> . . .  gets  fixed  part  of  the Bonn  area  will  be  unreachable,
>DDAGMD11 will only  be accesible thru the  DDAGSI1-DDAGMD11 line, and
>LISTSERV distribution passing thru DBNGMD12 will be delayed.
 ---------------------                       ---------------
 
Knowing that Eric is running out of things to do ( :-) ), I had
an idea after reading Jose Maria's note to the LINKFAIL List.
 
A few weeks ago, POLYGRAF's link to BITNET was down for a couple of
weekends while the 3705 was on the operating table. Powering it
on was a real sight! LISTSERV's reader filled up and LISTSERV took
10 - 15 minutes to process ererything. That wasn't too awful, but
as in the case above, distributions were DELAYED ( :-( ).
 
** and now the idea **
 
Wouldn't it be great if the other LISTSERV's recognized ( after
a while ... ) that one LISTSERV was unreachable, and dynamically
modified the weights so that whatever could be kept away from
the "down" LISTSERV was kept away?
 
Peered lists could play a new role, handling the traffic for
the "down" peer ( although each would need to keep a complete
list of subscribers ).
 
I'll leave the idea here ... comments anyone?
 
 
Tony