Order Tray | Contact Us | Home | SIG Lists

[aprssig] Bob, Address the WIDE7-6 Possibility

Robert Bruninga bruninga at usna.edu
Tue Feb 22 13:58:33 UTC 2005


I think that once we get everyone's attention that only
WIDEn-N should be used with a small value of N,
or SSn-N for in-state traffic,   It will take several months,
and give them all time to adjust...

I dont think we will have any large N problems... and
the ones that remain, we can address those problems
individually....  I have already seen a big change to
WIDE2-2 in our over-saturated Maryland.  Down in
NC I woiuld think that WIDE3-3 or NC3-3 would be
fine in most cases..?

Bob


>>> "Christensen, Eric" <CHRISTENSENE at MAIL.ECU.EDU> 2/21/05 3:08:09 PM
>>>
Bob, 
Respectfully I disagree with your statement that our problems were not
with large WIDEn-N users.  We saw a SIGNIFICANT improvement in our
network when we removed WIDEn-N from our digipeaters.  I think we
will,
again, see an improvement when we remove WIDE and TRACE as well.

There is a fundamental problem with your trapping method.  I'll use
the
example of two digipeaters, Alpha and Bravo, setup as you say in your
plan.  Alpha is trapping WIDE7-7, WIDE6-6, and WIDE5-5.  Bravo is
trapping WIDE4-4 and WIDE3-3.  (Remember that you can't trap ALL of
the
problem packets in one digipeater because of the limitation in the
KPC-3s.)  So a packet that is WIDE4-4 comes through Alpha and is
propagated as WIDE4-3 and is now free to spam away.  Bravo will not
stop
it because it heard it as WIDE4-3 not WIDE4-4.  Even better...  Bravo
hears a WIDE7-7 and propagates it as WIDE7-6.  Now Alpha will digipeat
that one too because it heard it as WIDE7-6 not WIDE7-7.  So the
trapping hasn't really worked...  Now we get the guy that doesn't
believe in all this stuff and still believes he needs his position
seen
in TX from FL.  So he runs WIDE7-6 and circumvents the entire trapping
solution.  These examples are the reasons your plan doesn't work.  At
best it is a band aid for a problem that needs more attention than
just
that.

I stand by removing WIDEn-N, WIDE, and TRACE (and TRACEn-N) from all
digipeaters in the problem areas.  If you are a frequent traveler I
suggest changing you path as you should when entering a new area.  I
do
as do other ops.  As much as everyone thinks there is a universal path
(other than RELAY), there just isn't.  And we haven't created any kind
of new education system that would make people think.  I actually
don't
think education works in most cases.  There are too many people out
there that just don't care or don't care to listen.  Yes, there are
folks out there that do listen but too many don't.  The only reason
there are folks out in my area that are listening now is because we
got
their attention because their equipment didn't work.  That is a bad
thing to have to do but it was necessary, unfortunately.


Eric KF4OTN


> -----Original Message-----
> From: Robert Bruninga [mailto:bruninga at usna.edu] 
> Sent: Monday, February 21, 2005 08:59
> To: kf4otn at earthlink.net; aprssig at lists.tapr.org 
> Subject: Re: [aprssig] Bob, Address the WIDE7-6 Possibility
> 
> 
> >>> Eric Christensen <kf4otn at earthlink.net> 02/20/05 3:23 PM >>>
> > I wish you would address the possibility of a user... path
> > of WIDE7-6 to get through our filters...
> 
> There are several important aspects to the New n-N Paradigm:
> 1) Its a very big change from the past.  The last 13 years of
>     evolution had lots of legacy settings that not only are no
>     longer needed but in fact were holding us back from a better
>     network.
> 2) The biggest QRM problem it turns out in the East coast
>     area was not large N's from the few abusers, but from the
>     70% of us that were still using the multiple-dupe-generating
>     paths of RELAY,WIDE and WIDE,WIDE or WIDE,WIDE,WIDE
> 3) Now users are told one thing:  Use WIDEn-N with-a-small-value-
>     of-N for your area.  In NC that should probably be 3-3 in 
>     general.
> 4) So bottom line, is that users have now ONE receommendation.
> 5) DIGIS have a *consistent* way of operating
> 6) DIGIS have a consistent display (R,W3,NCn) alerting all users
>     in a consistent abbreviation set as to what path systems are
>     supported.  In this case RELAY, WIDEn-N up to 3 and NCn-N
> 
> Yes, for a while there may be some people who individually 
> may try to circumvent the system.  But they are not going to 
> significantly impact the big performance improvement of this 
> system when everyone else gets on board.  Give them all a 
> chance to learn the new simpler and better way of operating.
> 
> ALso a bulletin should be floating around in your are with 
> the recommended path.  SOmethingn like "WIDE3-3 or NC3-3 is 
> best in this area."  And this does not mean you should go 
> after the occasional western trucker that drives through with 
> WIDE4-4 since he wont be around long...
> 
> >messages to this SIG that complained because when the
> >SSn-N LANs came into play they could no longer send messages 
> >to far away locations...
> 
> Then they need a local IGate.  Very simple solution. or they 
> can specify the route as in DIGII1, DIGI2, DIGI3..  So we 
> need to educate the users.  Also any message further than 
> about 3 hops is a complete waste of time.  The probabliitites 
> are so low as to be useless...
> 
> >..if we put WIDEn-N back in... they would be the ones that
> > would circumvent our filters and reduce the reliability of 
> > our LANs. 
> 
> Yes, but they are a vast minority.  Give them time to adjust 
> and they will probably come around.  Again, in the past 13 
> years we had no consistent set of education that would guide 
> users.  Now we do.  And because it is consistent, it is 
> better for everyone to see what is "normal" and what is not...
> 
> >Your plan will not work to prevent spamming of our LANs;
> >it only makes offenders become more creative.
> 
> I dont think they are intenionally spamming.  I think it is 
> because they dont understand the network and have not
> been informed of the New n-N Paradigm.  And their software
> does not alert them of the damage they may be causing...
> 
> It will take a while, but I think most hams want to do the 
> right thing, as long as there is a consistent and easy way
> to tell them what the right thing is.   WIDEn-N and SSn-N
> I think will be great tools.  The way I see it, WIDEn-N is
> for small values of N.  If one wants to use N one step
> larger, then he should use SSn-N. etc...
> 
> de WB4APR< Bob
> 




More information about the aprssig mailing list