Could you please elaborate in the aforesaid upcoming project?. Have you considered other DHT models like Kademlia?.
I had not heard of Kademlia, I will read the paper tonight. Many of these schemes (CAN, Chord, Pastry) are quite similar -- what I planned to do was pick whatever I like best from each, rather than arbitrarily pick one. I will add Kademlia to that lst.
As for my project, search through my posts to get an idea. Basically a new client/new cointype to address some technical and philosophical issues I see in bitcoin. Technically: network topology, storage requirements, instant transactions, built-in pool. Philosophically: all parameters tweakable and set by miners, such as # of coins generated per block. Generated block includes extra "notes" space, intended for miners to write their parameters like "generated coins/block I'd like to see, maximum generated coins/block I will accept as valid". There could potentially be many other parameters. The idea is then one could maintain a "scoreboard" of current mining shares scored by how low the hash is. The scoreboard + the notes gives you a verifiable, trustworthy "current state of mind of coin miners". This allows "rapid dissent", where a group of people who are unhappy with something can very quickly signal to one another via their notes that they are ready/willing to make changes to their parameters if there are others who agree with them, effectively solving the problem of "I want change but I don't want to waste my vote" (where vote here means mining a block that others wont accept as valid).
Anyway I don't want to derail your thread further so let that be the end of that. It will be 6 weeks until I can begin work on this project due to real-work deadlines; when I'm ready I will make a thread for it.
I've said before that saying "DHT" should trigger an automatic ban in IRC. It's an almost perfect metric for identifying people who don't understand bitcoin, don't have any interesting in understanding bitcoin, and whom are going to just spout distracting buzzwords rather than contribute.
What problem exactly do you plan to solve with your DHTs. What measurements have you made to determine that the problem exists? What analysis have you performed that supports that the DHT's will actually improve that problem?
Absent the existence of lite clients bitcoin is a flooding network: All participating nodes need to hear about everything eventually. The randomly wired graphs appear to work perfectly well at this. Because of the inv process any inefficiency in the topology doesn't actually increase bandwidth usage substantially, and global coverage of connected nodes seems to be no worse than a few seconds, which is well below what is strictly required.
Moreover, the 'random' wiring makes the network fairly resistant to various attacks short of isolating a node or adding tons of well distributed sybils. This kind of resistance is hardware in more structured topologies.
I agree that if you are going to change nothing else about the architecture, structured overlay adds little (it is still helpful though, as there are broadcast protocols for structured overlays that are perfect in the sense they avoid
repeat messages).
But it is not clear at all to me that having every client's tx pool being arbitrary and inconsistent is a wise architecture. There seems to be room for improvement here.