Subj : *RVIA
To   : Carol Shenkenberger
From : Carl Austin Bennett
Date : Thu Apr 19 2001 06:36 am

 CS> There is good and bad in this method.  The good is the net controls 
 CS> the entry directly.  The bad is that an NC may not have an accurate 
 CS> entry and it takes time to figure out the upper link of the net if 
 CS> the link is to a 'between site' but Dale's working on a software 
 CS> solution to link it all together.

Actually, I'd expect the NC-supplied nodelist information to be more reliable 
than the manually-prepared regional routing lists for a couple of reasons.

One is the potential for human error (avoided by automating the process) and 
the other is that the regional lists are in some cases dreadfully out of date.

 CS> Dale's first test version came out about 3 weeks ago. He didnt make 
 CS> it secret that he was working on it.  He asked a few people to assist 
 CS> in a non-backboned echo called ERN and I was one of them.  I accepted 
 CS> and am assisting with the RVIA/MRVIA keyset development to support 
 CS> it.  This will however take a long time to effect, possibly a year or 
 CS> more?

 CS> I'm hoping Carl will join the echo (he's been asked) as it needs some 
 CS> wide angled input as the solution Dale uses requires mail movers and 
 CS> *C's work in concert.  I'm apparently the only one in there with RC 
 CS> experience and a competing view from Carl would be *most* welcome to 
 CS> assist the process.

Unfortunately, I don't have REC background; I wonder if it would've made good 
sense to ask StephenM to join this one? If the routing info's already in the 
nodelist, a program such as Dale's could be of use to an REC and could almost 
automate things to the point where the regional route list hatches itself. :')

 CS> Oh, I still support the 'single list is best' idea but the folks in 
 CS> ERN havent beaten on me for it.

Different regions are approaching *RVIA proposals from different directions. 

While admittedly R12 has started using the tags relatively recently (as 
compared to R10 or R13) the approach here has been to use them to provide 
complete info for the region instead of implementing them just to deal with 
cases (like nodes with independent feed) that can't be flagged any other way.

We're not at the point yet where the nodelist even flags the default route for 
each region with a tag in the RIN section like:

,1,Eastern_Canada_REC,Montreal_QC,Stephen_Monteith,1-514-367-1505,9600,V34,CM, 
XA,IFT,IVM,ITX,IBN:juxtaposition.dynip.com,UREC,MRVIA:1:379/1 

Were at least this much available, *RVIA-aware software could automatically 
default any nets with missing info back to the default route for the region.

--- Msged/LNX TE 06
 * Origin: Pause-Cafe Kingston 613-549-5599 telnet:cafe.dyndns.org (1:249/116)

.