Date: Sun, 26 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 #130 To: tcp-group-digest TCP-Group Digest Sun, 26 Jun 94 Volume 94 : Issue 130 Today's Topics: NOS and the PC Routing Project 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: Sat, 25 Jun 1994 11:41:46 -0600 (MDT) From: Klarsen Subject: NOS and the PC To: tcp-group@ucsd.edu It has been interesting watching the emergence of thought about using a stripped down PC as the TNC of the 1980's. I did like the cost estimates that made sure the tnc-2 from TAPR would be successful. Running these same cost estimates now says run a PC and load the switch software. Here is where we have a problem. The PC is cheap and we can all write code for it with our Borland or Microsoft systems. But in fact we can't get the data from the computer to our RADIO with software... So the parts of the tnc which are not a part of the PC need to be made in the hardware sense. I bought a drsi board and it is in operation at this moment. This is a form of the tnc parts the PC left out. But it is expensive. I know this could be made available in kit form for less cost. It is just a problem to get even this done. Now if you were building a switch for some mountain top today what would a smart person do? On a mountain top a smart person would buy 1 to 4 tnc-2's and radios and hope the lightning isn't too bad this summer. This is because making a PC survive in the wilds is an art not reported on much. There is one on Mt. Bigello above Tucson, AZ that is running the G8BPQ switch and a drsi card and some tnc-2 in kiss mode. The guy who did this a few years ago is not sure it was a good idea. But then PC's were not cheap then... Assuming we can make this PC live in the mountains with the equivelent of,if not exactly an autoexec.bat file to bring it back on line when power returns we face the same problem: do we drive tnc-2's with the kiss e-prom or use a drsi card? Where is the economy coming from? We get the PC cheap. We get the tnc or drsi not-so-cheap. The node gets more expensive... So it is plain the problem comes down to making a plug-into the PC card that has the modem and low level packet forming/reading hardware. I suggest this is where a lot of free time from us hams can pay off. If we get thinking about making high speed modems on a pc card it gets to be fun. No problems trying to get clocks and stuff. It's all there on the pc bus. Think about sending really fast packet. But make it possible to use the available 1200 baud modem chip if needed. Now NOS. The experiance I have had says nos is not stable under high use enviorments. We have a internet chat system in Las Cruces, NM that gets a lot of use forwarding bbs traffic, allowing Russian and other DX stations to use my bbs and the more common chat mode. Both aa5dl and f6cnb have daily problems keeping NOS running. I have no problem with my NOS since it is only used as a bbs and ocassional smtp or ftp; it runs for weeks without crashing. But those with super heavy use will crash in a matter of hours. So what's needed if not already here but unknown to me is a nos that is as stable as dos. If this can be done then nos will work fine as a switch from my experiance. But out there right now is the G8BPQ switch which I have running and has NEVER CRASHED! Unless we need something better than netrom under ax.25 this is the best thing available. To long already but hope it makes sense. 73 de karl k5di@k5di.nm.usa.na ------------------------------ Date: Sat, 25 Jun 1994 10:34:44 -0500 (CDT) From: ssampson@sabea-oc.af.mil (Steve Sampson) Subject: Routing Project To: tcp-group@ucsd.edu agodwin@acorn.co.uk (Adrian Godwin) writes: > 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 ? We're only talking 30k here. 30k is low enough to be insignificant (5%). Anyway DOS is the "Glorified Bootstrap Loader" we need :-) I've since revised the limits to include FTP. That way you could FTP a new version and then reset remotely. But in any case, I don't want to write low level code for the PC components. This is supposed to be cheap, not high tech. Remember, we decided that mass production CPU boards are to be used. If I wanted to write low level code for a floppy or other such necessities I'd start with a new design... Alan Cox reports that his group is nearing completion of a Linux design that could run in 2 Meg and a floppy. I like this idea much better, being an advocate of using better code. For example the TCP/IP is probably no better than Phil's, but the support code is (domain, NFS, etc). Hopefully they will be able to release a floppy image. > Serial ports have some utility, even if you ban KISS TNCs - you can handle > your telephone line comms on the same (shack) router. That was my intention. To get off the single character at a time logic in NOS, and do bulk transfer. That serial code is terrible, and the performance is terrible (I've been able to NOS up to 4800 baud without overruns). I looked at redesign of that part, but it's a major change. > 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 .. I think it's more than a whisker. I was thinking of the Gracilis P-10 when I specified that there would be no "interactive mode" such as used in NOS. It doesn't have to be Telnet (just an idea) and maybe a simple AX.25 connection would be good enough. Anyway the remote command mode in NOS is the only area that needs to be modified in order to bring it out as the main console. This designs advantages are that anyone could build a P-10 with off the shelf parts. The advantage of doing it this way is that you won't have to wait for Gracilis to implement anything. You have Ethernet and PI2 off the shelf now, and the complete design is probably half the cost of the P-10 and does exactly the same thing. I don't mean to denegrate Gracilis, but the group here has found that "Just use a PC" is the only economical way to go. Waiting for others to improve or design a switch using modern technology has proven too expensive. Just look at all the designs in the ARRL Computer Conferences. The way they were written was that the gadget would be in production by the time you read about it. The actual case was that they never left the garage design stage, and did not exist. But even if the design never reached economic feasability, at least they expounded a design philosophy that was correct and interesting to follow. The AWACS radar computer design was developed in 1969 and is a one-of-a-kind DMA driven core memory general purpose computer (I think the designers are all dead - the logic devices "MECL" certainly are. This was my first experiance to never buy digital devices which start with 'M') device. It runs at 10 MHz and could probably run rings around any 66 MHz PC in I/O. But the upgrade computer was taken off the shelf. Rather than re-invent the wheel (which Westinghouse really wanted - they designed a new computer from scratch) the Air Force specified that it should be a current Mil-Spec computer. Westinghouse just plugs two of these into the equipment rack and saves millions. It's not what they want, it's all they can afford. On that note, I'll agree, it's all we can afford - to use a PC. -- Steve ------------------------------ End of TCP-Group Digest V94 #130 ******************************