Posted by
Phil Henshaw-2 on
Mar 30, 2008; 4:25pm
URL: http://friam.383.s1.nabble.com/Wed-Blender-Stereo-and-Computational-Photography-Videography-for-Cultural-Preservation-tp525972p525995.html
Marcus,
[ph]
>
> Phil Henshaw wrote:
> > I'm trying to compare the use central managed
> > solutions and user negotiated solutions in this fairly simple problem
> to
> > develop a way of discussing the more complicated situations where
> efficient
> > and fair central resource management is not possible. For lots of
> things
> > central control is going to work well and be naturally more
> efficient.
> >
> The "user negotiated solutions" reduce to the question of a shared values.
[ph] Well, not as I see it. That would be the central manager's assumption
for setting up rules that are to be made self-consistent, and excluding the
environment's inconsistencies from consideration. From an inclusive view,
though, the individual users will have different needs, interests,
information and perceptions and that's the general problem.
> When shared values can be identified amongst N people, then
> conceptually
> we can replace those N people with a single person that plans a larger
> array of work and then again it's a question of scheduling, load
> balancing, and optimization. Normally, though, people know what they
> want and are competing to get as much of it as possible.
[ph] Yes,... when shared values can be identified. In addition to
individuals not choosing to cooperate as you suggest, there are also lots of
times when the long established shared values become inappropriately agreed
to, because of a change of circumstance. One present glowing example is the
idea of agricultural resources being 'renewable' and the world's
environmentalists and governments committing themselves to a plan of
increasing non-renewable mining of 'renewables', permanently setting aside
growing areas of the surface of the earth to replace the growing energy uses
that used to be supplied from holes in the earth. It's then kind of
quintessential mistake that helps us examine the real blind spots in our
thinking about nature.
>
> When there are no shared values, than all that can be done is to
> dynamically divide the resource into a virtual resource and use the
> strengths of the resource to make up for the weaknesses in the
> resource. It's a design question, whose solution may be central or
> distributed in nature but still algorithmic. Virtualization can
> prevent hogging, although the resources will divide in power as more
> and more users draw upon it.
[ph] There are lots of things that virtualization might work for, and that's
a good way of saying it. It still requires the global "God's eye view" of
things that no one naturally has... though. Still, there are some things
for which one can set up useful controls based on information sharing if one
has run into a necessity of rationing due to not having seen some new
circumstance coming... That can be done either with central regulation, or
better with informing free markets about the truth of some long neglected
problem, where knowledge is more out-of-date than usual, so they can catch
up. An example is the fact that the economies are widely acting to
accelerate the depletion of under priced resources to maintain a growth of
output displays. That seems to be another kind of quintessential error that
helps us examine the real blind spots in our thinking.
> In contrast, a political solution requires
> trust (or at least policing). Without trust, there will be
> instabilities created when people pretend to have consensus in order to
> get preferred access and then soon defect on one another when it
> actually comes time to use the thing.
[ph] yes I agree. I think managing social and economic conflict with
political conflict is nearly a complete waste of time. Governments also
have a tendency of getting fed up with stalled negotiations and going to war
instead... I think it would be much better if they devoted their limited
resources to giving people better information and discovering comprehendible
rules that would tend to be self-enforcing.
> > Without a central operator different users connected to a bus would
> need
> > some way of telling what the load on it in the near future would be,
> in
> > order to be ready to use it when it wasn't busy.
> Sure, it's called a scheduler. For schedulers like are in the kernel
> of
> an average Windows/MacOS X/Linux system one resource is made to look
> like many, and during the period that a user sees their resource, there
> will tend to be minimal contention for resources, although waste may
> still occur due to mismatches in latency between the different
> components in the system (as would also occur in the non-virtualized
> case). There are some costs to time slicing, in particular that CPU
> caches have to be invalidated on context switches, but the idea is to
> run long tasks enough to amortize these costs.
[ph] The management of users on a controlled resource provides the users
with predictable windows of opportunity for using the resource. There are
some network schemes I've heard discussed that would let subscription users
have ownership of a certain amount of bandwidth that they could then offer
to the highest bidder. That makes the resource allocation more liquid and
still not 'managed'. That's something like the carbon "cap and trade" idea
that seems rather efficient for the big users.
It still does not answer the basic problem though. On a network one of the
signals of something wrong, and the system of balances going haywire, is
price wars where none were expected, or the promised bandwidth simply not
materializing. If you sell a certain amount of bandwidth and don't deliver,
then you have trouble. If you recall, I pointed out examples of regular
'fishtailing' behavior in the financial markets that is definitely not
supposed to ever occur. In natural systems it appears a variety of 'signs
of trouble' tend to grow exponentially for long periods of time before whole
systems are bought to a stand-still by them.
> For large scale compute environments, this is taken further to have a
> high level scheduler that operates on the time frame of days to
> months. Jobs run through such a system get the full machine for a
> long period of time without any friction. The policies for such a
> system are to some extent a subject for negotiation. In academic
> environments, it can be a matter of peer review and politics, i.e.
> people write a grants to get access to the queues. In commercial
> environments, scheduling can be market driven, or run at a fixed rate
> for CPU time / hour.
[ph] It's when core policies turn out to be mistaken that it is hardest to
perceive and adjust them as needed, since the whole negotiation model is
predicated on them. The world is discovering more and more core policies
that are fundamentally in error though, so we need to start looking for what
we should start looking for.
> > Would there be any way
> > for users to sense that other than to sense increase or decrease in
> > electrical load on the bus somehow? In open systems the usual way to
> tell
> > if something else is using a common resource is finding disturbances
> around
> > it, and signs of depletion in what's available for yourself. Bees
> might
> > skip flowers that have been recently visited, for example.
> >
> More in the philosophy of ethernet or software transactional memory --
> wait for a conflict to occur and then retry..
[ph] Yes, that's the usual solution, that natural systems are not in a big
hurry and seem to work well that way, and can just wait for their usual
opportunity to drift bye. That's not the trend in the design of the complex
industrial system changes for responding to the growing failures of our
complex industrial systems, though...
The basic circumstance in which we're operating seems to have changed and we
don't seem to know what kind of information we should be looking for to tell
us what to do. I think developing measures to indicate where conflict is
blowing up might indicate the location of lines of conflict being crossed
that we were unaware of. I'm thinking that might be better understood in
terms of the simplified model of users of a controlled system, and help
frame the discussion for the tougher problem for users of uncontrolled
systems.
Phil
>
> Marcus
>
> ============================================================
> FRIAM Applied Complexity Group listserv
> Meets Fridays 9a-11:30 at cafe at St. John's College
> lectures, archives, unsubscribe, maps at
http://www.friam.org