| |
ZL3AI > APRDIG 03.10.03 09:42l 322 Lines 13974 Bytes #999 (0) @ WW
BID : 2063-ZL3AI
Read: GUEST
Subj: TAPR Digest, Sep 30, 6/16
Path: DB0FHN<DB0RGB<OK0PPL<DB0RES<ON0BEL<VK7AX<ZL2BAU<ZL3VML
Sent: 031002/1412Z @:ZL3VML.#80.NZL.OC #:9560 [Chch-NZ] FBB7.00i $:2063-ZL3AI
From: ZL3AI@ZL3VML.#80.NZL.OC
To : APRDIG@WW
Subject: How to locate someone?
From: "Michael J. Wolthuis" <wolthui3@msu.edu>
Date: Tue, 30 Sep 2003 14:16:53 -0400
X-Message-Number: 23
The following packet shows up in my area:
APRS>APRS,TCPXX*,qAX,APRS:@301811z4248.00N/08542.00W_325/014g...t050P002
h39b10055.DsVP
However, it is not on RF and is on TCP only. Driving to the location
reveals nothing there that resembles a computer or transmitter. How do
I determine this incorrectly setup station's actual location so I can
have them put their corrected callsign inplace of APRS?
Please advise,
Mike
Kb8zgl
----------------------------------------------------------------------
Subject: cannot see others out there using APRSpoint
From: "JBCrafts" <jbcraft@adelphia.net>
Date: Tue, 30 Sep 2003 14:21:17 -0400
X-Message-Number: 24
with sincere thanks to Michael, I got APRSPoint working again... well,
sorta.
I can connect to the TNC. I can cause the TNC to transmit my data, I see my
posit on the map, but I do not see others!
My DCD led tells me that the TNC is seeing traffic...
I opened Hyperterm and I see the traffic thru the TNC, I can send and
receive from the CMD line in Hyperterm...
I then close hyperterm, reload APRSPoint and still nothing is posted onthe
map.
What am I forgetting????
Bob K8YS
----------------------------------------------------------------------
Subject: Re: Digipeater Duplicate Checking is broke
From: Bob Bruninga <bruninga@usna.edu>
Date: Tue, 30 Sep 2003 14:25:18 -0400 (EDT)
X-Message-Number: 25
On Tue, 30 Sep 2003, Scott Miller wrote:
>Ok, here's a physics problem for the list... given two digipeaters 20
>miles apart, both at the same HAAT, with the same gain and power output,
>how large is the zone between them where the difference in amplitude
>between the two signals is less than 2 dB? In other words, how big is
>the dead zone where receivers won't decode either digipeated packet?
Ah, but you need to throw in a 6 to 20 dB multi-path fade margin in as
well. At any instant there can be that much fluctuation on either path.
Thus even perfectly inbetween, sometimes you will get one, sometiems the
other, and sometimes neither befcause they are within 10 dB and neither
captures a the FM discriminator..
This is why SOME dupes were considered as fundamentl to APRS, to improve
the local reliablity of delivery.
TO answer your question, if there was no such thing as multipath (it
exists in SPADES in APRS) then it would extend from 1/4th to 3/4ths of the
distance between them. Or about half the coverage area.
BUT Of the 6 to 20 dB of VHF multipath that exists EVERHWHERE, that means
that you can ADD or Subtract 6 to 20 dB to ANYEHERE inbetween the digis
and find out one thing:
1) there will always be some collisions.
2) There will always be some success from DIGI A just about everywhere
3) There will always be some success from DIGI B just about everywhere
So APRS was designed to allow FRATRICIDE kill out most dupes at the local
level, and still MOST of the time, even those in the KILL area will get
one OR THE OTHER of the packet. THus APRS was designed taking all of
these details into account and that is why some DUPES and SOME FRATRICIDE
are part of the design...
de WB4APR, Bob
----------------------------------------------------------------------
Subject: Re: Digipeater Duplicate Checking is broke
From: Bob Bruninga <bruninga@usna.edu>
Date: Tue, 30 Sep 2003 14:27:34 -0400 (EDT)
X-Message-Number: 26
On Tue, 30 Sep 2003 wes@johnston.net wrote:
>Although I like the idea of fracticide, I have found recently with one
>digipeater (ie one copy of each packet), if I have any flutter in my mobile, I
>don't decode a packet. At least with two digipeaters, I double my odds of
>getting a copy of a given packet.
Yes, that is why APRS was designed to assume some dupes. WIth out some
dupes, mobiles are going to miss a LOT of packets. SO it is th ebalance of
fratricide, some dupes and the 6 to 20 dB of multipath that were all
considered when we designed APRS... (Back before WIDEn-N or WWWWWWW
FLOODING killed the network)... ;-)
Bob
----------------------------------------------------------------------
Subject: Re: Take Control of your DIGI part 4-2
From: Bob Bruninga <bruninga@usna.edu>
Date: Tue, 30 Sep 2003 14:29:35 -0400 (EDT)
X-Message-Number: 27
On Mon, 29 Sep 2003, Tim Cunningham wrote:
>It takes three hops to cover North Alabama since the IGate is centered
>in the Northern part of the state.
I think it is better to solve that problem with a SECOND IGate than to
find more ways to improve the flooding of your system with MORE packets
from furthere away...
>If one needs to pass traffic to Fort Payne or the Shoals reliably, it is
>going to take 3 hops to get there. That the reality based on a cellular
>structure of of digipeters that do not cover a wide area intentionally.
Yep, but those in florence can easil use a WIDE,WIDE,HTSVL path and those
in Huntsville that need to QSO with those in FLorence can use a
WIDE,WIDE,MSC path.
[an example of 11 copies followed]
>
>You are not dealing with reality.
Ah, I AM! By saying wht we both agree is that we have a TOTALLY broken
network that we MUST fix. And my proposal to change the WIDEn-N paradigm
and get back to generic local callsign substitution is the ONLY way that
gains all the benefits and makes APRS work like it was originally
designed and the ONLY way it can handle the 20,000 users on a single 1200
baud channel. We CANNOT affor the FLOODING algirthm of the old WIDEn-N
concept.
[Bob said get rid of the non-substituting RELAY digi]
>Shame on yuo Bob. That RELAY could have been a perfectly functioning
>callsign substituting digipeater using a callsign in place of the RELAY and
>WIDE it injected and the results would have been no different. Do you
>agree with that statement?
ABSOLUTELY! But because we have NO CLUE where it is, we CANNOT go turn it
off! ANd it is the CAUSE of the FLOODING of dupes because it should not
be an INFILL RELAY digi if it is in an area where mobiles that hit it can
ALSO hit 3 surrounding digis as this example shows! This is why we must
get to CALLSIGN SUBSTITUTING RELAY's only.
Wow are we on different pages..
>Nothing is going to change in the example as you indicated with what
>was presented. The RELAY and the WIDE would simply become the
>callsign of the digipeater instead of RELAY and WIDE. We intend to
>turn off the WIDE at WC4ADG in the near future. It was activated to
>support traffic downtown in beteen the buildings and the recent hamfest.
Shooting yourself in the foot so that you can have IMPROVED WIDEn-N
flooding in your area, worse performance, poorer reliability and LOOSE all
the 7 or 8 advantages of switching to the NEW n-N paradigm...
Seems to be a failure to open your eyes to reality and realize that your
focus on local DUPES is causing you to ignroe all the other parts of the
problem.
>This only works in areas that have an Igate within 2 hops in most cases.
>Some areas may need the 3 hops.
Duh, then put in an IGate. That is far better for your network than
encourageing everyone to be sending ALL their traffic out with 3 hops and
letting it propogate with 21 to 36 copies (TIMES all the local dupes).
Your approach of keeping the old WIDEn-N as it currently is WILL NOT WORK
and you can see how it takes a local DUPE and then FLOODS the network
with the original and all the original dupes. Whereas, if you just
accepted the LOCAL dupe, and it would not go any further and you would
get orders of magnitude improvement in performance...
>>2) DX stations more than 2 hops away (the DOZENS of users surrounding
>>EACH of the 6 digis that are out in the 3rd tier) are no longer QRM'ing
>>your LAN!
>
>I agree, but 2 hops may not do the trick for everybody.
Ah, ANYONE can use 3 or more hops, but they just CANNOT use a 3 hop
FLOOD with impunity. THey should use at least one or more specific
CALLSIGNs in the longer path to ACCOMPLISHE the need, but NOT QRM the
other 21 or 36 surrouiding digis that are not part of the need.
>That is the problem. Educating all the users wit ha moviing target that
>is already moving again. When the network takes care of itself it will
>be a better system and only when it does not depend on the end use to
>take any action to make it work. That is a goal worthy of exploration.
Ah, but it is moving BACK to basics, and TO a system that takes care of
itself. Break the old WIDEn-N FLOOD EVERYONE EVERYWHER paradigm and use
my new n-N paradigms and you get all of the good and hardly any of the
bad.
And it can be DONE locally! You dont have to wait for everyone else, just
do it one digi at a time. Swapping the new n-N paradigm is a LOCAL option
that has all of the advantages locally. This is what makes this approach
so different from all the other LOCAL hacks that have been ateempted.
THey KILL the functionaity of APRS or require total cold-turkey universal
adoption (that will never happen in the anarchy of ham radio).
So like I said. "take control of your LAN"...
1) Enable RELAY/WIDE callsign substitution,
2) Change your WIDEn-N paradigm to LINKn-N or LANn-N or WIDEn-N in the
wilderness or OFF.
3) Educate users that the GENERIC hops are only encouraged for 2 hops.
Anything beyond 2 hops must include directed DIGI's or LANn-N or LINKn-N
or WIDEn-n(in those areas that can afford it). Yours can't...
>>Once you accept that premis, arguing over dupes is wasted energy. You
>>gain so much, in reliabilty, and lack of QRM, with full path tracing of
>>every packet, and redundant delivery to improve reliablity all come for
>>FREE...
>
>So, are you saying that we really no longer need the TRACE function
>since the packets will be traceable?
Yes. But the reason we can get rid of the UIDIG parameter of TRACE is
more because there will no longer be any WIDE-ONLY digis that do not do
callsign subsititution. Thus there is no need to distinguish between
WIDE and TRACE>..
>Two hops are not going to cut it when a packet must go to the extreme EAST
>and WEST boundaries in this part of the State. Why, they would require much
>taller digipeaters killing the concept of many more lower digipeaters.
1) THe vast majority of those people in the middle can use 2 hops and
cover a total of 4 hops worth of distance.
2) Those at the ends can use 3 hops to cover the entire area, but as long
as ONE or more those hops, either the first, or last is a single digi by
callsign, then you can use 3 hops just fine with only a few dupes. Not 21
to 36 like a true WIDE3-3 system would generate.
>OK, so you do agree that dupe checking is broken. Thus we will have
>to tolerate dupes until the digipeater firmware is corrected or they are
>replaced with systems that have better dupe checking.
Dupes are part of APRS. the only way to get a single packet to the 12
digis that are part of your north alabama system is to have 12 copies
digipeated. If you are high enough, you will see them ALL. But they are
not dupes. They are how APRS distributes packets.
The LONG propogation of multiple DUPES is a problem of "LONG propogation"
not of the dupes themsleves.. which DIE OUT at N-minus-1 hops. Thus if we
get back to 2 hop generic paths (RELAY,WIDE or WIDE,WIDE), then the one or
so dupes ONLY occurs whewre it originated, and does not go to any second
much less 3rd tier digi. It just is not that big a problem if we
fixeverything that is broke...
>Yes, they can implement callsign substitution, but dupe check is still
>going to be broken in most cases.
Agree, but if we stop FLOODING the network, then the dupes dont go
anywhere. and only improve local reliablilty in the immediate vicinity of
the starting point. No further...
>The callsign substitution will cause packet duplication in the current
>network. There is no disputimg the reality until the digipeters are
>upraded or replaced.
But continuing the mind-set that WIDEn-N is the SOLUTION is what is
killing APRS and dooming us to failure! As your examples have clearly
shown, APRS is BROKEN as a local, reliable real-time system.
Getting back to short 2 hop generic paths, making sure that callsign
substitution works on the two RELAY and WIDE generic paths, and changing
the WIDEn-N paradigm are the way out that can be done NOW, LOCALLY, and
with only one sysop at a time, finally seeing the light...
Problem is they what to see everything from hundreds of users over
thousdands of square miles on a single 1200 baud channel and that is
impossible. Thus APRS is broke in their areas....
>OK. It reminds me of the Bob Riley Tax Plan for Alabama. You vote for
>all 19 House and Senate bills in one package or you leave it... ;-)
>
>In the end we are not disagreeing at all.
Yep. I agree. I am saying we have lots to fix, but my approach can be
started by a single SYSOP, and most of the changes can all be done
locally until we fix it. That is what I hope will catch on. Becuse no
one has to wait for someone else to do something.
Now. about North Alabamna. You konw the topology well. I wonder if a
NALA3-3 (LAN3-3) system would work? Can you visualize X digis that would
consider themselves as part of the North Alabama LAN?
DRUMM ROLL!!! NEW IDEA!!!!
Instead of changing UIFLOOD to NALAn-N, change the UITRAFCE parameter to
NALA. This way, you can STOP all of the original DUPEs you are concernted
about (yes they are a pain) and now, since NALA3-3 automatically TRACES
every path ther is NO need to begin with the WIDE. Thus, you get all of
the advantages of my ideas with all of the advantages of NO DUPES!
Yes, packets carry up to 3 digis calls, but NO FARTHER, even if someone
uses NALA7-7, it will still only hit the X digipeatres ONCE. and STOP.
THus, they dont grow without bound... AND we get all the advantages of
injection point and path tracing!
kBob
----------------------------------------------------------------------
Read previous mail | Read next mail
| |
