Hello Andrew!

Mar 17 00:56 2015, andrew clarke wrote to Kai Richter:

 ac> It can be days (if not weeks) before you're reconnected to the echos 
 ac> you were connected to previously, and that's if you can find a 
 ac> reliable system that has all the echos you want. And that's only 
 ac> after you've noticed your upstream mail provider has gone down, then 
 ac> waited a few days for it to come back online (only for it to just 
 ac> disappear forever).

This only applies if you got a single echo uplink. You could set up a
second  stand-by uplink for a fast switch-over or build echomail
distribution rings, i.e. each node got a left and a right uplink. The
latter requires a dupe-proof system, since the ring produces dupes to
provide redundancy. In R24 we're running rings for echomail distribution
for quite a while and it really works fine. When one of my neighbours (a
major system) crashed during the sysop's vacation I still received R24
echomail. Of course we could go for a fully meshed distribution, but the
ring is a nice way to reduce the number of links which brings several
benefits. BTW, it would be nice if we could set up echomail rings with
other regions and zones too. Unfortunately some still don't get the idea.
And hpt has some nice features to deal with complex echomail distribution
including zone gating now ;) 

Regards,
Markus

---