Some Support, on 08 August 2013 - 08:00 AM, said:
DavidXanatos, on 08 August 2013 - 06:24 AM, said:
Well, as we see form the above I don't dictate anything, I only prevent others form dictating me.
Because you convienently skipped the "Addition clause for networking applications" part.
Which is not longer part of the actual draft, but placed separately at the bottom with other source material links.
but lets take a look on it anyway:
Addition clause for networking applications:
5. Every in any way altered version of this software, must identify itself appropriatly during the initial phase of any network communication with other instances of this software.
a. The identification string must be unique and persistent for each distribution.
Usage of a random identifiers or an identifier already used by a different
distribution is strictly prohibited.
b. The identification string may be concatenated with a version or build identifier.
c. The placement and notation of the identification string must be implemented by
all modified versions of this software following the precedent set forth by the
original version.
Why does it limit your freedoms if you are prohibited form miss labelling your product as my product?
Its basically a trademark restriction nothing more.
Quote
I would like it if would take yourself a bit more time for this, but present a detailed, worked out concept for Kad which can really be evaluated rather than one which leaves too many questions open to actually understand it. Of course I'm not judging anything yet, I just doubt that a few hours are enough for that.
The concept is not being developed in a few hours just a synopsis formulated in clean English.
We have been working on the ideas for quite a long time now, actually.
EDIT: And here it is short explicit and in hopefully clean English:
The concept of using Bitcoin techniques is quite simple, we have to increase the cost of creating new identity's, sufficiently to make it not feasible for an attacker to provide a substantial portion of the networks nodes.To make this not to cumbersome for new users to use the network, we allow new nodes to participate with limited rights just from the start.So that for example the users can run searches and also publish them selves as sources, but they will be allowed to participate in the routing only in a limited way and will not be used to index content.This way for example we can already prevent eclipse attacks, as unverified nodes wont be able to send anyone on a wild goose chase.An other problem publishing spam entries can be tackled by making each file publishment cost the node some credits, to be obtained by prof of work. Simmilary to how NameCoin does it:
http://de.wikipedia.org/wiki/NamecoinA central difference would be of cause that all publishments would have only limited lifetime, and our block-chain would not go back infinitely in time but would have to be truncated.The Complexity of a problem would also not be exponentially increasing with time for the network as bitcoin does, but on a network level constant growing only linearly with hardware capabilities. Also having a going long time back consensus record (out block chain) of node behavior would allow us to monitor node behavior and exile misbehaving nodes, as well as actually reducing the cost for nodes that are in for a reasonable time and are behaving well.
This system can be understood as a kind of a decentralized web of trust, as long as at attacker tries to infiltrate an already large network he will fail, though if he tries to infiltrate it form the very begin he will likely take it over. I don't see yet how this particular bootstrap issue could be solved algorithmically, a blond solution is to start the network closed and open it only once enough users have been found.
An important design aspect would be to make it unable to be faster than a normal users by employing Dedicated hardware, like litecoin.org did, by employing a prof of work problem that can not be simplified a lot using ASIC's, in their particular case they employ a problem that needs a loot of memory.
http://en.wikipedia.org/wiki/Litecoin
Now please note that this is a concept that is under development, it is currently not implemented in a productive way.
Also thee system will have to be a little bit more complicated in order not to break the NeoKad inherent anonymity, that is obtained using recursive lookups. But thats not so difficult, a node will have in addition to its main identity, also a set of so called releaser identity's used to associate publications with. This ID's will not bee persistent, but only used temporarily, and it will not be a mater of public record what nodes are operating which releaser ID's.
Cheers
David X.
This post has been edited by DavidXanatos: 08 August 2013 - 11:19 AM
NeoLoader is a new file sharing client, supporting ed2k/eMule, Bittorent and one click hosters,
it is the first client to be able to download form multiple networks the same file.
NL provides the first fully decentralized scalable torrent and DDL keyword search,
it implements an own novel anonymous file sharing network, providing anonymity and deniability to its users,
as well as many other new features.
It is written in C++ with Qt and is available for Windows, Linux and MacOS.