Date: Sun, 12 Dec 93 04:30:02 PST
From: Advanced Amateur Radio Networking Group <tcp-group@ucsd.edu>
Errors-To: TCP-Group-Errors@UCSD.Edu
Reply-To: TCP-Group@UCSD.Edu
Precedence: Bulk
Subject: TCP-Group Digest V93 #320
To: tcp-group-digest
TCP-Group Digest Sun, 12 Dec 93 Volume 93 : Issue 320
Today's Topics:
Extended ANSI Characters...
netrom and datagram and ...
UDP Lock-Up & Sig File...
UDP Locking Things Up
VC vs Datagram (2 msgs)
Send Replies or notes for publication to: <TCP-Group@UCSD.Edu>.
Subscription requests to <TCP-Group-REQUEST@UCSD.Edu>.
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, 11 Dec 93 21:30:04 UTC
From: n8wei@N8WEI.AMPR.ORG
Subject: Extended ANSI Characters...
To: kf8nh@kf8nh.ampr.org
"Brandon S. Allbery" <bsa@kf8nh.wariat.org> Writes:
> Unix systems support real standards. The 8-bit graphics characters are NOT a
> standard, except insofar as they aree present on all 80x86 PCs; but not on
> Mac or Amiga, much less Sun, AGI, DEC, IBM workstations, etc., etc., etc.,
> unless someone decides to cater to DOS PCers who think they have a standard
> just because they're DOS PCers.
>
> There IS an ISO graphics standard (ISO 2022). PC graphics aren't it.
>
> ++Brandon
> --
> Brandon S. Albery kf8nh@kf8nh.ampr.org bsa@kf8nh.wariat.org
> "MSDOS didn't get as bad as it is overnight -- it took over ten years
> of careful development." ---dmeggins@aix1.uottawa.ca
> Do not taunt Happy Fun Coder. (seen on the Net...)
>
Point well taken Brandon. Although, this was not the point of my Comment.
The point was that Ron had sent me a message concerning the fact that my sig
file was the way it was, and to ask a question concerning my previous message
to nos-bbs (UDP screwing up the display updating). Then, after a paragraph or
two, he just puts in this "plug" for Unix. It did not realy make sense in the
context that it was in.
Let me note that no offence was ever taken at Ron's letter. He had a right to
send it, and I am glad that he did. I had already knew that the High-ASCII
characters were unique to 80x86 family of computers, but had inadvertantly put
them in my "sig" file without thinking. If it wasn't for him, I would not have
even thought twice about it, and it would have stayed that way.
The main point of my reply, and my reason for sending it to the same groups
that I sent the original message concerning UDP, is the fact that Ron did not
understand which problem I was speaking of. I assumed that there might be
others out there who had not remembered the discussions about this problem, so
I thought that they might get a use out of the explination also.
My point was not to discuss why Unix is a better operating system. As a matter
of fact, I believe that myself (even though I run Messy-DOS). I too would be
running Unix if it wasn't for the fact that Most everyone that I know runs
MS-DOS as well, and I do a lot of custom programming for those people.
i.e. Log-File sorting program for our 80x386 based Repeater Controller,
and other helpfull Ham-Related software programs.
73 DE N8WEI...
C-Ya'...
*---------------------------------*------------------------------------*
| ((N8WEI) TCP/IP) PBBS 147.560 | Todd W. Powers (N8WEI) |
| ------------------------------- | 4245 Stonebridge Road SW |
| Packet Radio: | Wyoming, MI 49509 |
| N8WEI @ N8WEI.AMPR.ORG | ---------------------------------- |
| N8WEI @ N8WEI.#SWMI.MI.USA.NA | |
| Internet: | Borland C++ & FoxPro Programmer |
| N8WEI @ HAMGATE.GVSU.EDU | |
*---------------------------------*------------------------------------*
------------------------------
Date: Sat, 11 Dec 1993 08:11:54 -0700 (MST)
From: Klarsen <klarsen@acca.nmsu.edu>
Subject: netrom and datagram and ...
To: tcp-group@ucsd.edu
I think a little understanding of the netrom system is needed. A
recient message was about which is better datagram or vc and it was said
that with datagram a mss of 256 is correct. It isn't. If your using a
netrom link then you must pay the penalty of a max data packet size of 230
bytes. I see 256 used a lot in the .doc's and it is fine IF your a direct
connection and don't use a netrom node, or, if you digipeat you can use
256 bytes.
A few months ago I did a study of our links between Tucson, AZ and
Las Cruces, NM. I started with mss=150 and datagram mode and with time
went to 256. I saw a major slowdown at about 240 bytes. It got worse as I
went towards 256. So the data is clear: don't use 256 in a netrom network.
73, karl k5di
------------------------------
Date: Sat, 11 Dec 93 15:12:53 UTC
From: n8wei@N8WEI.AMPR.ORG
Subject: UDP Lock-Up & Sig File...
To: ron@chaos.eng.wayne.edu
ron@chaos.eng.wayne.edu (Ron Atkinson N8FOW) Writes:
>
> Can you be more specific on the problem and the exact sequence leading up to
> the occurance or your suspected problem? Nothing was mentioned on what the
> problem was and the info isn't enough to indicate. Looked pretty normal to
> me. Do you have smtp usemx on ? Are you querying a domain server? Is
> the *pause* when you enter something like: s nos-bbs@hydra.carleton.ca and
> it takes a while for the Subject: line to come up? If so then this is
> 100% normal while querying a nameserver. What's the problem you're having./
> Another thing, is that supposed to be an ANSI signature? If so can
> you please get rid of it and replace it with regular ASCII characters? Us
> folks using Unix systems typically can't decode the PC ANSI junk and the
> signature files look completely hosed up. And I feel that the Unix
> system rules everytime over a DOS system.
>
> Ron N8FOW
>
UDP Locking Things Up...
------------------------
Well, Ron...
We have discussed this numerous times in NOS-BBS, & TCPGROUP. The "locking up"
in effect, screws up the Screen Updating. Say for instance, when I had
telneted to my self and was recording all of those values, When I type the
SOCKET command, about half the list appears on the screen, sometimes leaving
the cursor in the middle a line somewhere in the vacinity of 2 in. from the
actual bottom of the text. (Maybe 10-15 lines) Well, if I keep hitting the
socket command, it will scroll, but still cutting off the bottom so many
lines. If I go to annother session (including the Net> prompt, or the trace
screen), then come back, the "buffer" of characters has now appeared and
been updated to the screen. So, say when I send a message, I have to keep
switching out of, and back into that session so I can see if I have reached the
end of the line.
As far as a sequence leading up to this pheonomina, there isn't one. It just
happens whenever it wants to. (Notice the UDP socket. Whenever this happens,
is when my screen update gets screwed up).
Hopefully you will understand this.
The SIG FIle...
---------------
I am totally sorry, I wasn't even thinking. I have been using that Sig arround
G.R. for a few days, and didn't remember to change it back...
Anyway, It's just a picture of Santa Clause in his sleigh, flying over G.R.,
and saying Merry Christmas.
P.S Can you tell me what your comment about Unix being superior has to do
with the fact that your computer won't decifer the Extended ANSI
characters?
73 DE N8WEI...
C-Ya'...
*---------------------------------*------------------------------------*
| ((N8WEI) TCP/IP) PBBS 147.560 | Todd W. Powers (N8WEI) |
| ------------------------------- | 4245 Stonebridge Road SW |
| Packet Radio: | Wyoming, MI 49509 |
| N8WEI @ N8WEI.AMPR.ORG | ---------------------------------- |
| N8WEI @ N8WEI.#SWMI.MI.USA.NA | |
| Internet: | Borland C++ & FoxPro Programmer |
| N8WEI @ HAMGATE.GVSU.EDU | |
*---------------------------------*------------------------------------*
73's DE N8WEI...
C-Ya'...
*---------------------------------*------------------------------------*
| ((N8WEI) TCP/IP) PBBS 147.560 | Todd W. Powers (N8WEI) |
| ------------------------------- | 4245 Stonebridge Road SW |
| Packet Radio: | Wyoming, MI 49509 |
| N8WEI @ N8WEI.AMPR.ORG | ---------------------------------- |
| N8WEI @ N8WEI.#SWMI.MI.USA.NA | |
| Internet: | Borland C++ & FoxPro Programmer |
| N8WEI @ HAMGATE.GVSU.EDU | |
*---------------------------------*------------------------------------*
------------------------------
Date: Sat, 11 Dec 1993 07:07:47 -0600 (CST)
From: ssampson@sabea-oc.af.mil (Steve Sampson)
Subject: UDP Locking Things Up
To: TCP-Group@UCSD.Edu
I've seen this problem I think when I was using DNS to look up a name that I
just typed in on the mailbox. The name wasn't in the table and everything
locked up waiting for the name to be resolved or timeout. In your case you
seem to be indicating "timer" and you have a Timer Listener process. Are these
two the same? What's a timer listener and do you really need it. If not,
maybe the timer is just a side affect of the DNS problem I saw. Probably
a DNS table design or record error.
---
Steve
------------------------------
Date: Sat, 11 Dec 1993 07:46:11 -0600 (CST)
From: ssampson@sabea-oc.af.mil (Steve Sampson)
Subject: VC vs Datagram
To: TCP-Group@UCSD.Edu
> 1. VC vs Datagram
The problem I've seen is that you need different parameters for each
interface. For example I try to send datagrams to a guy who mostly does
VC over Rose. His parameters (MTU, MSS, IRTT) are so small and times so
large, that any datagram failure puts me back to the initial. If there was
an interface structure which had different parameters for each, then this
problem would go away. One advantage NOS and NET have is that they use
AX.25 segmentation. That is, you can set the MTU (or MSS - I forget) very
large, and the program will fragment at the AX.25 level. That is, only one
IP header and all the data fragmented to the PACLEN. This doesn't work over
Net/Rom (Net/Rom is obsolete so we don't care) but does over Rose. Rose can't
handle very many fragments though. It can handle about two, and after that
it falls on it's [expletive]ing face (as normal). AX.25 segmentation is too
slow, so it's not a design issue any more.
Pure datagram on the other hand is fast.
The trouble is hams are cheap. The existing circuits in the central U.S. are
Rose based. They're Rose based because a commercial telephone circuit
donation is used between Texas, Florida, and Mich (I believe). A NOS node at
each end would quickly convert everyone and finally put Rose to rest, but that
is probably in the far future.
The X-1 node chip (TheNet) is the current state of the art remote site node
switch. Since most Rose sysops are past Net/Rom sysops, they are not really
interested in doing something they've already done, and Net/Rom was mostly a
failure (a nice idea in theory - doesn't work in practice - too slow).
What they miss though, is the fact that it's two nodes in one. If I was on
the design team I would [expletive]-can the whole Net/Rom part of the code
and keep the IP router part. Maybe make it a dynamic router, rather than a
static one.
I think new networking systems should be based on datagram, and the remote
sites should be X-1 (or X-2 when it's released) and the fixed sites with the
DNS and other servers should be JNOS based, or Unix based (preferably Unix
based).
With this we can quickly kill off the current BBS systems and switch to NNTP,
and provide intelligent email services. My personal experiance with datagram
is that I wouldn't hessitate to send a 1 Meg file over 9600, but I don't even
send 2k over a VC path.
---
Steve N5OWK
------------------------------
Date: Sat, 11 Dec 1993 13:13:45 -0800
From: karn@qualcomm.com (Phil Karn)
Subject: VC vs Datagram
To: ssampson@sabea-oc.af.mil
>The problem I've seen is that you need different parameters for each
>interface. For example I try to send datagrams to a guy who mostly does
This has been the case for my own NOS code for some time now. I guess
nobody's tracking it anymore.
I got rid of the "mode" command, and created different attach modes to
control this. For example, the KISS TNC is connected to a serial port
on my machine with the command
attach asy 0x2b0 s kissui com3 2048 512 9600 c
The "kissui" parameter sets the default AX.25 encapsulation mode to use
UI frames. As before, it can be overridden with bits in the IP TOS field.
The "kissi" parameter could have been specified, in which case connected
mode (VC) AX.25 would be used by default.
BTW, the values 0x2b0 for the address and "s" for the IRQ are because I'm
using a 4-port card with shared interrupts. I've had support for that for
some time too.
Phil
------------------------------
End of TCP-Group Digest V93 #320
******************************
******************************