Re: what is the cost of bb deployment

[David Farber <dave () farber net>]

Begin forwarded message:

From: Rahul Tongia <tongia () cmu edu>
Date: October 2, 2008 11:09:22 PM EDT
To: dave () farber net
Cc: ip <ip () v2 listbox com>
Subject: Re: [IP] what is the cost of bb deployment

Dave/Tom,

Before we can answer it, we should clarify what the Q is.  Is it  
merely fiber, in which case some fiber designs are vastly more shared  
(think PON), or otherwise constrained (e.g., FiOS) to less than their  
theoretical potential (perhaps for commercial reasons)?  Should we say  
minimum broadband of speed X? 100 Mbps? Bidirectional? With no port or  
usage restrictions? etc.

It's actually easier to give 100 Mbps wireless than some other  
solutions if you're willing to oversubscribe (statistically multiplex)  
that across lots of users since wireless is inherently shared.
IMHO, mandating 100 mbps for everyone (or even 98% of the population)  
might be a tough goal. Unless we can severely price discriminate, it  
may make things expensive for everyone.  Of course, meaningful  
broadband would still be many Mbps, but ones that are achievable  
through copper if not wireless.
What I do think is required is lots of fiber, at least to points  
nearby the last mile.
Interestingly, there is way more fiber in the ground (or on poles)  
than people realize.  The bigger challenge is who gets to use it, at  
what price?
There are several options for how to use existing fiber better, from  
regulated (say, costs-plus) all the way to pure market, with options  
in between.   The challenge with market systems is two-fold. One,  
there is limited physical competition in many places. Second, why is  
that model optimal? Do we have three sewer lines to the home in the  
name of competition? We can compete for who gets to operate the system  
(see my webpage below, publications, FiberAfrica for more on an  
analogy to roads).  A single fiber has enough capacity for not just  
one user but even multiple if well designed.
To answer the Senator's Q, I would frame this as: what's the  
incrementally cheapest way to get meanginful broadband to the  
overwhelming majority of Americans?

If I have to get fiber into rural areas, then the cheapest manner is  
aerial pole based.  Order of magnitude lower than digging up and  
trenching.  Let's say 20 million households (??) are in rural areas  
not near a fiber or CO (excluding the furthest 2%, who might get  
subsidized satellite connections or simply competitive bundled  
satellite internet, instead of today's slow, $70/month service).  Say  
they are on average 0.5 miles from a fiber or incremental fiber (i.e.,  
their neighbor, who we are paying for in this new system).  = 10  
million miles.  @$1000/mile inclusive, that's $10B total cost. (note,  
pole rentals are a policy decision, in some cases regulated. See my  
FCC testimony for more on this, in the Discussion:
http://www.cs.cmu.edu/~rtongia/web2/FCC.html#FCC-Personal_Reactions

This is a very conservative (high) cost given poletop fiber could be  
much cheaper, e.g., $100/mile/strand, plus 20 poles per mile of  
incremental labor.  Add in some simple layer 2 switching gear, and  
it's still cheaper. Secondly, I haven't done the GIS analysis, but  
each home is not (??) half a mile from a neighbor in rural US  
(excluding the most remote 2%).
All of this excludes using copper via DOCSIS 3.0 or advanced DSL, both  
of which need some fiber for backhaul.
Thinking out loud...

Rahul

--
************************************************************************
Rahul Tongia, Ph.D.
Senior Systems Scientist

Program in Computation, Organizations, and Society (COS)
School of Computer Science (ISR) /
Dept. of Engineering & Public Policy

Carnegie Mellon University
Pittsburgh, PA 15213 USA
tel: 412-268-5619
fax: 412-268-2338
email: tongia () cmu edu
http://www.cs.cmu.edu/~rtongia




David Farber wrote:
> Begin forwarded message:
>
> From: Tom Bleha <tbleha () earthlink net>
> Date: October 2, 2008 7:05:48 PM EDT
> To: dave () farber net
> Subject: Re: [IP] GENI Spiral 1 - startup
>
> For IP, if you think it's appropriate.
>
> Late last month, Senator Inoyue held a hearing on broadband and  
> asked those testifying at least twice how much it would cost to  
> deploy broadband across the entire country.  No one had an answer  
> for him.
>
> If we expect to have nationwide broadband, if would be helpful to  
> have an answer.  Does anyone on the IP list have a good estimate for  
> the cost of:
>
>  -- nationwide fiber broadband (reaching 98% of the population or  
> so) and
>
>  -- nationwide ultrafast (100 mbps+) wireless?
>
>  Many thanks, Tom Bleha
> ----- Original Message ----- From: "David Farber" <dave () farber net>
> To: "ip" <ip () v2 listbox com>
> Sent: Thursday, October 02, 2008 11:06 AM
> Subject: [IP] GENI Spiral 1 - startup
>
>
> > GENI Project Office at BBN Technologies Announces $12M in Funding  
> > for  29 Academic-Industrial Teams
> >
> > Academic and Industry Powerhouses Kick Off GENI Prototyping
> >
> >
> > CAMBRIDGE, Mass., September 29, 2008 - BBN Technologies, an  
> > advanced technology solutions firm, announced today subcontract  
> > awards totaling $12M for 29 academic/industrial research teams to  
> > build, integrate,  and begin to operate the first prototypes of the  
> > GENI suite of network research infrastructure. GENI prototyping is  
> > sponsored by the National Science Foundation to support  
> > experimental research in network science and engineering.
> >
> > GENI prototyping will be conducted using a "spiral development"   
> > approach, with simultaneous development and trials giving rapid   
> > feedback to help guide evolving designs.  Spiral 1 focuses on ways  
> > to  discover, schedule, and control resources for large-scale  
> > research  experiments and to measure GENI capabilities. Multiple  
> > competing  approaches are being funded to provide design insights  
> > for the  evolving suite of experimental infrastructure. Successive  
> > spirals will  refine and extend the GENI suite in response to the  
> > research  community's evolving interests in network science and  
> > engineering.
> >
> > "GENI Spiral 1 has begun!" said Chip Elliott, GENI Project  
> > Director. "This first spiral will federate a wide variety of  
> > network research infrastructure, ranging from optical backbones to  
> > disk farms to sensor networks, with the very first prototypes up  
> > and running in six to  twelve months."
> >
> > Awardees include the following universities, colleges, institutes  
> > and centers: Arizona, Carnegie Mellon, Columbia, Georgia Tech,  
> > Houston, Indiana, Kansas, Kentucky, Langston, Maryland,  
> > Massachusetts-Amherst  (2), Massachusetts-Lowell, Ohio State,  
> > Pittsburgh Supercomputing,  Princeton, RENCI, Rutgers (2), Southern  
> > California, Stanford, SUNY  Buffalo, Utah, Washington, Washington  
> > University in St. Louis,   Wisconsin-Madison, and Williams.
> >
> > Corporations, including Ciena, Cisco, CNRI, Fujitsu, Hewlett- 
> > Packard, Infinera, Microsoft Research, NEC, Netronome, SPARTA, and  
> > Qwest will  work with these academic teams.
> >
> > The complete list of proposals funded in GENI Spiral 1 is as follows:
> >
> > * Carnegie Mellon University - David Andersen
> > * CNRI (Corporation for National Research Initiatives) - Larry Lannom
> > * Columbia University - Keren Bergman
> > * Georgia Tech - Nick Feamster
> > * Indiana University - Jon-Paul Herron
> > * Langston University - Pierre Tiako
> > * Ohio State University - Anish Arora
> > * Pittsburgh Supercomputing Center - Matt Mathis
> > * Princeton University - Larry Peterson
> > * RENCI (Renaissance Computing Institute) - Ilia Baldine
> > * Rutgers - Dipankar Raychaudhuri
> > * Rutgers - Marco Gruteser
> > * SPARTA, Inc.- Steve Schwab
> > * Stanford University - Nick McKeown
> > * SUNY Buffalo - Chunming Qiao
> > * University of Arizona - John Hartman
> > * University of Houston - Deniz Gurkan
> > * University of Kansas - James Sterbenz
> > * University of Kentucky - James Griffioen
> > * University of Maryland/Mid-Atlantic Crossroads - Peter O'Neil
> > * University of Massachusetts-Lowell - Yan Luo
> > * University of Massachusetts-Amherst - Brian Levine
> > * University of Massachusetts-Amherst - Jim Kurose
> > * University of Southern California/ISI - John Wroclawski
> > * University of Utah - Jay Lepreau
> > * University of Washington - Thomas Anderson
> > * University of Wisconsin-Madison - Paul Barford
> > * Washington University in St. Louis - Jon Turner
> > * Williams College - Jeannie Albrecht
> >  About GENI and the GENI Project Office
> > GENI is a suite of experimental network research infrastructure   
> > sponsored by the National Science Foundation. As envisioned by the   
> > community, this suite will support a wide range of network science  
> > and  engineering experiments such as new protocols and data  
> > dissemination  techniques running over a substantial fiber optic  
> > infrastructure with next-generation optical switches, novel high- 
> > speed routers, city-wide experimental urban radio networks, high- 
> > end computational clusters,  and sensor grids. All infrastructures  
> > are envisioned to be shared  among a large number of individual,  
> > simultaneous experiments with  extensive instrumentation that makes  
> > it easy to collect, analyze, and  share real measurements. The GENI  
> > Project Office, operated by BBN  Technologies, is responsible for  
> > project management and leading system  development and prototyping  
> > efforts. Visit www.geni.net for more  information.
> >
> > About BBN Technologies
> > BBN Technologies is a legendary R&D organization that leverages its  
> > substantial intellectual property portfolio to produce advanced,  
> > repeatable solutions such as the Boomerang shooter detection  
> > system.  With expertise spanning information security, speech and  
> > language  processing, networking, distributed systems, and sensing  
> > and control  systems, BBN scientists and engineers have amassed a  
> > substantial  collection of innovations and patented solutions. BBN  
> > now employs  approximately 700 people in seven locations in the US:  
> > Cambridge,  Massachusetts (headquarters); Arlington, Virginia;  
> > Columbia, Maryland;  Middletown, Rhode Island; San Diego,  
> > California; St. Louis Park,  Minnesota; and O'Fallon, Illinois. For  
> > more information, visit www.bbn.com .
> >
> > # # #
> >
> >
> >
> >
> >
> > -------------------------------------------
> > Archives: https://www.listbox.com/member/archive/247/=now
> > RSS Feed: https://www.listbox.com/member/archive/rss/247/
> > Powered by Listbox: http://www.listbox.com
>
>
>
>
>
> -------------------------------------------
> Archives: https://www.listbox.com/member/archive/247/=now
> RSS Feed: https://www.listbox.com/member/archive/rss/247/
> Powered by Listbox: http://www.listbox.com




-------------------------------------------
Archives: https://www.listbox.com/member/archive/247/=now
RSS Feed: https://www.listbox.com/member/archive/rss/247/
Powered by Listbox: http://www.listbox.com