[aprssig] New Digi Settings
aprs at kd4rdb.com
Thu Jun 9 10:38:31 CDT 2005
Two issues here....
1)I've seen users run crazy paths because they either didn't know any
better, or they were locked out by the sysops (budlisted) due to abuses
they did not resolve. This in turn leads them to run really really
stupid paths. Case in point was a station in NC who was so aggrivating,
he ended up budlisted by all the local digipeaters. His response was to
run RELAY,RELAY,WIDE,RELAY at 160watts. On the one hand, this
"creativity" is a testament to how robust the aprs "p2p" network is! On
the other hand, the entire situation could have been avoided if the
NETWORK was smart enough to re-write the guy's path to something more
acceptable. The end result would have been that his packet would have
been seen as far away as prudent, and there would have been no feelings
hurt because he never got budlisted.... matter of fact, many many
newbies don't pay attention to paths at all, so would be blissfully
ignorant of the on-the-fly path changes that could have been made.
Pete's NSR system does just this.... it takes whatever path is found in
a packet, removes it and substitutes the callsign of the digi that heard
the packet first. All surrounding digipeaters that consider that first
digi's geographic area to be local to them will digipeat the packet.
His system even permits overlapping coverage areas! For example,
packets originating inside a city could be digipeated outward, while
packets originating outside the city would not be brought into the busy
city. He's even figured out how to use IGATES to still permit messages
into areas that normal position reports would be too much.
I did say in the above paragraph that "packet would have been seen as
far away as prudent".... One of my first objections to Pete's NSR system
at first was what if I don't want my packet to go out the full number of
hops permitted in the aloha circle... but when I really thought about
it, who really does this? For all but a few APRS operators, paths are
"set and forget".... and often times set higher than needed for those
occasional cases where you might need it. Heck, I'm a good example of
this.... I set my path to WIDE1-1,WIDE2-2 to cover my occasional trips
to the upstate of SC. Only yesterday did I realize that 3 hops got me
nearly to Tampa FL! With Pete's system, the NETWORK can adjust your
hops as needed.
The really slick part is that it can be implemented smoothly over
time.... the users don't have to do anything.... you can run whatever
path you wish in your car or your home... if you are in a normal aprs
network, the path is utilized. If you happen to wander into a NSR
network, the network will remove your path and route your packet over an
acceptable area (ie your aloha circle)
Just think, if we had NSR everywhere _today_ we wouldn't be screaming
about the long haul truckers running W4-4 W5-5 here on the east coast
for those cases when they are out west and need 4 hops! With NSR in
place, let 'em run crazy paths.... the digipeaters will politely fix the
paths and keep the stations 'in' the network.
While implementation of NSR can't be done on your run of the mill KPC3,
it can be done with digined and a cheap laptop hooked to your kpc3. I
suppose Kantronics could implement it in a firmware revision though....
it's simply a matter of removing the path from any packet heard direct
and substituting the digipeater's callsign. A list of "friendly"
digipeaters needs to be maintained so that when a packet comes in that
has been digipeated by a neighboring digi, it is digipeated. That's
really it in a nutshell.
2)forcing packets out an igate has been discussed before... the problem
is that inevitibly you'll have 10 people "pushing" packet out into some
large city like LA where the network is already saturated. I think a
better solution is to let someone in the area register (or subscribe)
with a local IGATE for certain data products. What would be slick is if
the tracker in my car could send this subscription message periodically,
say mixed in with my positions every 10 minutes. Upon hearing this
subscription request, the nearest IGATE to me would send my requested
data for the next 30 minutes. The end result of this is that the data I
want to see from a far away place would appear (to me) to follow me
around. From a fixed station's perspective, the data I requested would
appear for a short time and when I left town, my data would not clutter
the network any more. Also, subscribing to data "pulling" would allow
the person on the recieving end the option of stemming the flow of out
of area data if they felt the network was nearing saturation. IGATES
already keep dynamic lists of what they will and won't gate out to
RF.... they keep a list of stations that have been seen locally and gate
messages out to those stations as the messages come along. This
subscription thing would be no different than that except that users on
RF would be able to tell the igate other things (other than just
messages) they wanted to see on RF.... like the location of another car
in another country.
AE5PL Lists wrote:
>What "limitations in the current Internet backbone" do you think flexnet
>would fix? The problem with flexnet is that it is still source routing
>under a different name. Let me ask you a simple question regarding your
>IGate example: can you, in today's telephone and Internet networks,
>force a connection to a remote speaker phone telephone or to a remote
>computer if those remote devices do not answer you request? The answer
>is no. Yet, you think it is perfectly all right to force that remote
>IGate operator to gate your packets to RF just because you say so in
>your path? Again, this is source routing and should be eliminated
>_completely_ from the APRS network so that the APRS network can go to
>the next level and be a truly useful mode of communications instead of a
>just a toy for a few hams to play with.
>As I pointed out, the way you get there is to eliminate the need for any
>user to have to configure more than their callsign-SSID and how they get
>their position information into their software/hardware. From there,
>the network should be able to distribute that information without
>further interference from the user. The second you introduce any type
>of paths, including destination "areas", into the equation for a
>broadcast-type protocol, you will fail to achieve a stable and reliable
>network because it now depends on thousands of users getting it right
>(which won't happen nor has it ever happened).
>Pete Loveall AE5PL
>mailto:pete at ae5pl.net
>>From: Andre PE1RDW
>>Posted At: Thursday, June 09, 2005 9:08 AM
>>Subject: Re: [aprssig] New Digi Settings
>>As long as we are talking about "new" routing options, why
>>not throw in flexnet routing to overcome some limitations in
>>the current internet backbone.
>aprssig mailing list
>aprssig at lists.tapr.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the aprssig