On Mon, Apr 17, 2006 at 06:32:41PM +0530, darshan ashpal wrote:
> Hi,
>
> I`m, a noob here but nevertheless am planning to give my final semester
> seminar on Freenet , to spread the good word here.I`ve gone through the
> basic white papers by Ian Clarke,the Darknet pdf given at CCC, and also the
> IEEE transcations,and the Tech FAQ section at
> http://e-privacy.firenze.linux.it/2002/freenet/freenet_0.3
>
> But a few doubts still persist.Any help would be most appreciated, and links
> to updated FAQ`s for v0.7 , ot to previous mailing lists on such topics will
> be most helpful. I` d like to list the key questions here :
>
> 1. Which cypher-schemes are being useda & where?
> I`m assuming its AES Rjindael for the STS protocol & Twofish for
> datastore encryption,is that correct?
No, AES everywhere, and DSA, and Diffie-Hellman (soon to be STS).
> 2. Are the node ID`s long integers / IP hashes? and what exactly is meant by
> "swapping a node ID" when we`re trying to make our network topology to
> follow the Small World Characteristics based on key clustering? swapping
> the DHT`s is a no-no but then whats the alternative to retaining existing
> links?
Identity = 32-byte opaque node unique identifier. Not used for routing.
Location = Double precision number between 0.0 and 1.0. Swapped by the
location swapping algorithm. We ASSUME that the network topology is
correct, and use a variant of the Metropolis-Hastings algorithm which
Oskar came up with to assign locations such that greedy routing (always
route to the node whose location is closest to the target) works.
> 3. How exactly would you implement premix-routing in v0.8? in other words ,
> how do prevent one node from knowing your identity & another node from
> seeing your requested object if you are directly sending a file request
> query to them directly?
A very good question! The answer is "with difficulty". Probably we will
divide the network into cells and expose the topology within those
cells; then the node would pick two nodes at random from within the
cell, and onion wrap the request to the two nodes, and pass it to a
random direct peer.
>
> In any case i`lll be uploading my efforts onto freenet itself & also the
> web, to serve as a beginner-friendly guide for tyros
> Looking forward to hearing from you guys.
Cool.
>
> Thanks in advance,
> BiGFooT
--
Matthew J Toseland - toad@???
Freenet Project Official Codemonkey - http://freenetproject.org/
ICTHUS - Nothing is impossible. Our Boss says so.
