how-to block ads
|
|
Share Topic
 |
 |
|
|
| reply to B
Re: So How's It Different From Gnutella? said by B:My point above was that, I think, BitTorrent's speed is DIRECTLY attributable to the success it has in enforcing sharing AND to the single tracker model, allowing many, many people easy access to different bits of a file. (This ensures multiple uploaders and sufficient, decentralized bandwidth.) I tend to agree. Do you want anonymous file-sharing? Or do you want fast and efficient file-sharing? Because I believe that network graph theory would tend to suggest that the most direct pathways through the network will be the fastest, but also the least anonymous (from a network-connection-identity POV - there are other steps that you can take, in terms of creating real-world anonymity and "disconnection" from your online information).
said by B:Decentralizing that would seem, to me, to cut off most of the reason that BitTorrent currently works so well (many people coming to a single source for finding each other and different sections of a file, all of whom are effectively prevented from leeching). I totally agree. It is designed to be basically centralized, like a web site, but every leaf-node that attaches to that "root" of the tree (aka "Seed", I guess, in BT-terminology), creates further branches, that each themselves offer upload bandwidth off of, thus the more people downloading, the more uploading too!
It works kind of like how many popular web sites subscribe to "edge node content caching server networks", like Akamai, so that when you access www.microsoft.com, you're actually hitting XYZ.akami.com, except that BT is all handled client-side, and doesn't need to have an explicit server-side support or fancy geo-DNS or round-robin DNS mappings.
I'm a bit curious in terms of what direction Intel and PlanetLab have been doing, in terms of adding content-caching nodes directly into the core fabric of the network itself - that's a model that I do believe will be the most efficient overall, but it will take getting content producers to accept a non-centralized distribution/server model before it could ever take off in the real world.
I see it as a model (for commercial sites/information) akin to television programming production and distribution. Rather than the television producers including the ads directly, they "wholesale" that content to redistribution server nodes (local to various ISP networks), and then they add "local ads" to the content, ala local television/cable programming affiliates, and pass on a portion of that revensue "upstream" to the content producers. Likewise, for non-ad-supported content, (almost like PPV cable programming) the customer would make micropayments to their ISP for that content, and a portion would be funneled upstream to the content producers.
Interesting how well BT might also map to a similar model like that, if micropayement or dynamic ad-insertion technology were added. The problem is that such a dynamically-forming network topology does best with larger files (not smaller files such as are served as part of web content), because of the requirement to amortize the costs of node connection/disconnection over the actual file-transfer times. | |
BPremium,MVM
join:2000-10-28
1 edit | Well said, Larry.
There's some new discussion over at »yro.slashdot.org/article.pl?sid=···7&tid=17 where some people seem to be rationalizing that trackers aren't really liable as central distribution points. I think they're dreaming.
-- B
--
In a realm outside causality and function | | |
|
| an interesting fact i was told when in a lecture on copyright was that under law (at least uk law, not sure about us) it is NOT illegal to share a copyrighted piece of material, as you are not directly giving the material to the leech (they are taking it from you). it is, as we all know, illegal to own copyrighted material without copyright.
by that model, a tracker would not be breaking the law.
i personally can't understand how people are ever procecuted for copyright infringement. copyright law is unbelievably ambiguous. | |
BPremium,MVM
join:2000-10-28 |
Funny you should resurrect this thread right after major tracker pointers including Suprnova were just shut down!
But your statement sounds backwards -- "they are taking it from you" when you distribute copyrighted material? So that means every person selling copies of new Hollywood releases on the street corner is completely innocent? It's the people buying who are the only criminals?
And I never heard it was "illegal to own copyrighted material without copyright". In fact, I really doubt it is. (And I'm no lawyer.) It would kind of make things like, you know, BOOKS, illegal to possess.
-- B
--
In a realm outside causality and function
| |
| The difference is, your street corner seller is selling for money - that's the illegal act, not the providing of copyright goods, but the providing copyright goods *for sale* | |
|
