[ax25-layer2] IP Routing

[Christensen, Eric H]

I think it would be neat to do NSR but have the NSR variable based on the
instantaneous ALOHA circle and knowledge of the existing LAN.  If that could
be automated, then you would have a very powerful tool that wouldn't need a
lot of SYSOP interaction.

Eric KF4OTN

 
-----Original Message-----
From: ax25-layer2-bounces at lists.tapr.org
[mailto:ax25-layer2-bounces at lists.tapr.org] On Behalf Of Robert Bruninga
Sent: Tuesday, September 20, 2005 9:30 AM
To: ax25-layer2 at lists.tapr.org
Subject: RE: [ax25-layer2] IP Routing

>>> pete at ae5pl.net 09/19/05 11:44 PM >>>
>For instance, the NSR UI algorithm supports (and improves the 
>performance of) APRS.

That is the #1 problem I have with NSR, in that

1) Yes, it improves performance when "performance"
is defined by the sysop as "your packets can only go here"...

2) It totally breaks "performance" when the SYSOP is not around to make
instantaneous real-time tweaks in the presence of urgent changing, real-time
situations.

3) It totally breaks "performance" when the originator's immdeiate
communication need is to get to a point not previously defined by the SYSOP.

Now, I have no problem with someone choosing to use NSR, by choosing the
path "VIA NSR" and letting the all-knowing-all-seeing SYSOP decide where his
packets can go, but I strongly object if then the ability to use a path of
VIA DIGI1,DIGI2 is not then allowed in the protocol.

Notice I said "in the protocol".  It is the SYSOPS choice to disable
point-to-point for his users (let them fight it out),  but to go about with
the intent to break AX.25 to make it impossible for the end user to
determine his path under certain conditions, I will have serious issues
with.

de WB4APR, Bob

>yet it does not support digipeater usage as a layer
>3 router.   This is why a layer 2 RF architecture 
>that is a known quantity is so important to reliable and transparent 
>operation of higher levels, just like the Ethernet architecture is a 
>known quantity...

Yes, as long as a SYSOP is always available to always re-configure the rules
of the network each time someone
new turns on a radio...   Ham radio on RF in some areas
can be legislated to do what you want, but in many cases, such
do-it-this-way appraoches fail due to lack of responsiveness of the sysops
and render the network useless in some cases...

It is inate in the HAM spirit to want to have control over where their
communications go.  In some cases, the NSR approach is like giving a HAM a
microphone and saying " here, talk.  We will make sure you can talk to
someone and try to get your message to the right person.. trust us..."...

If the objective is to set up a PHONE system, yes that
is the right approach.   If the objective of NSR is to
set up a stable digital infrasctructure, I have no problem with that either,
but it must allow for exceptional use in the protocol if not in the
individual implementation too...

Bob, WB4APR



_______________________________________________
ax25-layer2 mailing list
ax25-layer2 at lists.tapr.org
https://www.tapr.org/cgi-bin/mailman/listinfo/ax25-layer2
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 3116 bytes
Desc: not available
Url : http://www.tapr.org/pipermail/ax25-layer2/attachments/20050920/272f2a36/attachment.bin