[p2p-research] About "Failing business models for user-generated content"

Tue Apr 28 18:58:30 CEST 2009

(note: below I refer to several URLs and threads passed here recently,
ie potentially already familiar to all the subscribers of this
list. If anybody thinks this is worth putting online somewhere, I'll
put together an HTML version with all the links)

The last post at the P2P Foundation blog
[http://blog.p2pfoundation.net/failing-business-models-for-user-generated-content/2009/04/28]]
is a good occasion to share some thoughts and doubts I've had for
years about the "fast, universal Net access for everybody" dogma.

I deliberately focus on what is actually doable, good and necessary
**now** (0 to ~30 years), not on how life could be or will be for our
grand-grandchildren.

There are a few concepts to re-analyze together here:

1) production, replication and distribution of digital goods costs nothing, or practically nothing
2) centralized infrastructures (think YouTube, Flickr, etc...) are much more expensive than distributed infrastructures (file sharing networks from Napster to TPB and beyond)
3) regardless of its architecture, the costs of the infrastructure should/could be distributed (socialized)
4) fast, cheap, unmetered internet access is good, necessary and a human right, like education, etc...

Concept #1 is, very often, a myth. Producing and keeping up and
running all the material infrastructure needed to copy and distribute
willy-nilly "immaterial" goods has a huge cost. Huge. Read for
reference:

http://blog.p2pfoundation.net/one-loaf-per-child/2007/06/14
http://p2pfoundation.net/Thoughts_on_P2P_production_and_deployment_of_physical_objects
http://www.infoworld.com/d/green-it/report-us-companies-waste-28b-year-powering-unused-pcs-758
http://bits.blogs.nytimes.com/2008/05/01/data-centers-are-becoming-big-polluters-study-finds/?ref=technology
http://www.socialroi.com/interesting-facts-about-digital-waste.html http://sitemaker.umich.edu/section002group3/e-waste
http://www.afro.who.int/heag2008/docs_en/New%20and%20emerging%20threats.pdf
http://www.ban.org/photogallery/

With respect to concept #2, bandwidth is bandwidth and storage is
storage. I am not sure that changing from concentrated to distributed
would reduce so much the order of magnitude of the cost mentioned in
the links above. More on this in a moment.

I am not sure that performance and reliability would always be at
least equal in the distributed case. Downloading a video from a
professional server farm like YouTube's takes more or less the same
time now and one month from now. Whereas downloading via torrents is
faster only if you want what everybody else is downloading, or at
least seeding, in the same moment. This is a minor point however.

A more important issue is if the real cost of one billion desktops and
fast bidirectional connections, all "forced" to be servers/sources
because there is no centralized structure built in a much more
professional, highly efficient way, using much less raw materials and
energy. Are there any data about this?

Back to the cost of distributed vs "centralized content distribution"
networks. I recently found some numbers about a very efficient, highly
decentralized content distribution networks, ie spam. Spam today is
mostly generated and redistributed by infected desktops but still
dissipates 33TWh/year:

www.net-security.org/secworld.php?id=7316 www.softlist.net/press/pandalabs_says_half_a_million_computers_are_infected_with_malicious_bots_every_day-62.html
www.guardian.co.uk/technology/2009/apr/08/spam-malware-online-security

So, the first two points could be summarized as "no matter how you do
it, production, replication and distribution of digital goods of the
scale advocated by proponents of universal access, worldwide free
flowing culture and communication, etc...has a very big, very material
cost. Better wake up and accept it"

At this point, we can move to "who should pay for it". It's an
important decision, because the cost **is** huge but the benefits to
society of fast, universal connectivity should be even bigger.

Personally, I am sure that there are a lot of important things that,
from a human/civil right standpoint, everybody should have the
possibility (someetimes the duty!) to do online: communicate, study,
check what the government is doing to keep it under control, publish
opinions, work, etc... I'm convinced that the more people actually use
the Internet in this way, the better, so I could agree with a law that
says "doing these things is a basic right, so everybody must be
connected and the related costs must be entirely socialized"

The point is, to do all those things you don't need broadband. Not the
kind the ISPs want to sell these days, to transform the Internet in
the next TV.

I think that to make your own life and the world better, a stable,
flat-rate connection at 512KBps or so is enough these days, so I could
support socializing the costs of connecting everybody in that way. I
would be happy if flat rate Internet offers where I live didn't start
at ~20 Euro/month for 5 MBps but from something like 5 Euro/month for
512KBps. Why don't they?

But is it really necessary and fair to subsidize the cost of both the
access and (above all) the backbone infrastructure that would be
needed to keep everybody hooked at 5, 10, 20 MB?? I'm not saying it
would be bad, but if it costs too much (especially in these times) and
it isn't really necessary...

There are tons of "content" online that could take much less space and
bandwidth. I find all the time podcasts or video tutorials which are
in that form ONLY because the author wouldn't take the time to
transcript what he said. So we're all left with something that takes
much more to download, isn't indexable as efficiently as text, is much
slower to navigate back and forth, much harder to mix and mash with
other content...

Here are a few cases where real fast broadband isn't, IMO, something
whose costs should be socialized (remember that you can't really
separate the costs of access from those of backbone)

- if you're stuck with cloud computing or software as a service,
  including any webmail or online office suite, that is if you gave up
  control of who owns and sees all your data. If so, you have a
  separate problem which would remain even if bandwidth were really
  gratis

- online games

- digital bulimia, that is downloading everything you find, even if
  you'll never need it, "just because it's there"

Please note that, even if you're a "producer", ie a private citizen
who blogs on whatever argument, a small business or an artist who puts
online his/her works, it can MAKE much more technical/ecological/
economical sense to rent for that a virtual server in a professional
datacenter in a country which gives a minimum of warranties about
civil rights... than to run a website from the second-hand PC in your
closet which could break every second and consumes much more power
than a virtual one (Virtual servers are NOT cloud computing!).

The solution? Maybe it is to subsidize "basic flat connectivity" as
described above, and let everything else to be paid directly by those
who have or feel the need for it.

More exactly: instead of an infrastructure based on the dogma that
everybody needs and wants 10+ MBps to everything else worldwide,
dimension network and fees so that the basic connectivity costs much
less than today, but the "heavy users" pay by themselves the burden
they place on the network (ie on everybody else and the environment),
since very often what they're doing is NOT bad, but isn't really
necessary to society as a whole.

This is NOT an attack to net neutrality, of course.  In this context,
a good read is "Net neutrality: An American problem?"
http://news.cnet.com/8301-1035_3-10053045-94.html (I agree with those
Australian ISPs!).

This approach is just recognizing that advertising like "for just one
low fee you're entitled to unlimited traffic" (which, by the way, is
exactly what keeps running today's decentralized file sharing
networks!!!) never was nothing more than an unsustainable marketing
scam.

This isn't even anything new, by the way. It has been a well known
fact of life among webmasters for YEARS that "if an hosting provider
gives you unlimited bandwidth, he's either incompetent or not 100%
honest. There can't be anything like "unlimited bandwidth". All
hosting providers cut traffic, because their network would melt or
cost too much otherwise. It's just that the good ones tell it to you
fair and square in advance, in the contract, how and when they will do
it". Search in any webmaster forum and you'll find posts repeating
these concepts since the nineties, and they never changed.

So the only news here may be the wish that literally everybody (not
just "professionals") can become a producer, or at least a
(re)distributor, without analyzing if the goal is (at least in the
medium/short term) intrinsically sustainable, or generally worth of
being completely socialized.

Marco Fioretti
Digital Rights writings: http://mfioretti.com

-- 
Your own civil rights and the quality of your life heavily depend on how
software is used *around* you:            http://digifreedom.net/node/84