Re: Controllerless distributed automation

["David Paterson (Pixi)" <david.paterson@xxxxxxx>]

On Wed, 14 Feb 2001 01:26:54 -0000, you wrote:

>This is exactly the problem with all such distributed systems: if
individu=
al=20
>components of it are to co-ordinate their behavior in order to operate
as =
a=20
>coherent "system" (rather than a collection of individual
control systems)=
,=20
>then what is required is actually quite a bit more than mere
communication=
=20
>between those components. - Each and every component needs to be aware
not=
=20
>only of the presence of other components, but also of the capabilities,
an=
d=20
>the state of them. This would mean that each component would require
local=
=20
>storage not just of it's own operating code, but also of the above=20
>information.

I don't think it needs to be that complex, as long as devices are aware
of _relevant_ neighbours and their status, they don't need (or want)
complete information.  Would the room stat need to know which TV channel
I'm watching?  (OK, having said that, somebody'll come up with a really
good reason why it might, but you get the idea :-)

At the moment I'm trying to figure out just what is useful information,
and what constitutes a "relevant neighbour" before even starting
to
worry about communications protocols or physical networking.  There's a
lot to think about....

> (snip)
>
>THAT is why you haven't seen such systems around, - it just plain
makes=20
>sense to have a central controller as it negates all the above
problems, i=
t=20
>keeps the intelligence in one place, keeps the management and
programming =
in=20
>one place, and thereby reduces the overall cost (the "dumber"
modules are=
=20
>obviously cheaper), eases setup, programming, management of the system
and=
=20
>future expansion.

Erm, I thought you were in favour of distributed systems, or are you
just playing devil's advocate? :-)  [cf the "Flight of fantasy"
thread]

>THIS is exactly what I proposed a month or two ago in the thread I
started=
=20
>called "flight of fantasy" - I even went as far as to
describe (in loose=20
>terms) how such elections could be coordinated and controlled, with=20
>heartbeats, assigned priorities and so on.

I missed/skipped this originally as things were incredibly busy at work
(lots of late nights) plus I was house hunting at the same time.  I've
gone back and read all the messages from then, and it looks like I
missed a lot of interesting discussion.  Unfortunately your doc isn't
available from the Yahoo archive - any chance you could fire a copy off
to me?

>This in my opinion, is actually the better "ideal" system -
any objections=
=20
>(which are valid ones, don't get me wrong!) to having a "single
point of=20
>failure" are negated by providing redundancy of the central
controller rol=
e.

The reliability argument is really a minor one.  As you rightly point
out, these systems are very reliable, and I'd actually expect my
home-brew bits and pieces to be worse as far as MTBF goes.  It's more of
an experiment/research idea really, just to see what might be possible.

> (snip)
>
>Just my =A30.02

Cheers!  Any comments are useful, especially from someone who's been
thinking about this for longer than I have.

DP

--=20
David_Paterson =3D david.paterson@xxxxxxx =A6 david.paterson@vissci.=
com;

R & D programmer        There are three kinds of people in the world -
Visual Sciences Ltd.      those who can count, and those who can't...