On Wed, Aug 24, 2011 at 12:56 AM, Volodya <volo...@whengendarmesleeps.org>wrote:
> > I think that your argument is 100% valid and sound. It is (with a single > exception) impossible to have both 1) Guarantee that the content that you > wish > to distribute will remain visible after you go offline and 2) Guarantee > that you > store only content that you are "interested in". (The exception i am > mentioning > comes from some sort of mandate that all the nodes/computers remain online > all > the time, which is unrealistic and undesirable for ecological reasons). > > At least some of the content will be interesting only to the people who > appear > online after you go offline. And if you refer to the "whole of the > humanity" > that segment will become larger as you are talking about different > time-zones > and thus patterns of online activity. > > Well, fortunately, real world p2p systems generally don't have to offer guarantees of any kind, unless you want to offer services like backup that are probably better provided for pay by commercial providers with SLAs. All that really has to happen is that your data "usually" remains available. If you store content you're interested in, this actually provides a fairly good property: a piece of content's availability is proportional to its popularity. This is, in fact, how content distribution works in the real world. I think it's OK to also store content your friends are interested in to improve availability, but I'm standing firm on my claim that requiring people to store content they don't care about, at least a large amount of it (versus, say, a DHT that only has metadata, which people are obviously willing to do because it's how Bittorrent works nowadays), is a dealbreaker. It's possible to create complex schemes that attempt to work around this problem ala mnet, but then the complexity of the system becomes your dealbreaker.
_______________________________________________ p2p-hackers mailing list p2p-hackers@lists.zooko.com http://lists.zooko.com/mailman/listinfo/p2p-hackers
