Felix Stalder on 3 Sep 2000 17:19:48 -0000


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

<nettime> (file)sharing and hoarding


[This text was triggered by Steve Cisler's comments on the same study here
on nettime a while back. It is perhaps interesting to remember that in the
same research center a lot of work is being done by Mark Stefik and others
on "trusted systems" for copyright protection.]

Sharing and hoarding: Are the digital commons tragic?

A recent study, entitled Free Riding on Gnutella [1], reveals lots of
taking and little giving among users of the file sharing system Gnutella.
The study presents important data but draws questionable conclusions from
it.

Conducted by researchers of the "Information Ecology Area" [2] at Xerox's
Palo Alto Research Center (PARC) [3] this study is based a 24 hour traffic
analysis of a single node in the Gnutella Network. Through this traffic
analysis, the researchers "established that 70% of Gnutella users share no
files, and 90% of the users answer no queries." Effectively, this means
that only 30% of the users contribute any files to the common resource
base. The study goes on to say that even among those who do contribute,
the concentration at the top is heavy. The top 10% of hosts contribute 87%
of all files, with close to half of all files (40%) provided by the top
1%. Furthermore, 90% of all users either provided no files or the files
they provided were never requested.. The files that were actually of
interest, hence downloaded by others, concentrated on only 10% of all
hosts.

This data questions some commonly held assumptions about the nature of a
distributed system such as Gnutella. First, the system is much less
distributed than the number of hosts indicates. A relatively small number
of hosts constitute, in effect, a central repository for a large part of
all files, particularly the popular files. Second, this concentration
(re)introduces into the system a number of vulnerabilities that were
thought to be avoided by its supposedly distributed nature. The system is
more vulnerable to censorship or hacking attacks (Distributed Denial of
Service attacks, for example) than usually claimed because it is possible
to identify the relatively small number of hosts that contribute the
majority of resources. 40% of the resources, as the study shows, were
contributed by only 314 hosts. While this is significantly more than the
single central directory of Napster, it still might not be too difficult
to enforce copyright/intellectual property/censorship laws against most of
them, once a clear legal regime has been established and precedences have
been set.

This heavy concentration of resources might introduce another weakness:
the unequal use of bandwidth throughout the system. If only 10% of hosts
contribute those files that are actually downloaded, then this small
number of hosts will have to carry 100% of the bandwidth used in the
system, potentially introducing bottle necks, slowing down transmission,
and burdening the most valuable contributors with the lion share of the
(bandwidth) costs. Hence the system punishes those who contribute the
most.

The researchers conclude: "These findings have serious implications for
the future development of Gnutella and its many variants. In order for
distributed systems with no central monitoring to succeed, a large amount
of voluntary cooperation is required, a requirement that is very hard to
fulfill in systems with large user populations that remain anonymous."
Consequently, an open file sharing system is likely to be affected by "the
tragedy of the digital commons."

The tragedy of the commons refers to a tendency of freely available
resources to degrade over time. In his classic 1968 article, Garret Hardin
gave the archetypical example of this:

Picture a pasture open to all. It is to be expected that each herdsman
will try to keep as many cattle as possible on the commons [and] the
inherent logic of the commons remorselessly generates tragedy. As a
rational being, each herdsman seeks to maximize his gain. Explicitly or
implicitly, more or less consciously, he asks, "What is the utility to me
of adding one more animal to my herd?" This utility has one negative and
one positive component.

1. The positive component is a function of the increment of one animal.
Since the herdsman receives all the proceeds from the sale of the
additional animal, the positive utility is nearly + 1.

2. The negative component is a function of the additional overgrazing
created by one more animal. Since, however, the effects of overgrazing are
shared by all the herdsmen, the negative utility for any particular
decision-making herdsman is only a fraction of - 1.

Adding together the component partial utilities, the rational herdsman
concludes that the only sensible course for him to pursue is to add
another animal to his herd. And another.... But this is the conclusion
reached by each and every rational herdsman sharing a commons. Therein is
the tragedy. Each man is locked into a system that compels him to increase
his herd without limit -- in a world that is limited. Ruin is the
destination toward which all men rush, each pursuing his own best interest
in a society that believes in the freedom of the commons. Freedom in a
commons brings ruin to all. [4] The difficulty of this example, apart from
assuming a rather limited rationality, is that it is not clear how it
applies to the sharing of digital goods, if it applies at all. Rishab
Ghosh, in his Cooking-Pot Markets [5], argues why it might not apply: With
a cooking-pot made of iron, what comes out is little more than what went
in -- albeit processed by fire -- so a limited quantity must be shared by
the entire community. The Internet cooking-pots are quite different,
naturally. They take in whatever is produced, and give out their entire
contents to whoever wants to consume. The digital cooking-pot is obviously
a vast cloning machine, dishing out not single morsels but clones of the
entire pot. But seen one at a time, every potful of clones is as valuable
to the consumer as were the original products that went in.

The difference stems from the fact that taking from the digital resource
base does not diminish what is available to others. Since digital data is
not tangible, giving it away to others does not imply being dispossed of
it. As a consequence digital commons can tolerate a much higher degree of
consumption than physical commons, where everything that is consumed needs
to be replaced before it is available.

The fact that none of the file sharing systems has been negatively
affected by the "tragedy of the commons" suggests that it might not apply
to digital goods. So far, every increase in diversity of the files
available is an increase of the attractiveness of the system for all
users, even if the number of users grows quicker than the number of files.
This suggests that these systems are much more stable than implied by the
PARC study and much less in need to be converted into "market-based
architecture."

While these systems are unlikely to collapse under their own weight, the
study does indicate that the centralization, introduced by the extremely
unequal distribution of resources throughout the network, makes the system
vulnerable to hostile attacks, both on a technical as well as on a legal
level. This shows, that even cleverly designed system cannot guarantee the
free flow of information in an environment that is either not willing to
support this goal or downright hostile to it.



------------------------------------------------------------------------

[1] http://www.parc.xerox.com/istl/groups/iea/papers/gnutella
[2] http://www.parc.xerox.com/istl/groups/iea/
[3] http://www.parc.xerox.com/
[4] http://dieoff.com/page95.htm
[5] http://www.firstmonday.dk/issues/issue3_3/ghosh/index.html


[Originally published by Telepolis <http://www.heise.de/tp>]



--------------------++-----
Les faits sont faits.
http://www.fis.utoronto.ca/~stalder




#  distributed via <nettime>: no commercial use without permission
#  <nettime> is a moderated mailing list for net criticism,
#  collaborative text filtering and cultural politics of the nets
#  more info: majordomo@bbs.thing.net and "info nettime-l" in the msg body
#  archive: http://www.nettime.org contact: nettime@bbs.thing.net