require a 386 or better). JB> Then Admiral Grace Hopper would no longer have to worry about the JB> rut we have fallen into... I have some bad news for you about the admiral, I'm afraid. -- Mike ------------------------------ Date: Fri, 24 Jun 1994 10:10:09 -0400 From: hprice@bektek.com (Harold Price) Subject: IP-TNC, the beginings? To: tcp-group@ucsd.edu > Jack Brindle said: >>Phil Karn said: >>> What's wrong with PC clones? They're widely available, cheap and give >>> a pretty good bang for the buck. >> >I must agree... >Gosh, the chorus of "I agrees" is quickly beginning to sound like "Because >it's always been done that way." There actually are better and cheaper >ways to do the task, plus some folks (myself included) actually ENJOY >designing hardware. There may be better ways. There will only be a cheaper way if someone throws in the hardware development costs for free, and gives the hardware away (that is, sell it in hundreds quantities but priced to compete with a commodity sold in the millions). I mention this in June QEX. Remember why we have the TNC legacy. In 1984, a computer cost multiple thousands. I bought an IBM PC (4 MHz 8088) for $3000. The TNC-2 gave practically the same performance, and came with built-in I/O, for $219. A 1200 baud phone modem cost more than that. TNCs were the cheap way to get on the air, by an order of magnitude. Now, you can buy a 386 motherboard (50x the tnc-2 performance?) for less than the original TNC-2. 14.4kb phone modems are $99. We get software solutions like NOS because people are willing to donate huge amounts of development effort, and give away the resulting implementation. We seldom get free hardware solutions because, while people are willing to donate the development time, they are unwilling to give away assembled and tested hardware boxes. So it goes. >Perhaps it's again time for someone to invoke the "code rule" by designing >better, cheaper hardware and software. As you say. It won't happen any other way. The cheaper hardware will be a problem, though. Harold ------------------------------ Date: Fri, 24 Jun 94 19:01:22 UTC From: eb3aod@albinyana.etse.urv.es Subject: MOSAIC Server To: tcp-group@UCSD.EDU Hola a todos!! Does anybody know where i can find the latest version of Mosaic Server for Macintosh?? I would like do a prove with Mosaic and amateur radio world ... Saludos de Curro eb3aod e-mail : curro@etse.urv.es AX25 : eb3aod@ea3rdt.eat.esp.eu "Data Highway" : In my Dreamland ... ------------------------------ Date: Sat, 25 Jun 94 11:53:38 UTC From: eb3aod@albinyana.etse.urv.es Subject: MOSAIC Server To: tcp-group@UCSD.EDU Hello again : Thanks a lot to all who answer me about Mosaic Server for Apple. But becarefull: the file NCSAMosaicMac.200A2.sea.hqx is a Mosaic Client if any want to up a Mosaic Server he/she must get the file (as Ian said) : MacHTTP.sit.hqx in 129.106.30.1 /public/mac/MacHTTP Yes, i Know that a Mosaic csn work via radio. (I proved it with a "Front-End Radio"). But i want to up a Mosaic Server which ha s a Amateur-Radio "side". I will try to link with the Mosaic server my PR - Internet Gateway and all of services that this gateway (albinyana.etse.urv.es) are giving ... Well thanks all to read me Saludos de Curro eb3aod e-mail : curro@etse.urv.es AX25 : eb3aod@ea3rdt.eat.esp.eu "Data Highway" : In my Dreamland ... ------------------------------ Date: Sat, 25 Jun 1994 04:11:21 -0500 (CDT) From: ssampson@sabea-oc.af.mil (Steve Sampson) Subject: Router Project To: tcp-group@ucsd.edu Gerard J van der Grinten writes: > But to just "blast" a monitor prom who [that] forgets all about a keyboard > is a task greater as [than] writing the router code. There are a LOT of > different chip sets doing mmu, refresh, and other tasks for the zillion > versions of Mother(f...)boards [:-)]. FYI - I've had the AMI BIOS for the last few PC boxes I've owned, and they have a toggle for keyboard and video in CMOS. -- Steve ------------------------------ Date: Fri, 24 Jun 94 13:48:40 +0100 From: agodwin@acorn.co.uk (Adrian Godwin) Subject: Router Project (TNC from Hell) To: tcp-group@UCSD.EDU > From: ssampson@sabea-oc.af.mil (Steve Sampson) > > Develop code for an MS-DOS computer using Borland C++ which > has the following components: But encapsulate the OS dependencies with a view to removing the OS altogether : on a system that has no applications, what does DOS do to justify it's memory usage after it's loaded NOS ? Freedom from DOS would give freedom to use the hardware as a real embedded system - cross-compiled, protected mode. > There will be no other options, serial ports will not exist. Prior > code from NOS or JNOS can be used. Serial ports have some utility, even if you ban KISS TNCs - you can handle your telephone line comms on the same (shack) router. > > Comments > > Ummm. I know your proposal is a gnat's whisker from a NOS switch anyway, but if it's better to spend money than redevelop existing solutions, there's already a NOS-in-a-box from Gracilis. No DOS dependency, good performance, ethernet Real Soon Now. Would it be as cheap as that PC if everybody in the market for a TNC from hell bought one ? 68302s are finally available in cheap packaging .. -adrian ------------------------------ Date: Fri, 24 Jun 94 15:20:00 -0000 From: mikebw@bilow.bilow.uu.ids.net (Mike Bilow) Subject: Router Project (TNC from Hell) To: tcp-group@ucsd.edu Cc: gvdg@nlr.nl GJvdG> PC boards make a easy platform, TRUE. GJvdG> But to just "blast" a monitor prom who forgets all GJvdG> about a keyboard is a GJvdG> task greater as writing the router code. There are a LOT of different GJvdG> chip sets doing mmu, refresh, and other tasks for the zillion versions GJvdG> of Mother(f...)boards. And the simple logic of "No GJvdG> keyboard present, strike GJvdG> F1 to continue" indeed withholds mountain top GJvdG> operations. Bact to the X820... This is a simple problem. You don't replace the original BIOS ROM, but leave it installed. Then you burn your new ROM as a BIOS extension ROM that sits in the address space between the C800 and EC00 segments, using a header that makes the BIOS ROM pick it up during the POST scan. With modern BIOS versions, especially the AMI, you just set the keyboard to "Not Installed" on the ordinary setup screen. This causes the machine to use the keyboard if it is found, but not to halt the boot process if no keyboard is found. Other BIOS makers use a "Halt on any errors" setting that can be selected or not as you choose. -- Mike ------------------------------ Date: Fri, 24 Jun 94 13:58:06 From: jks@giskard.utmem.edu Subject: Router Project (TNC from Hell) To: tcp-group@UCSD.EDU Steve Sampson writes: > This device will interface the RF Local Area Network with the Baseband > Local Area Network. It will allow connected and unconnected AX.25 protocol > encapsulated IP to be converted to Ethernet encapsulated IP and vice versa. > There will be no FTP, or SMTP, but a command oriented Telnet port for > configuration will be available (both RF and Baseband access). Other RF > protocols above Level 2 will not be used (Net/Rom). *BINGO* > Use KA9Q NOS rather than JNOS. JNOS has such a large development group that > it is almost unreadable. Exact and major problem! > Once the code is developed it should be frozen. Gotta Do! > Any other > modifications such as BBS and interface to the washing machine and dryer should > be taken to a new baseline and version name and extended from there. This > version should be fixed as a small router with a command interface for > configuration purposes. The intent is to operate as a backbone on both RF > paths, LAN paths, and Internet paths. If a BBS is desired then a JNOS box > should be interfaced to the network using Ethernet. Well summed up! I think this should be a very do-able project. This would leave anyone elses NOS preferences alone but allow for the kind of stable performance needed in a gateway. Let's do it! Hardly a TNC from Hell tho' KD4IZ ------------------------------ Date: Fri, 24 Jun 94 10:46:47 EST From: "Shannon, Lew" Subject: Telnet Client for JNOS To: tcp-group@UCSD.EDU From: rtorres@tazz.coacade.uv.mx >Hi!! Does somebody knows about a good Telnet Windows client for the KA9Q or >JNOS?? All the telnet clients works perfectly on Unix machines but for the NOS >it needs the \127 key as backspace and for enter. Try Winqvt. You can specify the backspace character, remap the enter key, and get local echo. That combo works very well with Jnos. Lew, K0RR ------------------------------ End of TCP-Group Digest V94 #129 ******************************