My group has been given the task of developing applications that remain in
constant synchronization with other instances of themselves running at
remote locations (very remote locations on a WAN, e.g., some in Florida,
some in Washington state). For example, Process A and Process B know what
Process C is doing, so that if Process C dies, Process A and/or Process B
can pick up *exactly* where Process C left off. When Process C comes back
on line, it will get back in touch with A and B and all with get back in
synch. There are potentially millions of "jobs" to do each day and none can
be missed. It would be great if none were done more than once, but that is
not a strict requirement, and may not be possible.
Use of message queues, replicated databases, other third party solutions,
and home-grown solutions are all options.
I realize that this is probably the wrong newsgroup for this post, but I
don't know where to start with this. Flame me if you must, or tell me this
is science fiction, or point me in the right direction if you can. Any help
would be appreciated!