[Asterisk-Users] DUNDi Not Able to HandleComplexFailoverSituations

[Patrick]

On Mon, 2006-06-26 at 15:33 -0700, Michael Collins wrote:
> > > I get annoyed Stephen when Digium goes around calling Asterisk
> > 'enterprise grade', which in my opinion it really isn't. I'd consider
> > distributed ACD queues to be a requirement for an enterprise grade
> > product, but it's becoming apparent that there is no mechanism for
> > implementing this. I'm being told that DUNDi isn't the right man for
> the
> > job.
> > 
> > I'd suggest you ask Digium for your money back.
> > 
> > Leif.
> 
> 
> Question: isn't there a bigger picture issue here?  I've seen a lot of
> bashing going on in this thread but not very much useful dialogue.
> (Doug bashing DUNDi and Digium, other people bashing Doug for his
> "annoying" posts, etc.)  
> 
> Whether or not we like or dislike Doug's tone is, IMHO, irrelevant.  How
> about we tackle the REAL questions: If DUNDi isn't the answer, what is?
> Is Asterisk even capable of doing what Doug needs, namely, distributed
> ACD queues?  If so, how?  If not, why?  Is it even feasible to try to do
> it?  Will it require an Asterisk add-on, or can the core be modified to
> do this? (This leads to the question for the dev list...)
> 
> These questions, of course, lead to other questions: If * can be
> programmed to do distributed ACD queues, does that mean there are other
> features that might benefit from a "distributed" model?  Etc., etc.  I'm
> just throwing out ideas because maybe one of these ideas can turned into
> a killer app, just like Asterisk itself.  Just think of the advantage
> you would have if you wanted to sell Asterisk against one of the big
> boys.  How much would a fully redundant, HA Asterisk system cost
> compared to the same thing by Cisco, Avaya, Nortel, NEC...  You get the
> idea.  
> 
> The moral of this post: a little good-natured bashing is just fine, but
> let's not lose sight of the ultimate goal, which is to keep making
> Asterisk a better product.

Nod. There have been many posts on the mailinglist about redundancy/HA
but I have never seen a detailed explanation how to create a fully
redundant HA Asterisk system. Would be nice if that recent DUNDi article
posted on the list were expanded by Digium and/or the community with
some of this magic. There's "clustering & scaling" mentioned at
http://edvina.net/training/schedule.shtm (day 5 at the end) so I guess
it's possible but you have to pay a few thousand euros to get the
info :)

Regards,
Patrick