Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Own your own Bitcoin Exchange (bex.io)
58 points by snitko on April 14, 2013 | hide | past | favorite | 41 comments


What is the appeal of this when the rates are so high? The fees on most exchanges are in the .4%-.6% range for low volume. If I tried to run an exchange and had to pay a 3rd party 1% of all transactions, I'd be running at a loss.

Furthermore, you can pretty much already do this by buying/selling on localbitcoins.com[1]. They provide an online wallet, escrow service, and also charge a 1%, but they don't charge a monthly fee.

[1] https://localbitcoins.com/cash_exchange_howto


I almost don't want to mention this, but it could be a "selling shovels" business model. Since they keep their $199 whether your exchange succeeds or fails, it looks like they are offloading all risk onto their resellers.


By hosting the platform, are they trying to avoid licensing/regulation? After all, they are not providing money transfer services, but are providing a platform for others to do it.

I looked into legally building an exchange in Texas, but I would have to get licensed as a money transmission company or a currency exchange company. The fees were outrageous! $2,500 application fee, $300,000+ security deposit and requires the applicant to have over $500K net worth.


This...

The service seems to imply a business model where there are lots of little operators would open an exchange if they weren't so capital constrained such that they can't roll their own.

But how many with 500k to play around with are going to then want to pay 1 percent? Even if there are a bunch of people just waiting to sign up - then the competition is going to see their margins evaporate - with no air to breathe over and above that 1%. What else are they going to compete on besides price?

Either that or I just don't understand this business model at all. Best of luck to them though.


Maybe they won't target US customers.


>I looked into legally building an exchange in Texas, but I would have to get licensed as a money transmission company or a currency exchange company. The fees were outrageous! $2,500 application fee, $300,000+ security deposit and requires the applicant to have over $500K net worth.

Doesn't it cost $30 million to get licenses in all the states that required it?

That certainly seems like a broken system. I'm not sure why there isn't a federal law or an agreement between states that one license can be used across states.

I've read that laws like this exist because of lobbying done by Western Union.

Contrasting with the EU. One license in any EU state means you can provide services across the EU.


Even though it is licensed to allow it, not giving credit to the creator of the hero image on this site is a little unkind.

http://redd.it/1c4er5


At the bottom, in the footer, they give credit...

"Background image is a fantastic work by cubrbeat, licensed under CC-BY-SA"


That wasn't there when I made the comment.


There's an opensource exchange being built, albeit with a terrible name (IMO)

https://github.com/buttercoin/buttercoin


I don't understand why they want to build a high-performance exchange engine, in Node.js. Even I wouldn't choose my language of choice, but would rather look at either Java, C/C++, hell even Go or Erlang!


They're pretty clueless.


I agree, but yea, it's there.


Not only that but the people building it have no idea what they're getting into. I tried to provoke them into a discussion of getting some design/spec docs for the matching engine online and everyone insisted it was "obvious" until I pushed them for 20 mins.


I don't understand the point of having a bunch of little exchanges running on the same backend. Why not just have one?


Agreed... an exchange is more useful the more volume it has.


Although this might not be the right way to do it, I think the goal of the project is to exemplify bitcoin's "distributedness" and minimize the reliance on a single exchange (ie MtGox).


Several exchanges running on the same backend aren't that distributed; there are many scenarios where they would all go down at the same time.


I think the problem isn't distributivity but decentralisation.


"... our technology is secured by state of the art techniques such as OS-less deployment, eliminating a large variety of exploits."

OS-less deployment? Are they using Erlang on Xen?

I'm just guessing since I know Yurii is an Elixir (a flavour of Erlang) core developer.


"Yurii is a software craftsmanship enthusiast with 15 years of professional experience in the industry. Noted speaker in the Erlang community, a core contributor to the Elixir programming language and an eager explorer of distributed systems."


Could be Google AppEngine, or the Node.JS on Google chrome Minimum +V8. (not that I know many people using that). Or it runs as a bot net :-)

Sound like the kind of line you put on your home page to make it to the top of hacker news.



Halvm doesn't (yet!) support recent ghc versions. For high concurrency workloads the ghc rts recently got some nice improvements merged in that will make the mega scale concurrency performance story even more awesome once ghc 7.8 gets released sometime this fall


If you can do so much, why not build a bitcoin exchange directly? It's not like currency exchanges are that much a commodity for you to sell. Mmm... aren't you a bit like making excuses for not trying? ;)

No seriously, if you made a bitcoin exchange site with the same copy, that's the one I would use. Well. If I wanted to buy Bitcoin again. You know, a superhardened style, paranoid exchange. Maybe with its paranoid wallet too, dunno'. So do it.


>If you can do so much, why not build a bitcoin exchange directly?

Because the margins are awful unless you are doing huge volume.


Do I still need to get licensed as an authorized money-transmitter? Unless you're getting me around that monumental obstacle, you're solving the wrong problem.


But it's bitcoin! ...laws don't apply! Jump right in!


Where does it say that bitcoins are money?



This seems incredibly unnecessary, especially since I might actually use their exchange if they ran one. (More design chops than most of the abhorrent exchanges.)

Though an alert() is a rather inelegant confirmation...


Not sure I quite understand... Is this so the kids on the corner can sell bitcoins for cash instead of lemonaide?


I'm afraid this kind of scenario is doomed to failure too. The issue with individual exchanges has been a lack of standardized security and hardware configuration. Many exchanges get broken into and bitcoins get stolen.

The issue with something like bex.io is that now we have many different exchanges, but all running on the same provider. Suppose that bex.io is able to put the best security around their systems. It still will possibly only take one vulnerability to get inside hundreds or thousands of exchanges.

Further, instead of DDoSing one or two exchanges like mtgox and btc-e but leaving the rest of the ecosystem free for trading, it only takes one massive DDoS against bex.io in order to take down hundreds or thousands of exchanges in one go.

Bad, bad, bad.

What's needed is both security and distribution. Someone else mentioned a node.js exchange product that's popped up in the last day or so, https://news.ycombinator.com/item?id=5545271. Its called Buttercoin, https://github.com/buttercoin/buttercoin.

The language its written in isn't as important as the fact that its open source. If we get an environment of several open-source exchange engines that can be hardened by their respective communities, we get security. (Yes, individual instances are susceptible to bad server configs...).

By having relatively easy to setup OSS exchange systems, we get distribution.

People have complained that mtgox is a problem because its the largest extant exchange. At least its a single point of failure that only affects btc and money on its systems. Something like bex.io is going to lead to catastrophic losses across any number of customers as soon as they start becoming a target for nefarious people.


DOSS attacks can be overcome with a solid infrastructure like Cloudflare (expensive but possible) and also with the help of modern browser capabilities such as webRTC and offline cache.


Services like cloudflare do nothing for API's, which is the core part of any exchange system. Browser technologies have nothing to do with building out secure, distributed trading systems.


You're mistaken about that. APIs cannot benefit from something like CloudFlare. You need the connection to be live to the backend at all times, which isn't how proxy caches work. They take over serving content when the load is high.


"We do the tech. You do the rest."

... So, my job is to press the button that says go?

After i pay the 200/800$ off course.


High exchange fees. Little info about what it is built on. Suspect timing (and likely too short of a build time) Likely to have regulation issues in the near future especially if lots of people do it.

All adds up to not how I would do it.


I've read over it twice and I'm convinced it's a joke, like MailAppApp. http://visualidiot.com/articles/mailappapp


I'm getting tired of this website design style already.


"Easy is not easy enough." Uh huh. Talk sexy to me baby. So now I've gotta buy a bitcoin exchange too?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: