Quote from: zhanlang13 on May 01, 2014, 08:02:32 AM

Max Coins is？I want to know the total number of this coin

"network will mine coins such that the annual inflation rate is kept permanently at 5%"

No maximum. Every year the coin volume will go up by 5%. We do it this way because we want the coins to be about the storage, not about the speculation. There will be explosive growth in the first 4 years however, so there's still a ton of opportunity to jump on the coin early.

Quote from: djm34 on May 01, 2014, 12:10:06 PM

actually it sound interesting (sound like a bit like torrent...)
But what about the network traffic generated by this ?
What kind of file gets into the storage ?
How do you protect the rest of the computer from being infected by crappy files downloaded there ?

It's a lot like bittorrent, but instead of each file being entirely stored on each seeder, each file is only partially redundant (you get to pick the redundancy). It's hard to guess at exactly how much network traffic there will be, but probably a lot for people who are hosts. There will be support for capping your bandwidth, though this will affect your income. (The model for charging for bandwidth isn't worked out yet).

Any file can get into storage. The clients will encrypt them first, but the network doesn't care. If you want to upload an unencrypted file (so it's available to the public), you will be able to do so. I'm not sure what you mean by "crappy" file, but it's just storage. If someone uploads a virus, it can't hurt your computer because the client doesn't open or read files, only stores them. If you try opening it with some program however, it might infect you.

Quote from: TheMightyX on May 01, 2014, 12:43:34 PM

Maidsafe has been developing for 7 years.
You think you and your anonymous team are going to be able to accomplish what they couldn't in 7 years with a large team?

This is not the first intended "proof of storage" system to grace our forums.
Do you have any proof of concept yet?
It would be nice if this was real and not a scam, but you are biting off a very big piece of pie.

1. Maidsafe has been working on this since before Bitcoin. A lot of solutions that are now available and have been thought about in greater depth were not available when maidsafe started. I'm pretty sure that if they started now, they would get a lot farther in the same 7 years.
2. The maidsafe project is considerably more ambitious than Sia. They are doing proof-of-bandwidth, proof-of-storage, proof-of-computation, and something related to uptime monitoring. It also seems like they intend files to be mobile (?). Regardless, from the best that I can tell, Maidsafe is much more ambitious than we are, and takes a more complex approach to many of the problems. This will increase the development time.
3. This project is still very ambitious, and is a few months away from a prototype. We are definitely trying to do some very difficult things. We are not asking for money at this point, nor are we giving away coins at this point. We plan to do both in the future, but will wait until we have something more substantial built.

I know that a lot of you are cautious about scams. If you look, you'll see that nowhere in this post or on our github page does a way to pay us appear (no bitcoin addresses, etc.). This project is very real and is very much not a scam; we aren't asking for your money.

Do we have a proof of concept yet?:

Well, we have a working demo that does not a whole lot. This currency is based on quorums, and we just finished building the base for quorums. A bunch of hosts can get online, engage in communication (following Leslie Lamport's solution to the byzantine generals problem) and kick dishonest hosts off the network. At the moment, there is no support for joining with a wallet or storing files. That's what we're working on now. Stay tuned.

Quote from: Touque on May 01, 2014, 02:03:26 PM

In this system, I wonder how the blockchain consensus will be reached.

Each set of 8GB that a miner offers to the network becomes a 'participant'. Each participant downloads and watches a portion of the network, and is paired with 128 other participants (chosen randomly). The 128 operate together reaching consensus using the Byzantine Generals solution provided by Leslie Lamport in his paper "The Byzantine Generals Problem." The pdf can be be found in the doc/ folder. We used the signed solution, meaning every honest host in a quorum is guaranteed to reach the same block as every other honest host within the quorum.

Quorums are arranged into a tree. Each quorum has a set of sibling quorums that all share the same parent.

When a block is found, it's announced to the sibling quorums, and to the parent quorum. The exact algorithm is not pinned down, but each participant will send the block to 2 or 3 other random participants. This keeps communication to a minimum. A group of dishonest hosts could announce a dishonest block, so there is protect against this; upon receiving a new block, each host will fire off a confirmation to the other participants in the announcing quorum. With very high probability, at least one honest host will see the dishonest block, and will be able to alert the receiving quorum about the dishonesty. After we pick the exact numbers, we will write a formal algorithms with proofs demonstrating what attacks are possible and with what probabilities the attacks will succeed. (Our target is to make the attacks as unlikely to succeed as double-spends on the Bitcoin network when an equivalent volume of hosts are dishonest).

Does that make sense? It's very different from any existing model, using a bunch of communicating quorums instead of using a single large blockchain. The primary reason for doing this is because Sia will have a lot more information in the blockchain than Bitcoin. For example, the header data of every single file on the network is stored in the blockchain. If every miner needed all of that data, you'd be looking at hundreds of gigabytes even before figuring out which miners had which files. A tree was absolutely needed.

We believe the algorithm to secure, but that needs to be confirmed by others. At this point, the algorithm is not formally written down, which makes it hard. Expect to have something more concrete regarding quorum consensus in the next 2 weeks.