| |
ZL3AI > APRDIG 28.09.03 12:24l 244 Lines 13632 Bytes #999 (0) @ WW
BID : 2007-ZL3AI
Read: GUEST
Subj: TAPR Digest, Sep 25, 7/11
Path: DB0FHN<DB0RGB<DB0FSG<OE7XLR<OE9XPI<DB0CZ<DB0GV<DB0SIF<DB0FHK<DB0ACC<
ON0RAT<ON0DXC<WA7V<WB0TAX<LZ3NP<YV6BTF<EA5AKC<KD4GBA<ZL2BAU<ZL3VML
Sent: 030927/1357Z @:ZL3VML.#80.NZL.OC #:9142 [Chch-NZ] FBB7.00i $:2007-ZL3AI
From: ZL3AI@ZL3VML.#80.NZL.OC
To : APRDIG@WW
Subject: I guess we're going to agree to disagree Bob. Part 2 of ?
From: "Jeff Thomas - KD4EVB" <kd4evb@hotmail.com>
Date: Thu, 25 Sep 2003 17:29:28 -0400
X-Message-Number: 29
AND ONWARD WE GO.....
"And that is with only a path of WIDE5-5 with all WIDE turned off."
AND WHERE THIS WIDE5-5 COMES FROM, I HAVE NO CLUE. A PURE WIDEN-N PACKET
THRU OUR NETWORK DOES NOT DUPE, I'VE WATCHED THE FLOW OF DATA RIGHT IN FRONT
OF MY FACE ON THE SCREEN LOOKING AT THE CHANNEL FROM THE PERSPECTIVE OF THE
DIGI'S WHERE DUPING OCCURS AND IS BEST SEEN. I KNOW THIS FILE IS OUT OF
DATE, BUT I THINK YOU WROTE IT, ADMITTEDLY IT'S DATED BUT THE DUPE CHECKING
FUNDAMENTALS HAVEN'T CHANGED MUCH IF AT ALL, I'D SUGGEST READING THE WIDEN
PORTION OF IT AGAIN:
http://www.aprs.net/vm/DOS/DIGIS.HTM
BEST I CAN RECALL, FOR THE FIRST 6-7 MONTHS OF THIS YEAR I RAN BOTH OF THE
NKY DIGI'S WITHOUT WIDE ON AND A GOOD PORTION OF SEVERAL SMALLER TIME CHUNKS
DURING THE PREVIOUS FALL/WINTER AS I TESTED IT. I DIDN'T JUST TYPE A
OPINION UP AND SAY IT SOLVES THE PROBLEM, I TESTED MY THEORY AND FOUND IT
HELPED 'PATCH' THE PROBLEM LOCALLY FOR THE TIME BEING. I'VE REPEATEDLY
ADMITTED THAT WHAT I'VE DONE AND ABOUT TO START DOING AGAIN IS NOT THE
ULTIMATE SOLUTION, NEVER PORTRAYED IT THAT WAY. THE LAST COUPLE MONTHS I
OPENED THINGS BACK UP AND THEN THE CONGESTION GOT BAD ALMOST INSTANTLY.
WITHIN A FEW HOURS I HAD ONE USER ASKING WHAT I CHANGED AND BY THE END OF
THE FOLLOWING WEEKEND MY NAME WAS BEGINNING TO BECOME 'MUD' FOR TURNING WIDE
BACK ON. I'LL TRUST MY REAL WORLD OBSERVATIONS OVER UNSUBSTANTIATED
OPINIONS OF WHAT SOMEONE SOMEWERE ELSE PERCEIVES AS THE 'RIGHT WAY' TO DO
THINGS AND TEXT BOOK THEORIES ANYDAY. I'VE NEVER ENDORSED ANY SYSOP MAKING
DRASTIC CHANGES TO THE APRS DIGI INFRASTRUCTURRE WITHOUT CAREFULL STUDY AND
CONSIDERATION OF THE IMPACT. I REALIZE I CAN MAKE EVERYONE HAPPY, SO I
CHOOSE TO MAKE THE MAJORITY HAPPY BY DOING WHAT SEEMS TO WORK FOR MY
SITUATION.
"Now if APRS users in that area would realize that EVERYONE FLOODING the
network across 4 states cannot possibly handle the 300 users that are
trying to do that, then you might make some progress.?"
AMEN BROTHER, COULDN'T AGREE MORE BUT WITH ONE EXCEPTION. IT'S NOT THE FEW
WIDEN IDIOTS CAUSING THE PROBLEM DOWN HERE. IT'S THE BIG PERCENTAGE OF
FOLKS RUNNING THE OLD DEFAULTS OR WORSE. I DON'T CONDEM WIDEN BECAUSE A
HANDFUL OF FOLKS (AT MOST AROUND HERE) ARE ABUSING IT, WIDEN IS A MUCH MORE
POWERFUL DIGI PATH WHEN DONE RIGHT. YEAH, 'DONE RIGHT' IS THE KEY WITH
ANYTHING. WHAT ARE WE GOING TO DO WHEN THIS HITS SAY 600 USERS? HUMM, MY
POINT EXACTLY WHY WE NEED TO USE WIDEN RIGHT TILL WE COME UP WITH A BETTER
WAY AND BETTER PROTOCOL.
I suggest that you transmit the ABOVE packet using only the path of
WIDE,WIDE and then count the result. In the SAME network where every digi
can hear 3 others, you will not see 94 DUPES, but only 4. THese four will
reach everyone within 2 digis of the SENDER and in most densly populated
areas that is enough to support the ALOHA circle of the 50 nearest
stations. In rural areas, try WIDE,WIDE,WIDE. In this same network,
the total number of packets will be 10. Now you are out 3 hops in ALL
direfctions and it is still only 10% of the QRM generated by the original
path.
SO YOUR SAYING WE SHOULD TOLERATE A FEW DUPES? I'M SAYING THAT WITH WIDEN
DIGIS AND FOLKS USING GOOD WIDEN PATHS WE DON'T HAVE A DUPE ISSUE AT ALL
EXCEPT UNDER THE RAREST OF CIRCUMSTANCES OR MISCONFIGURATION OF THE UIFLOOD
PARAMETERS. WHICH DO YOU WANT? I CHOOSE NO DUPES, WIDEN. NOT ROCKET
SCIENCE FOLKS. DOES THIS MEAN WIDEN IS BEST EVERYWHERE IN THE US, NOPE BUT
PROBABLY IN MOST AREAS. I'M SURE THERE'S STILL SOME OLD WIDE/TRACE/RELAY
ONLY DIGI'S OUT THERE THAT DON'T SUPPORT WIDEN, THEY PRETTY RARE AROUND HERE
THOUGH. FOR THOSE STUCK IN A 'LEGACY' DIGI AREA (NOTHING DEROGATORY ABOUT
IT, JUST A TERM WE USE FOR OLD WIDE/TRACE/RELAY ONLY DIGIS AROUND HERE),
WITH A LOT OF CHANNEL CONGESTION...I FEEL FOR YA :-) I'M SURE THAT IN SOME
OF THE MORE RURAL LESS CROWDED AREAS WITH FAIRLY PREDICATABLE PROPAGATION ON
2M AND GOOD DIGI LOCATIONS, LEGACY PATHS & DIGI'S PROBABLY DO JUST
FINE....FOR NOW. JUST NOT THE CASE IN THIS AREA. AT LEAST ON THE COASTAL
AREAS YOU GET THE PERKS OF HALF YOUR COVERAGE LIKELY BEING WATER, GRIN. MY
PHILOSOPHY IS ALL YOU NEED IS WIDEN-N WITH N= THE NUMBER OF HOPS TO RELIABLY
HIT YOUR LOCAL IGATE IS ALL THAT YOU NEED, MUCH MORE THAN THAT IS EXCESS
REGARDLESS WHETHER WE'RE TALKING RELAY,WIDE,WIDE, OR WIDE3-3 OR WHATEVER.
OF COURSE IF YOU DON'T HAVE WIDEN DIGI'S AROUND YOU, THEN THAT'S OBVIOUSLY
NOT A GOOD PATH.
"People are just not thinking clearly. The problem is TOO much traffic
caused by TOO LONG HOPS. A 1200 baud APRS network on RF whether it is
from 10,000 foot m ountains or lots of small 100' digis, can support no
more than about 50 users reliably. People must start thinking CELLULAR
and reduce the size of their HOPS to match that 50 user radius. Anything
else is just fuzzy thinking..."
AMEN BOB ON THE CLEAR THINKING. AMEN AMEN ON THE TOO MUCH TRAFFIC, TOO LONG
HOPS, BUT I'LL ADD TO THAT THE 'DUPE FACTOR' AS I'M NICKNAMING IT. AROUND
HERE FROM MY ON THE AIR MONITORING, THE FREQUENT DELUGE OF THE DIGI PATHS
WITH DUPING PACKETS IS THE PROBLEM, NOT A THE 2-3 GUYS WITH SAY WIDE5-5,
THEIR IMPACT IS MINIMAL COMPARED TO THE OTHER ISSUES. YOU TAKE ALL THESE
ISSUES PUT TOGETHER AND YOU GET A BIGGER PROBLEM AS A WHOLE. JUST BACKED MY
DIGI'S THEMSELVES BACK DOWN TO WIDE2-2 A COUPLE NIGHTS AGO MYSELF. DO US
ALL GOOD TO THINK ABOUT OUR OWN SETUPS, DO WE NEED THAT MANY HOPS, DO I NEED
TO TRASMIT THE SAME POSITION THAT OFTEN. EMERGENCIES AND SPECIAL EVENTS ARE
CLEARLY A DIFFERENT BALLGAME, I'M TALKING OUR DAY TO DAY NORMAL OPERATION.
DO YOU NEED TO BEACON YOUR TRACKER EVERY MINUTE NORMALLY, HEY EVEN AT 60MPH
ON A STRAIGHT LINE INTERSTATE PATH...YOU'VE ONLY GONE A MILE ON THE MAP, DO
YOU NEED BREADCRUMB TRACKING DETAIL OR DO YOU JUST NEED TO TRACK GENERAL
LOCATION? I'D VENTURE TO SAY, BUT HAVE NOTHING TO BACK IT UP YET, THAT TOO
HIGH OF A TX RATE OF LOCALS IS A BIGGER ISSUE THAN SOMEONE GOING A COUPLE
HOPS TO FAR OUT BUT WITH A REASONABLE TX RATE.
PUSHING THE PTT BUTTON BEFORE BRAIN IS ENGAGED COULD BE HARMFUL TO THE
CHANNEL CHUCKLE.
ON TO PART 3, YAWN....
----------------------------------------------------------------------
Subject: I guess we're going to agree to disagree Bob. Part 3 of hopefully 3
From: "Jeff Thomas - KD4EVB" <kd4evb@hotmail.com>
Date: Thu, 25 Sep 2003 18:23:57 -0400
X-Message-Number: 30
Yes, and the problem is NOT the RELAY,WIDE, it is the fact that the packet
is FLOODING the network 5 hops in ALL DIRECTIONS!
NO ARGUEMENT THERE IF YOU ONLY NEED TWO HOPS, WHICH MAY VERY WELL DO IN MANY
CASES. RELAY,WIDE,WIDE OR RELAY,WIDE,WHATEVER OR WORSE IS A GOOD CHUNK OF
THE PROBLEM AROUND HERE.
Yes, that was back in 1994. But everything changed, and I cannot believe
it took me till last week to realize that WIDEn-N was now the number one
problem in most areas.
THEN I HAVE TO WONDER IF IT'S REALLY THE PROBLEM IN YOUR AREA, IT SURE ISN'T
AROUND HERE AND FROM THE RESPONSES I'M GETTING, NOT IN MOST OTHER AREAS.
MAYBE YOU NEED TO PONDER IT LONGER, IF WIDEN WAS THAT BAD IT SHOULD OF BEEN
NOTICED A LONG TIME AGO. DON'T CONDEM A GOOD PROTOCOL THAT NEEDS SOME WORK
BUT A SMALL FRACTION OF THE USERS ARE ABUSING IT. LIKEWISE I'M AGAINST THE
OLDER PROTOCOLS CAUSE THE VAST MAJORITY OF USERS ARE USING IT BECAUSE IT'S
THE DEFAULTS CHOOSEN YEARS AGO AND DON'T KNOW ANY BETTER, NOT TO MENTION
FROM DUPE PREVENTION PERSPECTIVE, WIDEN IS CLEARLY SUPERIOR. I GUESS THE
QUESTION IS WHAT ARE YOU GOING TO TOLERATE, A FEW DX STATIONS ABUSING
WIDEN-N OR A PILE OF FOLKS CREATING COLLISION AND DUPE CITY WITH THE OLDER
PATHS? THE CHOISE IS ULTIMATELY THAT SIMPLE AND I CHOOSE TO LIVE WITH A FEW
DX WIDEN PACKETS TO GET AWAY FROM THE DUPE/LOOP MESS. CHOICE IS A WONDERFUL
THING, BUT OFTEN A COMPROMISE.
And I can assure you, that turning offg WIDE does NOTHING to solve that
problem.
WITH THE RELAY,WIDE,WIDE3-3 ISSUE OR RELAY,WIDE,WIDE, TURNING OFF WIDE DROPS
THIS BAD PATH PACKET TO IT'S KNEES, RELAY GETS DIGIED THEN THE REST IS
USELESS IF THERE IS NOT ANOTHER WIDE DIGI THAT CAN HERE IT. EVEN THEN IF
THERE IS DUPES OUTSIDE THE 'NO WIDE' ZONE OF DIGIS, THEY ARE NOT BOTHERING
US, GRIN. AT WORST, WIDE OFF HELPS REDUCE THE DUPING CAUSED BY THE WIDE
PATHS UNCONTROLLED DIGIPEATING BY EVERY DARN DIGI THAT HEARS IT, REGARDLESS
WHETHER OR NOT THEY JUST TRANSMITTED THE SAME PACKET A FEW SECONDS AGO AND
THEY HEAR IT AGAIN OFF ANOTHER DIGI A HOP OR TWO OR THREE AWAY. I'M NOT
GOING TO BACK DOWN ON THIS STANCE, I'VE SEEN IT WITH MY OWN EYES.
Trying to bring in DX packets from hundreds of miles away to fill up
static maps with no one moving is just not what APRS is about.
COULDN'T AGREE MORE, YET THAT'S WHAT A LOT OF FOLKS THINK IT IS. WHY WE
NEED A MORE ROBUST DIGIPEATING PROTOCOL AND FILTERING/LIMITING OPTIONS AT
THE DIGI LEVEL, THEN THEY EITHER COMPLY WITHIN REASONABLE LIMITS OR HIT
BRICK WALLS FOR BEING INCONSIDERATE. IF THEY WANT TO PLAY, NOTHING SAYS
THEY CAN'T PUT UP THEIR OWN DIGIS ON THEIR OWN, LETS CALL IT DX-APRS..THEN
THEY COULD JUST PLAY TILL THEIR HEART IS CONTENT.
WIDEn-N works well where the path length is greater than 2 or larger.
But that changed in 1997 when Steve Dimse and Sprouls briught in the
internet. We jst cannot afford LONG paths in BUSY areas. No matter what
you do. SO get back to basics. Turn off the long paths.
Since we have not been able to do it through user education, then it is
time for local SYSOPS to take control over their local area and TURN OFF
WIDEn_N if it will h elp them reduce the traffic load on their local digi
so that it can serve its LOCAL area...
WORKS PRETTY GOOD AT WIDE1-1 OR WIDE2-2 AROUND HERE. RELAY,WIDE2-2 MAKES A
WONDERFUL MOBILE PATH AROUND HERE. LONG PATHS IN BUSY AREAS, I COULDN'T
AGREE MORE. MAYBE WE CAN HACK THE KANTRONICS EPROM TO CHANGE MAX WIDE7-7 TO
SAY WIDE3-3. I'M SORRY, USER EDUCATION IS GOOD BUT IT'S JUST PART OF A
NEEDED SOLUTION, SOME FOLKS WILL NEVER LEARN AND THAT IS WHERE FILTERING AND
PATH LIMITING AT THE DIGI LEVEL WOULD CURE THAT ISSUE REAL FAST OR EVEN IF
THEY WANT TO TRY WIDE7-7 IT WILL NOT MATTER. I'D ARGUE THAT KILLING WIDEN
ISN'T GOING TO HELP, NOT EVEN FACTORING IN THE AMOUNT OF EFFORT, TIME, AND
MONEY SPENT BY THE MULTITUDE OUT THERE TO GET THESE WIDEN/TRACEN COMPLIANT
NETWORKS GOING OVER THE LAST FEW YEARS.
I hate to say this, but if you would listen to me, you would have a
single point of leadership and you would have the right answers.
NOTHING PERSONAL BOB, BUT WHO IS TO SAY ONE GUY SETTING ON THE EAST COAST
SHOULD BE THE ULTIMATE SAY SO IN WHAT IS RIGHT OR WRONG FOR THE WHOLE
COUNTRY? I'D PREFER IT BE DONE BY A NATIONAL GROUP, FEEDING STANDARDS AND
RECOMMENDATIONS TO THE LOCAL/REGIONAL GROUPS TO BUILD AGAINST AND MODIFY TO
MEET LOCAL NEEDS. YOUR NEEDS & ISSUES ON THE EAST COAST ARE LIKELY
DIFFERENT COMPARED TO MINE HERE IN THE MIDWEST/SOUTHEAST COMPARED TO THE
GUYS OUT IN THE PLAINS, COMPARED TO THE GUYS OUT IN THE MOUNTAINS, COMPARED
TO THE GUYS ON THE WEST COAST. SORRY, ONE GUY, ANY GUY, IN CHARGE OF APRS
AS A WHOLE SCARES THE CRAP OUT OF ME JUST TO THINK ABOUT IT. IT'S NOT AN
AUTHORITY ISSUE, IT'S A GROUP CONSENSUS ISSUE. NEVER GOING TO GET EVERYONE
TO AGREE TO A SINGLE STANDARD, BUT WE CAN COME UP WITH BETTER BASELINES, SET
GOALS, AND ADAPT LOCALLY FOR THE BENIFIT OF THE MAJORITY.
Yep, its easy. Either get the users to cut back their paths (has been
shown to never work) or turn off WIDEn-N at the digis that are
overloaded.
IN THE REAL WORLD, NEITHER IS AS PRACTICAL AS IT SOUNDS ON PAPER. AS YOU
SAID, TURNING THE DIGI OFF ALSO HELPS CUT DOWN ON CHANNEL LOADING. A BETTER
PROTOCOL AND MAYBE USING MORE THAN ONE FREQUENCY ARE GOING TO BE THE ONLY
LONG TERM SOLUTION, DANCE AROUND IT ALL YOU WANT. THAT'S MY NOT SO HUMBLE
OPINION AND CLEARLY THE CONSENSUS OF MANY IN MY QUICKLY FILLING INBOX, WOW
WHAT A HORNETS NEST I'VE STIRRED UP. GOT ABOUT HALF A DOZEN EMAIL
CONVERSATIONS GOING ON HERE, MY WEAK MIND CAN'T HANDLE IT, GRIN.
FOR ANYONE COMING THRU NORTHERN KY AFTER OCT 1ST, MAY I SUGGEST YOU RUN
RELAY,WIDEN-N OR WIDEN-N, AS WIDE WILL BE OFF ON THE NKY DIGIS. IF SOMEONE
HAS A BETTER, PRACTICAL, AND PROVEN SOLUTION, I'M ALL EARS, BUT I DON'T CARE
ABOUT OPINIONS THAT ARE ONLY THAT, BACK IT UP WITH ON AIR PROOF. IF I'M
WRONG, I'LL GLADLY EAT MY SHOE ON THIS LIST, I DON'T EXPECT TO BE NEEDING
NEW SHOES THOUGH AFTER WATCHING THE OPINIONS OF THE LAST TWO YEARS ON HERE,
GRIN.
LOT OF GOOD FEEDBACK AND IDEAS COMING IN. RATHER THAN PLUG UP THIS LIST
WITH ANYMORE JEFF VS BOB OR MY RANTS, WHEN I GET TIME TO GATHER IT UP AND DO
IT, I'LL PUT THE BEST STUFF ON THIS WEBSITE. I'LL ALSO BE POSTING SEVERAL
HOURS WORTH OF ACTUAL MONITOR LOGS FROM ONE OF OUR DIGI'S, SOME SCREENSHOTS,
AND MAYBE EVEN A MPEG MOVIE OR TWO OF ACTUAL PACKET FLOW DOWN THIS WAY.
I'LL BE OFFERING A WHOLE LOT MORE THAN JUST EASY TO STATE OPINIONS, WE GOT
SOME TALENT AMONGST US AND I'M SURE WE CAN ALL LEARN A SOME THINGS AND OFFER
DIFFERENT PERSPECTIVES....MY INBOX IS PROOF OF THAT....HOLY COW! THE PROOF
IS IN THE PUDDING, I'LL BE PUTTING SOME PUDDING UP ON THE WEBSITE BEFORE
LONG. KEEP THE IDEAS ROLLING GUYS, WE'VE GOT TO GO FORWARD INTO THE
FUTURE, NOT BACKWARDS!
www.qsl.net/kd4evb
73 DE JEFF, KD4EVB. MAY THE LIST SERVER HAVE A BREAK!
----------------------------------------------------------------------
Read previous mail | Read next mail
| |
