Topic: The next Mt Gox? (MtGox 2.0) (Read 15269 times)

There is more weed than there are Bitcoin people...http://www.meetup.com/bitcoin-denver/

Let's smoke

Colorado? I'm in Colorado... lets talk!

Anything is possible - with the right business plan, contacts, and $$$ - it IS possible

Agreed with previous poster.  It is TOO capital intensive to launch a site like mt gox.  It would cost millions and millions just to get through US law, let alone the entire free world's regulations...  
https://bitcointalksearch.org/topic/trading-with-non-license-money-transmitters-aka-lending-btc-to-anyone-215951

Well, first and foremost: follow the laws.  People talk about MtGox like it's a big government crackdown, but I think the only reasons that they're cracking down is that a) bitcoins are getting popular and b) they didn't follow the laws for creating an exchange like that.  That's going to be an issue for any exchange that deals with the US currency.  If you want to be legit and not get shut down, this is a must.

After that's taken care of, I pretty much agree with Viceroy on the issues needed.  Secure and scalable are the key points.  There was an article in one of my business magazines recently about Netflix using a third of all the internet traffic during much of the time to stream videos, so scalability is definitely possible.  How do they do it?  Amazon web services.  I think this, or something similar, is the key to a successful exchange. 

AWS is a great system because from what I've heard it's basically infinitely scalable, you need more resources, simply allocate more to it.  But it's not trivial either, whoever runs this system would likely need a sysadmin or several to keep it running.  If the person / people running the exchange is willing to put the time and money into figuring out AWS, however, I think they have a real good shot at making a system that's much more scalable and stable than MtGox.

I'm about to launch a new exchange, but it's for China. I was going to encourage USD IN too, but not any more -- at least not at first.

I think the big part missing from most current exchanges is real scaleability; I don't see many properly utilising cloud infrastructure, or parallelising operations.

Having looked at a few existing trading engines when I was getting started, and not being terribly impressed, I have the feeling that there are a few crufty old lines of code still doing the rounds on other exchanges.

This is the best example I found, and it appears a few exchanges adopted the code...

... lock all database tables
... match all possible orders
... Hope nothing goes wrong. If it does, just print an error.
... unlock all tables
... sleep for a bit
... repeat

Sigh.

Please, we need your help more than ever.  If you have anything useful to add, please do so now.

MAJOR NEWS:

http://www.zdnet.com/us-department-of-homeland-security-issues-court-order-against-mt-gox-7000015389/

Wow, human ingenuity, a community of like-minded people, there's no stopping progress! It amazes me that people can get together on something as trivial as Bitcoin (come on, have you seen what some people say about it? How divided the opinions?), yet we continue to wage wars, kill innocents...

Keep it up!

Back on track, I think the most important thing is low fees and a stable exchange rate. Really that only happens with regulation and central authority, which is the complete opposite of what Bitcoin stands for, so I think we have to rethink this!

SEPA shouldn't take more than 3 business days.

The reason why Mt.Gox has such massive delays (not days but often weeks) is because their accounts have limits and having a lot more customers now, their withdrawal orders are queued.

But everyone knows that a wire transfer should take around 2-5 days, and there are a lot of people saying it took more than 20 days, almost a month to receive their money.

If a sepa is anything like a wire transfer it WILL take days.  Not much you can do to expedite that, I don't think...

What is a better way to get FIAT into or out of bitcoin?

The international or sepa withdrawals should be more fast, they take a lot of days to get processed, that is ridiculous.

Per your suggestion I just added this to the OP:

5) More, better and faster withdrawal methods:
 a) wire transfer takes days
 b) instant cash deposit to my account at a major bank?
 c) sepa withdrawl (European Only)
 d) something like a money pak code?
 e) overnight precious metals delivery?

It will be realy interesting to see what they roll out.

As far as I can see bitcoin exchanges have two major unsolved problems with a huge potential for improvement:

* Actually being able to deposit money frictionlessly
* Getting shut-down by governments / banks

These two need addressing before anything else. In an ideal world, the exchange would be located in a part of the world sympathetic to bitcoin but at the same time, not somewhere associated with scams (Nigeria, for example) because the public has to feel confident enough to entrust real money.

Is there any scope for actually setting up some kind of genuine regulated bank of bitcoin upon which to form the basis of any exchange infrastructure?

Cheers, Paul.

Mt Gox are doing Litecoin shortly aren't they? Or is that not what you meant.

On BTC-E the new currencies have generally tanked when they've been added, but Mt Gox may be different.

One massive exchange medium for the major crypto currencies is an awesome idea. Just think if Mt.Gox picked up LiteCoin, millionaires over night.

Let me put myself out there, as well: I have a lot of love for bitcoin and I've been a professional software developer for well over 10 years.

If you find some investment for this project I would be interested in working with you because I believe the market is crying out for a better exchange than MtGox.

Cheers, Paul.

Would it be possible to design a p2p based exchange?  It would be in the spirit of Bitcoin, decentralized, and resistant to DDOS.  I'm not sure of the technical details it would take to design, but a basic idea would be similar to how p2pool keeps track of stats and payouts.  The client would display current market data.  You could submit trades and the network would process them.  When you wanted to cash out you would have to actually trade cash for bitcoin with an individual or an entity, but while trading you could have wallets like the exchanges have that keep track of your Bitcoin and USD.  Not sure if it would work, but it would pretty cool.

PB

The transaction fee schedule should be more intelligent than MtGox - their fixed percentage of value charge means they make *more* money the higher the price of bitcoin, which is a horrible conflict of interests.

The way to do it might be something like a fixed spread on top of order-book entries like the ECN Forex brokers have. Maybe.

Cheers, Paul.

IMHO, it should be an exchange much like the NYSE, NASDAQ or AMEX is and allow for all crypto currencies, the underlying trading engine is the same for all so why limit ourselves?

I disagree, NYSE, AMEX, NASDAQ all do this for pending news, if it is good news then the price may go up. Market Stops *should* definitely be put in place to avoid things like last weeks mini crash and the 2-3 previous to that, as well as some of you may remember the crash of 87

Agreed

Agreed; we can make something better.

Agreed

I disagree that we cannot do this. You are thinking short-term here. The market has proven itself with the nerd/geek class, by the time it hits main stream it will require a truly robust world class exchange. Actually to make it the main stream we should do this now so that it will make it easier for joe and jane to trade just like stocks, etf's etc are.

Hi All,

Should this new exchange take Litecoin and other Crypto Currencies?

K

 

Mtgox is the most trusted exchange right now due to them swimming in commission fees.  If someone has the fund and build a trade platform that provides peace of mine for traders (ie deposit insurance), that someone will be very successful in the long run.

The only reason that I use mtgox is that I know they are swimming in commission fees and so they have no incentive to close down shop and screw their customers.  They have also made enough money to cover potential losses without closing shop.

If another exchange can provide account insurance or just the illusion of safety, that exchange will take most of mtgox's customers. I doubt that any who has used mtgox will stick with them if another exchange can offer the same level of perceived safety.

1) Implement Halt Procedure - I don't think this helps, and it is actually harmful in some instances for a number of reasons.
  a) in the event of news
  b) in the event of mother nature <--- Maybe this one.
  c) in the event of extraordinary market activity

2) Highly Secure - +1
  a) even Nasdaq can be taken down
  b) site must be able to thwart ddos
  c) client money and btc must be secured
  d) can deposits be insured?

3) Scaleable +1
  a) does not fail upon large order queue
  b) multi-server system for redundancy

4) Fast +1
  a) upper limit 10,000 transactions per second?
  b) lowest limit 1,000 transactions per second?

5) API? +1

6) how about Deposit Insurance, like the FDIC? +1 I am not sure this would be possible without huge fees though, something like 3%.

I would happily run this in Colorado USA without fear of the FED.  There is nothing illegal about buying or selling or mining bitcoin in America.  The process of becoming a "money transmitter" which allows a company to trade in cash is relatively straight forward and quite inexpensive.  

Chad is a buddy who *might* fund it.  Do you have a specific proposal to create the beast?  As I say, I'm happy to host it and run it IN AMERICA.

Not sure how to promote it any more than this:
http://www.reddit.com/r/Buttercoin/comments/1d134o/tired_of_buttercoin_lets_start_again_with_a_team/

and in the comments here:
http://www.huffingtonpost.com/max-keiser/bitcoins-problem-with-liquidity_b_3140557.html

Max Keiser doesn't know that much about how exchanges operate. But, what's the question? who is chad? 


The guys at Buttercoin don't have clue. They might be decent computer scientists, but now they've gone dark, which is no surprise. Make an announcement, create some buzz, then find some guys with $$$...

The risks in terms of legal are huge IMO. If things get though a few people might go to jail, see Wikileaks & Megaupload.

>http://www.huffingtonpost.com/max-keiser/bitcoins-problem-with-liquidity_b_3140557.html
>Have you seen an bitcoin exchange for sale?
>Chad

>Research to build the best trading platform out there and I will
>consider funding it.
>Thanks
>Chad




Anyone got a proposal?  

Did anyone mention that from the 4 big exchanges, 2 had their accounts frozen in the last 2 weeks? Imagine if the network reaches 10x or 100x the size - then the governments of the world will probably attack bitcoin. Look at how quickly Visa/Mastercard/Paypal shutdown access to Wikileaks funds. Wikileaks was not a problem for the governments, before they perceived it to be a threat. I think these issues are more pressing  than just the volatility.

Before I would open an exchange I would think twice, depending on the jurisdiction. Iceland anyone? I wonder what Bitcoin could do in countries like Argentina, Zimbabwe, etc.

With regards to deposit insurance. Interesting thought. FDIC is kind of a joke, as the central banks are really the insurers. Banks, centralbanks and governments are much closer related than people think. Central banks are forced to insure banks and governments. So they increase the money supply, the whole point of which is to transfer funds from one sector to another.

I think instead of building towards a set goal (eg 10000 transactions/second), the goal should be scalability - constructing a distributed engine which will run on multiple servers, so that as new users sign up or volume increases more servers can simply be added. High throughput, efficiency and reliability should be prioritised. In addition, MtGox complained of an influx of new users contributing to the massive lag during the crash. MtGox needs to be aware of the approximate server capacity, and if they can't upgrade quickly enough to handle new registrations, then they should impose a registration delay/waiting list when necessary in order to prepare new servers.

I also agree that security and insurance are high priorities. I personally haven't been affected, and I'm sure that I'm very well protected via two-factor authentication. Nevertheless, the topics here about losing hundreds, even thousands of dollars worth of BTC due to hacks/attacks are worrying, and the community can't just blame people for being lax with their own security: Yes, if you leave an unencrypted wallet on a windows machine with no antivirus or firewall, you're responsible for any losses that result, but an exchange should be a secure place to store and trade money. If a browser exploit or a virus or something infects a user and transacts money from an exchange account, the user should be compensated. At the very least, new safeguards should be used, such as manual/delayed confirmation of any non-reversible withdrawals by a user. Receiving a deposit without verification is fine, no one would object to an unexpected donation But if you wish to send BTC from one address to another then there should be an enforced delay, checks for suspicious activity, and confirmation (e.g. by phone or other). Bitcoins already often take upwards of one hour to transfer due to the enforcement of 6 confirmations before most services accept the payment, so a little longer for added peace of mind would be acceptable and safer. (There are probably people who would object, and would want their transfers ASAP, to them it may be possible to have an override in the account settings to forego any checks/confirmations, but this would likely result in MtGox renouncing all liabilities for any losses). With the addition of these checks, losses would be reduced, however MtGox would also become liable for compensating users whose accounts have been hacked, but I think this would be a positive step for the bitcoin community and help it to gain wider acceptance.

Basically, if this new trading platform is to have any chance to succeed, if it's worth doing, it's worth over-doing. That means some substantial money needs to be thrown down on network security and anti-DDoS protection, as well as quite literally tens of thousands for server hardware alone. Their system needs to, as mentioned be able to handle absolutely insane transaction/second and site visits without flinching. This shouldn't be that hard, seriously. The bitcoin network is a t 817 PetaFLOPS. We should be able to get a decent system going.

Never halt because of news, this just causes more panic.

Never get lag, the engine must run on many servers alone, perhaps AWS has performance product they can offer you?

New York stock exchange handles more trades in a minute than Gox has ever handled in its life. Bitcoin exchanges are the most terrible thing in the world, they really are trite.

So you need to be designed to handle 10,000 trades a second, if you can't then you just going to be another Gox.

(I don't think you can do this though, there isn't enough money in Bitcoin trades yet to warrant someone designing a real trading platform that can handle real traffic, I would love to proved wrong)

We all know what happened during the crash, Mt Gox (and everything else) failed us.  My sell order at 240 was ignored for what seemed to be hours, eventually I cancelled and held my position awaiting another run up.  At this instant I, and you, realized that Mt Gox is a steaming pile of shit.  So what do you require in Mt Gox 2.0?  


Click here to read a whitepaper about: A COMPREHENSIVE MODEL OF DESIGN PRINCIPLES FOR ONLINE STOCK TRADING SITES (Adobe PDF File).


Initial thoughts:

0) Establish and STAND BY a KYC and AML directive.
  a) added in light of Mt Gox Asset Seizure on 05/15/2013
  b) establish and follow a know your customer directive per US law
  c) establish and follow an Anti Money Laundering directive per US law

1) Implement Halt Procedure
  a) in the event of news
  b) in the event of mother nature
  c) in the event of extraordinary market activity

2) Highly Secure
  a) even Nasdaq can be taken down
  b) site must be able to thwart ddos
  c) client money and btc must be secured
  d) can deposits be insured?

3) Scaleable
  a) does not fail upon large order queue
  b) multi-server system for redundancy

4) Fast
  a) upper limit 10,000 transactions per second?
  b) lowest limit 1,000 transactions per second?

5) More, better and faster withdrawal methods:
 a) wire transfer takes days
 b) instant cash deposit to my account at a major bank
 c) sepa withdrawl (European Only)
 d) something like a money pak code?
 e) overnight precious metals delivery?

6) API

7) how about Deposit Insurance, like the FDIC?

What would you like to see?  Am I wrong on the above?


FAQ:

What about Buttercoin?

Quick assessment from a forum post: "From the docs I've read so far, it seems like Buttercoin is only an attempt to address the performance and scale-ability of exchanges. It does not appear to fundamentally change the traditional role of exchanges.
It seems that what is being proposed is just an open source exchange platform that anyone can host and run, but the average trader will mostly likely still trade on one of these closed, corporate-controlled exchanges. Am I wrong?"

Comments: