| |
ZL3AI > APRDIG 11.10.03 10:01l 227 Lines 9930 Bytes #999 (0) @ WW
BID : 2127-ZL3AI
Read: GUEST
Subj: TAPR Digest, Oct 08, 3/5
Path: DB0FHN<DB0RGB<OK0PPL<DB0RES<ON0BEL<IR0NET<VK3TE<GB7YFS<GB7BHM<ZL2BAU<
ZL3VML
Sent: 031010/1115Z @:ZL3VML.#80.NZL.OC #:10120 [Chch-NZ] FBB7.00i $:2127-ZL3AI
From: ZL3AI@ZL3VML.#80.NZL.OC
To : APRDIG@WW
Subject: Re: Abusive User paths
From: Wes Johnston <wes@johnston.net>
Date: Wed, 08 Oct 2003 10:24:14 -0400
X-Message-Number: 12
To: Bob Bruninga <bruninga@usna.edu>
Subject: Re: [aprssig] Abusive User paths
At 09:11 AM 10/8/2003, you wrote:
>ALl these ravings about new digs and new software and new protocols etc
>have NOTHING to do with fixing the fundamental problem of everyone trying
>to FLOOD an existing network with 10 times more traffic than it can
>handle. And I have shown how with the new n-N paradigm, we can FIX
>that problem WITH EXISTING digs, with just a few simple commands WITHOUT
>having to FIX the users..
The problem is the 1000 digis which are running firmware which is
defective... and floods us with dupes....
For example, a station in NC is flooding with RELAY RELAY WIDE WIDE for a
path....
I typically see 8 or 9 unique copies of that packet down here in central
SC.... I'm not talking about children of the packet... I'm talking about 6
relay stations heard the packet, and the several WIDE nodes see a different
callsign (or RELAY) version of the same packet and digipeat it. Let's say
6 relay stations hear this packet... 6 wide nodes hear those 6 copies (36
packets)... each of those wide nodes SHOULD have the good sense to not
digipeat the PAYLOAD but once.... but they don't.... each of the
digipeaters repeats this packet each time it hears a different copy of
it. The network is broken and morons running RELAY RELAY in their path can
exploit it! This is nearly as bad as the morons who write software to
exploit microsoft outlook email program....
>But lets stop complaining and get out there and THINK about our LOCAL
>APRS networks, WHAT LOCAL Service we want it to perform, over WHAT AREA,
..
..
..
>Use the new n-N paradigm to limit FLOODING and make the network IMMUNE to
>user errors and abusive paths while making it universal.
I am thinking local.... I'm running the new LanN-n system, and staying
backward compatible with WIDE and RELAY... and I'm still flooded by a moron
running a silly path of RELAY RELAY WIDE WIDE.
Since we can't reverse engineer KPC software (and they are unlikely to
issue a fix for kpc3 which is obsolete), what if we unplugged the
microprocessor and installed a daughterboard with an AVR or PIC on it. Or
rewired John Hansen's TNC to use the hardware interface (modem chip and PTT
lines) to fit this daughterboard.
Wes
----------------------------------------------------------------------
Subject: varying the coverage of a RELAY
From: Wes Johnston <wes@johnston.net>
Date: Wed, 08 Oct 2003 10:29:33 -0400
X-Message-Number: 13
We seem to have a problem in our town (which shouldn't be a problem if dupe
checking worked).... but in a nutshell, our problem is that we have
overlapping coverage between relay stations. 30% of our area overlaps, but
70% does not... and that 70% is important to us.
KC4PL and I were talking on the way to work this morning. What we need is
to listen on a yagi, and TX on a 1/4 whip. I asked about using a relay to
switch antennas and he suggested using the PTT line from the TNC to pull
the relay in to swap antennas in addition to keying the radio. Then he
suggested we use a spare set of contacts on the relay to actually trigger
the PTT on the radio.... cascaded PTT line.. the Radio can't key up until
the relay is in the TX position.
TNC PTT -> relay (coax and spare contacts)-> RADIO PTT
anyone know of a relay which is suitable for this application?
Wes
----------------------------------------------------------------------
Subject: WIDEN-N copies
From: Bob Bruninga <bruninga@usna.edu>
Date: Wed, 8 Oct 2003 10:45:55 -0400 (EDT)
X-Message-Number: 14
One person took me to task for my sum of 111 copies of each packet being
generated in a 7 hop WIDEn-N path as posted on:
http://www.ew.usna.edu/~bruninga/aprs/fix14439.html
That total was a pesimistic assumption that 3 digis heard the initial
packet. I have re-done the assessment as shown below and the answers are
a total of 74 if only two initial digis hear it and only 64 if only one
digi hears the initial packet. (Still an awfull lot of packets).
Below is if the original packet hit TWO digis at
the beginning and each digi can hear 3 others.
SETTING 1st 2nd 3rd 4th 5th 6th 7th TOTAL
------- --- --- --- --- --- --- --- -----
WIDE1-1 2 2
WIDE2-2 2 4 6
WIDE3-3 2 4 8 14
WIDE4-4 2 4 8 10 24
WIDE5-5 2 4 8 10 14 38
WIDE6-6 2 4 8 10 14 16 54
WIDE7-7 2 4 8 10 14 16 20 74
Below is if the original packet only hit one digi
at the beginning and each digi can hear 3 others.
SETTING 1st 2nd 3rd 4th 5th 6th 7th TOTAL
------- --- --- --- --- --- --- --- -----
WIDE1-1 1 1
WIDE2-2 1 3 4
WIDE3-3 1 3 6 10
WIDE4-4 1 3 6 9 19
WIDE5-5 1 3 6 9 12 31
WIDE6-6 1 3 6 9 12 15 46
WIDE7-7 1 3 6 9 12 15 18 64
Again, ANYTHING beyond 2 or 3 is just far too much QRM for our system to
handle in most areas. ANd also it is much worse if most digis can hear 4
or more other digis. Rememebr that resulted in almost 200 copies!
de WB4APR, Bob
----------------------------------------------------------------------
Subject: Re: Abusive User paths
From: Bob Bruninga <bruninga@usna.edu>
Date: Wed, 8 Oct 2003 11:14:35 -0400 (EDT)
X-Message-Number: 15
>The problem is the 1000 digis which are running firmware which is
>defective... and floods us with dupes.... For example, a station in NC
>is flooding with RELAY RELAY WIDE WIDE for a path....
Yes, but we need to get real... There are 1000 digipeaters out there and
they are not going to be UPGRADED anytime soon..
Instead of fussing over some dupes from a few bad apples (which is a waste
of time), lets get on with simply getting the SYSOPS to figure out where
their ALOHA areas of traffic needs are and simply SEND the remote commands
to the digis to FIX THEM. We must PLAN our APRS and how we intend to use
it. We also have means NOW to solve the bad-apples too...
Do not interpret my position as holding back APRS from progress. I am all
for new digis and everything else, but I am trying to get us to FIX what
we can, TODAY!??? Its simply sending a few parameters to existing
digis...
AND changing the way we all think about FLOODING...
>I typically see 8 or 9 unique copies of that packet down here in central
>SC.... The network is broken and morons running RELAY RELAY in their
>path can exploit it!
Yes, Absolutely frustrating... but complaining isnt going to fix it.
The new n-N paradigm I am offering gives SYSOPS a solution to MOST of
the flooding problems. IF someone wants to fix his LAN today, then he
should take the new n-N paradigm. If he wants to just dream about the
future, then he can continue to dream on.... and/or work for new digis in
his area...(this is a great longer term goal)...
>>But lets stop complaining and get out there and THINK about our LOCAL
>>APRS networks, WHAT LOCAL Service we want it to perform, over WHAT
>>AREA, ... Use the new n-N paradigm to limit FLOODING and make the
>>network IMMUNE to user errors and abusive paths while making it
>>universal.
>I am thinking local.... I'm running the new LanN-n system, and staying
>backward compatible with WIDE and RELAY... and I'm still flooded by a moron
>running a silly path of RELAY RELAY WIDE WIDE.
Ah, yes, I certainly appreciate all you are doing. I AGREE with your
frustration... but here then is the next round of things that needs to be
done:
1) eliminate the RELAY alias from ALL home stations except those specially
designated as RELAY digs in FILL-IN areas which CANNOT hit a digi
otherwise...
2) This is a fun challenge, but it is easy to do in the case of a "moron
causing problems". SImply look at the map in the vicinity of the
offending station to see who is RELAYing his packets and target them with
messages and Email asking them to change their ALIAS away from RELAY.
3) You could even ask them to BUDLIST the guy.
4) Again, looking at your LAN, see what "firewall" digi is bringing his
floods into your area and BUDLIST him in those few firewall digis
5) Now then we DO need MORON filters, but now with the new paradigm, we no
longer have any problems with any value of N. The ONLY bad path under the
new paradigm is anything with more than 2 RELAY's or WIDE's. This is the
real advantage of the new paradigm, is that there is now only ONE RULE
for BAD PATHS. That is, No more than 2 RELAY's or WIDE's. If so,
TRUNCATE or IGNORE the packet and FEEDBACK to the user...
By removing the "N" from the determination of what is GOOD or BAD or what
is something that someone needs to use in a given situation, we have
boiled it down to a single RULE. This makes education and enforcement FAR
SIMPLER than it has ever been before.
>Since we can't reverse engineer KPC software (and they are unlikely to
>issue a fix for kpc3 which is obsolete), what if we unplugged the
>microprocessor and installed a daughterboard with an AVR or PIC on it.
>Or rewired John Hansen's TNC to use the hardware interface (modem chip
>and PTT lines) to fit this daughterboard.
Certainly great ideas. But great ideas often never get implemented by
the majority. Lets FIX what we can now with what we CAN do now... Yes,
continue to work on the new hardware/software fixes. But lets all think
about our LAN's and what we can do NOW with just settings and
forethought...
de WB4APR@amsat.org, Bob
----------------------------------------------------------------------
Read previous mail | Read next mail
| |
