Topic: P2P Exchange for bitcoin - page 16. (Read 42787 times)

anyone here at all familiar with hawala? money transfer system from the arabic world going back a long ways. i don't understand it fully, but it seems like it's applicable here.

http://en.wikipedia.org/wiki/File:Hawala.png
ie: (1) A customer (A, left-hand side) approaches a hawala broker (X) in one city and gives a sum of money (red arrow) that is to be transferred to a recipient (B, right-hand side) in another, usually foreign, city. Along with the money, he usually specifies something like a password that will lead to the money being paid out (blue arrows). (2b) The hawala broker X calls another hawala broker M in the recipient's city, and informs M about the agreed password, or gives other disposition instructions of the funds. Then, the intended recipient (B), who also has been informed by A about the password (2a), now approaches M and tells him the agreed password (3a). If the password is correct, then M releases the transferred sum to B (3b), usually minus a small commission. X now basically owes M the money that M had paid out to B; thus M has to trust X's promise to settle the debt at a later date.

wiki: http://en.wikipedia.org/wiki/Hawala

thoughts?

I know this is against the free market but this thread should be PINNED! A P2P Exchange is a MUST and gollum has started us in the right direction. This need the full focus of the entire community right now. Maybe we should even start a fund for the developers.

Gnunet has it's own naming system you can run servers inside it. A web front end for IRC is much easier, a phone app would be especially easy to interact with bitcoin-otc automating the whole orderbook and gpg authentication/rating system.

We're also forgetting all of our cities are chock full of corner stores and currency exchange booths that could easily buy/sell bitcoins to the public too. Just need to tell them how to do so, and automate that as well, or become their traders for them and settle up end of day.

This sounds like it could work. I also think retroshare has potential. At least we can borrow code from it and solutions if those solutions are proven to work.

This is the advantages algos have in the regular stock exchanges:

Physical Advantage
The stock exchange servers are physically located in one or a few server facilities in the same city.
Owners of the HFT-machines wants to be physically close the the stock exchange, every meter closer to the exchange server is an advantage.
When the exchange is decentralized, the current matchmaking could be done 100 meters from the HFT or it could be 10 000 miles from the HFT - the randomness creates unpredictable ping time to the exchange.

Order book Advantage
The order book of regular stock exchange follows a set of rules that can be used to the advantage of algos, the stock exchanges are also executing orders so fast that the algos can make lot of trading before a human even has reacted to a price change on the screen and clicks the mouse to buy/sell. When we got blocks of 10-second execution in combination with random order book matching it is a more fair play between algos and humans. It dont matter if you put your order first or last to buy 1000 BTC at 53$, randomness decides if you are the first to get filled.

An OTC client on top of RetroShare (http://retroshare.sourceforge.net/) would be another option... It's very user friendly as far as these tools go. And there would be sharing files, communication and trade in one application. It already has the web of trust, and already uses PGP just like OTC (but the user can remain oblivious of this). "Just" add in the reputation system and a way to post trades...

IRC would be rather limited because most people aren't going to want to use IRC. That is the main problem with that but also it's got it's own problems as well.

I have no idea how we can prevent high frequency trading but I suspect that we suffered a flash crash due to high frequency algorithmic traders entering. I think Wallstreet finally took Bitcoin seriously this week and the professionals entered the market with their statistical pair trading and other esoteric algorithms. I guess one of the requirements for the protocol design would have to be a mechanism to protect Bitcoin exchanges from high frequency trading and market manipulation. Just assume Wallstreet insiders want to kill Bitcoin and control the market and what can be done to stop it?

Gollum I hope your solution works. It sounds very simple but if it works it works. Order matching every block, it sounds like it could work or at least be part of a solution.

GNU Net has an interesting paper http://secushare.org/2011-FSW-Scalability-Paranoia

Don't know if it would work with Bitcoin for the exchanges or how you'd put Open-Transactions on top of GNUnet.

So what if the exchange protocol could be a chrome/firefox extension? Would this be simple enough? I will get some sleep and maybe someone smarter than me will solve these problems.

If you're getting into GPG auth ect, we already have this kind of decentralized exchange it's called bitcoin-otc

GnuNet is the only candidate for such an endeavor and it still is prone to different kinds of ddos they're working out. For instance you can flood a peer til they disconnect and then impersonate them. Not good for an exchange unless you had cryptosigning, and if you're going to go that route why not just use IRC what's the difference. If you're competent enough to do multisig and pgp signing then you can use bitcoin-otc.

It would be trivial to write an application that plugs into #bitcoin-otc and allows you to authenticate with the bot easily instead of reinventing the wheel. The app can list the order book and check for gpg identify or make gpg identify easy for total idiots

Or, run Open Transactions inside Gnunet. Great success

It might be immune to DDOS because it's decentralized but then we still have to worry about timing attacks, possibly bots with algorithms designed to game the system. It's distributed but then would it be vulnerable to botnets now that it's distributed? How can we verify each node in the network isn't a zombie node or part of a dark pool to manipulate the network and how would we detect certain patterns or at least design the system to be modular enough that the protocol can continually update it's defenses via plugins or extensions?

I fear if anything is thrown together without much thought about design that the hackers could find flaws in the distributed design. The three main concerns are establishing trust, detecting malicious trade behavior or patterns and limiting the damage, and knowing for certain that every node is who they say they are.

Digital signatures via GPG/PGP can allow for a web of trust and can allow you to know each node is who they say they are. A system where I vouch for you, someone vouched for me, and we all are continuously reviewed can produce a web of trust (just off the top of my head). It still doesn't prevent a team of hackers from conducting malicious trade behavior and I'm not an expert on trading platforms of the possible attacks against them.

I just know when we looked at what happened with mtGox we saw all kinds of unexplained trading behavior that didn't look just like a panic sell but looked algorithmic and coordinated. It didn't look very natural, and I don't know enough to know what natural trading behavior looks like. This is why we would at least need someone in here who is an experienced trader who has seen all the attacks that happen in Forex or Wallstreet and how those attacks have been defended against. The protocol would have to be modular enough that as new attacks are discovered a plugin can be written to manage or detect it similar to how Firefox and Chrome have extensions and the like. Scaling is important because if mtGox really did crash because of too many users and if Bitinstant really is forgetting to pay us our coins because of too much load then that scale and load is still going to be a problem if you decentralize it.

John even if we create a totally fucked up p2p exchange Im sure we cannot fail worse than MtGox
The concept that saves us is the p2p-protocol - at least our system will be less harmed by DDOS than the tradition bitcoin exchanges.

Enough words about this. I look forward to your perfect P2P exchange that runs without flaws from day one, has all the features people could want from day one, and won't have any scaling issues. In the meantime I'm going to look if I can concretely support some project. Sayonera.

Do you think in war that they just design the cheapest crap without doing any R&D or running any simulation on the small scale as a test to see if it can scale up? They test for years. We don't have years, I'm saying let's at least do some R&D and not rush something out just to get something out and have it attract a bunch of users and then collapse under it's own weight. That might be fine for music and movie sharing but when people are trading thousands of dollars they wont have that patience.

Just look at Bitinstant. Butterflylabs. All these companies doing exactly your method and look at the results? I don't want anymore amateurware but that is just me.

Tradecoin isn't a bad idea. Let tradecoin be part of the contest. Any idea should be considered and voted on.
Tradecoin seems to be a high level abstraction layer protocol similar to devcoin and the like. The problem is if you can't buy coins at all due to the exchanges being centralized and hacked what good is tradecoin?

I've given it further thought, we should do it where there are individuals/teams. We classify the teams with code names, like the red team, the blue team, or the viper team, or the chariot team, or whatever they want to call their team. We let each team come up with their own protocol but we as a community come up with the list of problems we want solved. We list those problems in plain English and the contest is about who can come up with the BEST solutions to these problems, the most efficient, user friendly, scaleable, etc.

As a community we can vote on the teams, reward the teams with points, etc. All of that could be done here and it can be tournament style or contest style. As far as what happens after that, once the winning team is found then we come up with a to-do list of what needs to be coded. We connect to each other via the app like astrid for instance. http://astrid.com/ Astrid allows individuals and teams to share lists. That could be to-do lists or task lists or whatever.

Beyond that github. Once a design is chosen then if people really want to save Bitcoin then they just have to code around that already designed protocol and follow certain rules in their coding so that it's well documented, and so well written etc.

Why should we create another alt-coin? I suggest we use existing bitcoins to reward the exchange-miners.

while your all talking we are already up just clone and go and build in more logic /apis/ code later

http://www.bteex.com/

see

https://bitcointalksearch.org/topic/annexchange-01-front-end-done-google-infrastucture-back-end-coders-lets-go-174117


about 100,000 of these should do it in the interim period

UPDATE GOX CRASHED AGAIN!!!AS SOON AS IT OPENED

http://www.bteex.com/

Google infrastructure is good

lets do it, cut out the bit exchanges problem solved

EDIT 1

Seriously a few good coders and there are api's google .gs scripts, javascript available and all sorts of integration, we can get something together, and en-mass clone it...local exchanges everywhere...and by local I mean in your street.

EDIT 2 design philosophy,

Ok a few layers are needed

[1] a more automated front end with scripts

order of (n) scaling
[2] A way for these to communicate with each other or some torrent like or rss feed or scaled data aggregate data

so feed data up to some sort of aggregate sheet or any number of sheets so that no one sheet is overloaded, eg 1 - 10 feed plus

10 exchanges feed up to one upper spread sheet they do calcs and 10 of those feed up to another upper sheet, this way the caac load is very distributed and easy to grow

Feed down same way 1 to 10 down......6 steps allows for 1 million exchanges each spread sheet only has to do 10 calcs

expand on these ideas

That's how innovation works. There is no way around it. It will suck in the beginning. People will find flaws in current systems, then improve on them, version after version, in smaller and bigger bursts.

The constraint area hasn't been mapped yet, let alone that we can know how an "optimal solution" would work. A "design competition", especially with so many non-technical people, is not going to solve that. Who is a judge of what is going to work anyway?

I am not saying that peer review is not good. It's good. That's why I posted links to current projects and collected some information. Why don't you peer-review bitcoinx and darkxchange? Criticize their mathematics?

I've had an idea for a while for a new coin. Tradecoin. Which are basically vouchers for a p2p stock market. btc / ltc would be on such stock market and the buying, selling, shorting, calls, puts, spreads ect would all be escroed by the "miners". The "ticker" price for any exchange values would be calculated and verified by the network. Miners would be paid by arbitrage within the network and or trading fees. I would also write a tradecoin constitution which would spell out the rules in the case of blockchain fork or other software glitches. Namely that the block chain is a voting / verification principle and that in the case of failure a new system would be created and shares/coins from the old system would transfer over from the most recent and verified copy of the blockchain. I also would propose an anomymous voting system which should be robust enough to handle even political elections and still ensure one vote per person. Perhaps one vote per share of the given asset on the network

The problem of course is that even though I am a programmer I find such a system to be a daunting task and quite the responsibilty. I prepose an agreed apon "pre mine" as bounty for the system.

This. All of these ideas skim over the most important part, getting fiat into the system. That requires trust, and that always leads back to a central point. Who do you give $10k to in a decentralized system? Unless you meet them in person and bring your baseball bat and some friends, no one. Our TRUSTLESS bitcoin network relies on TRUST based networks to give it value, so much for independence.