[p2p-hackers] Computer scientists develop P2P system thatpromises faster music, movie downloads

Serguei Osokine Serguei.Osokine at efi.com
Fri Apr 13 14:12:31 EDT 2007


On Thursday, April 12, 2007 Justin Chapweske wrote:
> Maybe the title is misleading, but "Computer scientists develop 
> P2P system that promises faster music, movie downloads" primarily
> applies to illegal file sharing as far as I'm concerned.

	Actually I wouldn't be so fast to label this technique. From
what I've seen so far, it seems to be primarily targeted at the rare
(single-copy) long-tailed content, and its best results seem to be
produced on it.

	No one needs an extra source for a Harry Potter movie download.
It already has 5,447. But some item in the long tail that is present
in just one copy on the network is more likely than not to belong to
some indy musician, who views P2P as a marketing tool - not as a
piracy one.

	Piracy seems to be unaffected by this feature. But the P2P
network usability for the legally shared obscure content might be
significantly improved by it (provided that this feature does 
really operate as advertised, of course).

	Best wishes -
	S.Osokine.
	13 Apr 2007.


-----Original Message-----
From: p2p-hackers-bounces at lists.zooko.com
[mailto:p2p-hackers-bounces at lists.zooko.com]On Behalf Of Justin
Chapweske
Sent: Thursday, April 12, 2007 11:46 AM
To: theory and practice of decentralized computer networks
Subject: Re: [p2p-hackers] Computer scientists develop P2P system
thatpromises faster music, movie downloads



On Apr 12, 2007, at 12:33 PM, coderman wrote:
>
> block level identification and caching/distribution is useful,
> effective, and proven for a wide variety of tasks.  riverbed [0] in
> particular is doing good *legitimate* business with this technique for
> WAN acceleration, and there are many other applicable domains.
>
> best regards,
>
>
> [0]  Riverbed RiOS Technology
>   http://www.riverbed.com/technology/

First, I don't want to diminish the work of the researchers: They  
chose a problem and came up with an interesting solution.

Delta compression and finger-printing techniques are already well- 
known in applications such as Riverbed's, and that technology is  
certainly interesting and useful.  I view this paper as an extension  
of those techniques for a use in a decentralized system with the  
constraint of uncorrelated sources.  If the uncorrelated sources  
constraint went away, and similar data tended to come from the same  
ingest points, then there are much simpler routes that you could take  
such as a simple extension of the rsync concept.  Since this research  
focuses on on the effectiveness of the technique for uncorrelated  
sources, I'm led to conclude that the focal use case is piracy.

Maybe the title is misleading, but "Computer scientists develop P2P  
system that promises faster music, movie downloads" primarily applies  
to illegal file sharing as far as I'm concerned.  These approaches  
will do little if anything to speed up one->many movie and music  
distribution systems.

>
> (Justin: surely you, of all people, would be aware of the many uses
> for powerful content distribution tools across the spectrum of
> ethical/unethical uses.  why so quick to dismiss a useful technique
> with the tarnish of piracy?)

Because I think there are simpler and more efficient techniques for  
delta-compression for most of the legitimate uses.


_______________________________________________
p2p-hackers mailing list
p2p-hackers at lists.zooko.com
http://lists.zooko.com/mailman/listinfo/p2p-hackers


More information about the p2p-hackers mailing list