[ppml] a modified proposal 2005-8

[Howard, W. Lee]

> -----Original Message-----
> From: ppml-bounces at arin.net [mailto:ppml-bounces at arin.net] On 
> Behalf Of Robert Bonomi
> Sent: Thursday, March 16, 2006 5:04 PM
> To: ppml at arin.net
> Subject: Re: [ppml] a modified proposal 2005-8
> 
> > Date: Thu, 16 Mar 2006 16:45:01 -0500
> > From: "Howard, W. Lee" <Lee.Howard at stanleyassociates.com>
> > Subject: Re: [ppml] a modified proposal 2005-8
> >
> > > -----Original Message-----
> > > From: ppml-bounces at arin.net [mailto:ppml-bounces at arin.net] On 
> > > Behalf Of Davis, Terry L
> > > Sent: Wednesday, March 15, 2006 10:29 PM
> > > To: Houle, Joseph D (Joe), CMO
> > > Cc: ppml at arin.net; bookeym at pachenalight.com
> > > Subject: Re: [ppml] a modified proposal 2005-8
> > > 
> > > Joe
> > > 
> > > Nope, you read it correctly!
> > > 
> > > The power company will require your electrical systems to 
> be on their
> > > PLC networks in order to control your electrical systems; it 
> > > would make
> > > a completely unworkable control and routing system for 
> the electric
> > > company to try to map the homeowners ISP assigned networks to 
> > > your home load controller.
> >
> > Why?  They have to have a table mapping (IP) address to 
> (home) address,
> > so why does it matter if the IP address is theirs?
> 
> *dynamic* IP addresses assigned by ISPs are subject to change 
> at any time, without notice.  <grin>

Do you advocate using dynamic assignments in IPv6?

> Or, how about an ISP that uses RFC-1918 space internally, w/ 
> NAT/PAT for
> that traffic that goes 'off net' to an external network?  Every 'SYN' 
> packet from your machine may have a _different_ source IP 
> address at the
> destination.  *WHICH* address should be in  the  'mapping table'.

Do you advocate the use of private address space in IPv6?  I
think an internet draft is needed for that to occur.


> If the utility does query/polling of your load controllers, 
> you're running
> *servers*.  this complicates ISP network architecture considerably.

Servers are hosts.  The bandwith may be lopsided a different way,
but how much data are you planning to pull?  Sorry, that's an 
engineering question, and not relevant.

> The utility has *GOOD* reasons to run a 'private' -- *not* 
> connected to the
> public Internet for power-control functions.  Contemplate 
> what could happen
> if a hacker started 'spoofing' directives from the power 
> company  -- whether
> it be to 'shut down all equipment immediately', or to 'run 
> everything at full power'.

Security considerations are a good answer.  So you expect there to
be nine separate networks to each home?


> > > You will need to interface to their control center 
> somehow to set your
> > > home systems/load controllers and that could be via any available
> > > networks but the actual controls will need to come in over their
> > > networks.
> >
> > I'm trying to understand your position:  The power company needs to
> > build its own IP network in order to manage power systems at each
> > home; their IP address assignments will be from their aggregatable
> > block.
> 
> *Yes*, they need their own _isolated_ network.
> _NO_, the IP address assignments on that network do not need 
> to be from a
> block assigned for 'public internet' use.  They can be from 
> IPv4 RFC-1918 
> space, or an IPv6 equivalent, for example.  

You will need to submit an internet draft to the appropriate IETF
working group.

> Or, they _can_ be from a block
> of public internet addresses -- but they don't _have_ to be.  
> Since it _is_
> an *isolated* network, they could even use addresses in 
> collision with use on the public Internet.

What policy are you advocating for or against?
I'm not trying to engineer networks, I'm trying to understand
what opinions are being expressed.  Should each network (service
provider?) assign addresses from its own block, or should each
residence be assigned addresses from a portable block?   Are we
talking about a /64 for a home, or some other size, and is that
per network, per resident, per device, or what?
 

Lee