NNSquad - Network Neutrality Squad

NNSquad Home Page

NNSquad Mailing List Information

 


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[ NNSquad ] Smart Grid vs Copernicus


Responding to http://www.nnsquad.org/archives/nnsquad/msg03865.html. Should
you do smart grid control over a shared network (or, to be precise, public
infrastructure)?

Of course. But first you need to understand best efforts and how do we
maintain edge-to-edge relationships despite (or taking advantage of) all
that stuff in the middle. This isn't to say that all control of the flow of
electrons can be done without regard to the particular characteristics of
flows of physical things like water or electrons. I'm not trying to
second-guess power engineering decisions.

But it seems to be that the smart grid advocates are taking the lock-step
control model used to keep generators in perfect harmony and extending it
inside your home by telling you when you can wash your dishes. One big
lesson of the Internet (and Y2K) is that this kind of rigid linkage is not
the norm in the real world. Today's homes are loosely coupled so why would
we create risk by creating a brittle dependency on network characteristics?

The idea of treating smart grid bits as something special touches upon a
deep philosophical issue that is similar to the idea of the ".xxx" domain.
It assumes that some bits are intrinsically metering bits and other bits are
intrinsically porn bits. I know it sounds silly when I state it this way but
this is the assumption driving so much of our policymaking. Silos work well
for politics by treating each problem as a campaign issue in the context a
sound bite.

The smart grid gives us a real example of the divide between the world of
"electrons" in which you build lighting systems atop a power grid (either an
AC world or a DC world) and the world of bits in which meaning is decoupled
from the layers.

We see this thinking in the flawed IEEE Spectrum article
(http://spectrum.ieee.org/energy/the-smarter-grid/wimax-for-smart-grids). It
assumes you control consumption by having the power company turn off the
electrons (AC -- Alternating Current) going to your AC (Air Conditioner)
instead of providing you with information, bits. With bits you convert
information into the policy appropriate for your situation.

The Internet has been transformative because we've been able to focus on the
relationship between end points decoupled from the path. The primary end
points have been mainframes and now PCs and cellular phones so we can get a
taste of decoupling by having a billing relationship with a broadband
service provider. This works because users understand the value of broadband
services for the web. We can then have other applications take advantage of
the opportunity. But the designers of devices and third party services
cannot assume that a path is available. And, as we see, even when there is a
potential path we have too many policy barriers and mental barriers in the
way.

When we try to connect things like power meters or medical monitoring
devices there is no direct or necessary relationship to particular billing
paths. These applications don't fit into the current funding model and thus
don't happen or we try to build special infrastructure for each
applications. We see this with trying to create "public safety" network at
700MHz. It's not only expensive but it deprives us of the ability to take
advantage of shared innovation while leaving us with a brittle dependency on
each such network.

Incidentally, I'm using the term power meter -- but we seem to be focusing
on just electric meters. What about water, gas and other meters? What about
measuring thermal consumption and carbon in the local atmosphere? This is an
example of how silo-thinking limits our opportunities.

The power (electricity providing) company has a relationship with the meter
-- does it need to provide its own path? Strangely enough that's what it
seems they are trying to do (and want to do?).  Why not use SMS? As Rahul
points out SMS pricing presumes that bits have a particular meaning and the
carriers' price for high value applications so as to maximize their profits
(ARPU) and in the absence of alternative bit paths they become as
rent-takers dictating prices.

The reason goes back to the days of telegraphy when the product was
telegrams and they built an expensive infrastructure for that one purpose.
It was an analog (electron-based) technology silo. You could associate the
path with and set a price on the value of a telegram. Better to keep the
telegraph line idle than use the spare capacity for inexpensive messages --
the same behavior we see in keeping copper lines idle rather than making
them available for even low speed connectivity as a community resource.

This approach doesn't make sense with bits. The obvious solution is to build
a common infrastructure for carrying bits without presuming a particular
purpose (or value). Instead of charging for the value of the bits you merely
need cover the costs. It turns out the costs for such a "bit commons" is
very low. Even if we put in new fiber at $1K/home we can quickly amortize it
an get the benefits of a very low marginal cost. If we didn't need to
contain bits in billable paths we wouldn't have to run fiber directly to
each home. If we didn't have the constraint of billing we could take
advantage of any path available at a very low marginal cost. Yet even with
an expensive infrastructure like California freeways we don't create a
billable event for each trip. So it's not really about cost as much as about
the value of having the infrastructure. It's not about cost, it's about
value. Without understanding we don't see the value.

We cover the costs of the common infrastructure by having the community fund
it. This is not necessarily government funding -- you fund your home's
connectivity and a building owner (or tenants) might do it for the building.
This is government as a community and not government as command and control
and anyone can contribute capacity. We have a real example in the history of
roads, http://eh.net/encyclopedia/article/Klein.Majewski.Turnpikes, in which
the business community would fund a road as an enlightened rather than
for-profit, investment. Once we get past the idea of rent-taking as a
business model we can welcome users as participating contributors rather
than competitors. If you run your own wires you're not stealing capacity but
are instead contributing to the community.

What's strange is that the market has forced us to share a common
distribution system. It doesn't make sense to have redundant power wires.
Yet these same companies seem to assume they need a different system for
sharing bits -- and we don't even consume bits! We create costs -- high
costs --without creating value!

I've compared this shift from electrons to bits with coming to terms with
Copernicus who merely shifted a reference point to simplify calculations
(and provide insight). It took generations to get past the idea that we were
at the one center of the universe. It may be even more difficult to shift
from thinking about electrons to decoupling bits from their interpretation.
Without this understanding we accepting paying a trillion dollars a year in
billable events and a higher price in preventing the development of vital
applications that don't profit the carriers.

We see an example of this in the so-called residential gateways or triple
plays. The carriers want to charge you (and third parties) for each service
because selling you Web, telephony and television allows them to charge for
the value whereas bits have no intrinsic value. Bits allow us to create more
value but that value doesn't accrue to the providers. This is the conflict
inherent in funding networking as a service with value-based pricing. No
wonder power companies want a cut of the action (as with BPL) even if, as we
see, they don't understand it.

http://frankston.com/public for more.

-----Original Message-----
From: nnsquad-bounces+nnsquad=bobf.frankston.com@nnsquad.org
[mailto:nnsquad-bounces+nnsquad=bobf.frankston.com@nnsquad.org] On Behalf Of
Rahul Tongia
Sent: Wednesday, July 14, 2010 01:53
To: Bob Frankston
Cc: Lauren Weinstein; nnsquad@nnsquad.org
Subject: [ NNSquad ] Re: Why it's hard to hack the power grid, and why NSA
is the wrong choice to protect it

As Bob has already shown, it's not "broadband" that is the issue.

Forget the enormous costs ($440/home) shown in the IEEE article.

A much greater issue is ownership, control, and boundaries (which come up
between utility/consumer and, in this mode, utility/telco).  Any system
operated by a 3rd party has implications for predictability and control.
SLAs are fine, but I know of a smart metering system that failed because the
carrier found the penalty lower than the profit margin on sending SMSs (text
messages) for paying (non-negotiated
bulk) consumers.

Meter reading is fine - you can wait, but real-time and *control*
applications over WiMax or any other carrier, shared network? One has to
think carefully.

Rahul

On Wed, Jul 14, 2010 at 4:35 AM, Bob Frankston
<Bob19-0501@bobf.frankston.com> wrote:
> Even without all the political over tones the word "perfect' seems
> antithetical to the idea of resilience. Perhaps one of the "skilled
> employees" at the NSA it trying to signal a concern?
>
> This reminds me of Y2K reasoning in assuming that every system is very
> brittle and the only fix is to harden it. In fact the real lesson of
> Y2K and the Internet is the power of loose coupling and the
> constraints of best efforts which force you to deal with exceptions and
verify information.
>
> It's not just about the NSA butting in but the larger problems with
> the so-called "Smart Grid" where all the old ideas seem to go to
> fester. There was a worrisome piece, WiMax for Smart Grids
> (http://spectrum.ieee.org/energy/the-smarter-grid/wimax-for-smart-grid
> s) in IEEE Spectrum. While much of the confusion may be due to the
> author of the piece it does reflect much of this rigid sloppy
> thinking. The power industry has long been enamored with the idea of
> building its own network though finally, it seems, the Broadband over
> Power Lines (BPL) cited in the article may have finally been put to
> rest. The article itself uses a big number --
> 300 billion bytes per year for the data from a million smart meters --
> as a rationale for requiring a high speed network. Yet 3e11 bytes is
> 3e12 bits
> (10 bits/byte to round up) or (3e12/1e6) 3e6 bits per home per year or
> 1e4 per day (300 days rounding the result up) or 1e4/1e5 bits per
> second
> (rounding) or 1e-1 bits per second. 3e11 is a WAG but for those who
> want false precision 3e11*8/1e6/365.25/24/3600=.08bps. That's not
> exactly high speed. While it's tempting to comment on the article as a
> blooper reel the real question is the extent to which it reflects
> politician/regulator understanding.
>
> If the government wants to protect us against attacks shouldn't it be
> working on improving the resilience rather fostering a climate of
> brittle dependency?
>
>
> -----Original Message-----
> From: nnsquad-bounces+nnsquad=bobf.frankston.com@nnsquad.org
> [mailto:nnsquad-bounces+nnsquad=bobf.frankston.com@nnsquad.org] On
> Behalf Of Lauren Weinstein
> Sent: Tuesday, July 13, 2010 14:22
> To: nnsquad@nnsquad.org
> Subject: [ NNSquad ] Why it's hard to hack the power grid, and why NSA
> is the wrong choice to protect it
>
>
> Why it's hard to hack the power grid, and why NSA is the wrong choice
> to protect it
>
> http://bit.ly/9YJocq  (Wired)
>
> --Lauren--
> NNSquad Moderator
>
>
>