My failed attempt at trying to explain BitShares and OL to C-CEX. Help needed.

[Akado]

We would either need them to use BitAssets or a common UIA like OPENBTC for example. Other than that it doesn't really change anything.

Now what if your reserve currency tokens were good in many other exchanges and you got transaction fees from everyone using your currency?  Transfers between member exchanges is a simple matter of exchanging IOU tokens, bypassing all the slow non-real-time blockchains out there.  And giving your customers another reason to do all their business with you.

In the current model the customer has to trust the centralized exchange but the exchange doesn't have to trust anybody.
They only credit a customer with an IOU once they've confirmed the actual alt-coins have been received in their wallet.

While I'm sure I've misunderstood something, does your paragraph above mean an exchange would have to redeem IOUs from other member exchanges?

If so that would create a situation where a Cryptsy default would fall on other member exchanges or worse a rogue member exchange could issue IOUs out of thin air and redeem them on other member exchanges?

No, each exchange has its own UIA and they compete on how credibly they back it.  Whichever exchange demonstrates the Exchange Backed Asset (EBA) with the most transparency stands to become a popular Reserve Currency.

Then users can trade against EBAs from other exchanges (perhaps at a slight discount or premium) if they want to exit/enter through that exchange to use any of its unique features.  Since BitShares lets you switch holdings in 3 seconds, you can spread out your risk using several of the EBAs or put your long term holdings into a counterparty free bitAsset if available.

The point is that something the exchange must do anyway (issue EBA IOUs for trading) can now become a new asset from which they can generate new revenue.  Especially if they innovate in how to prove credible, transparent backing.

Exchanges might even become a steady market for bitAssets if they chose a basket of them as a way to back their EBAs.

Then we won't have shared order books. We can have both but with the issue empirical mentioned, which I addressed on my post above.

[Stan]

We would either need them to use BitAssets or a common UIA like OPENBTC for example. Other than that it doesn't really change anything.

Now what if your reserve currency tokens were good in many other exchanges and you got transaction fees from everyone using your currency?  Transfers between member exchanges is a simple matter of exchanging IOU tokens, bypassing all the slow non-real-time blockchains out there.  And giving your customers another reason to do all their business with you.

In the current model the customer has to trust the centralized exchange but the exchange doesn't have to trust anybody.
They only credit a customer with an IOU once they've confirmed the actual alt-coins have been received in their wallet.

While I'm sure I've misunderstood something, does your paragraph above mean an exchange would have to redeem IOUs from other member exchanges?

If so that would create a situation where a Cryptsy default would fall on other member exchanges or worse a rogue member exchange could issue IOUs out of thin air and redeem them on other member exchanges?

No, each exchange has its own UIA and they compete on how credibly they back it.  Whichever exchange demonstrates the Exchange Backed Asset (EBA) with the most transparency stands to become a popular Reserve Currency.

Then users can trade against EBAs from other exchanges (perhaps at a slight discount or premium) if they want to exit/enter through that exchange to use any of its unique features.  Since BitShares lets you switch holdings in 3 seconds, you can spread out your risk using several of the EBAs or put your long term holdings into a counterparty free bitAsset if available.

The point is that something the exchange must do anyway (issue EBA IOUs for trading) can now become a new asset from which they can generate new revenue.  Especially if they innovate in how to prove credible, transparent backing.

Exchanges might even become a steady market for bitAssets if they chose a basket of them as a way to back their EBAs.

[Akado]

We would either need them to use BitAssets or a common UIA like OPENBTC for example. Other than that it doesn't really change anything.

Now what if your reserve currency tokens were good in many other exchanges and you got transaction fees from everyone using your currency?  Transfers between member exchanges is a simple matter of exchanging IOU tokens, bypassing all the slow non-real-time blockchains out there.  And giving your customers another reason to do all their business with you.

In the current model the customer has to trust the centralized exchange but the exchange doesn't have to trust anybody.
They only credit a customer with an IOU once they've confirmed the actual alt-coins have been received in their wallet.

While I'm sure I've misunderstood something, does your paragraph above mean an exchange would have to redeem IOUs from other member exchanges?

If so that would create a situation where a Cryptsy default would fall on other member exchanges or worse a rogue member exchange could issue IOUs out of thin air and redeem them on other member exchanges?

They shouldn't be able to issue assets out of thin air, because if they have a shared order book with the same asset, all services using that asset would need to have a weight on it to prevent that. I think that would be good in the sense that they would monitor each other too to avoid abuse, plus the would need everyone's permission.

As for the first point... you make a good case. Unlike you I only saw it from a user's pov, which means that for me if an exchange went rogue,I could still redeem my IOUs with another exchange meaning I would never loose my BTC, ever.

However, like you pointed, from an exchange's pov, it's a completely different view that I haven't thought about! I didn't give it too much thought but my first reaction upon reading that is it could be exploitable....

If we have 5 exchanges. All with 100 btc on their addresses and the equivalent on our chain, meaning 500 OPENBTC for example.
If one exchange goes rogue, there will be 100 extra BTC on the market that cannot be redeemed. I mean, they can but then we will see stuff like Gox and Cryptsy all over again which is what we're trying to avoid. I think we really should thought this through first since it's has the potential to jeopardize the entire DEX...

The only solution atm is having exchanges locking BTC in collateral... So when a client deposits BTC on their account, the exchange should use that btc to by the respective amount of BTS at lock it, maybe in a multi signature or account with permissions... Then we get to the same problem as bitAssets, what if the price of BTS goes way up or way down? That can mess up the collateral, that's why we need 200% to create bitAssets.... and now we're back to the same initial step, where exchanges might not be able to keep up with the collateral.

Even if they would place only 100% collateral, I doubt they would do all of that, it's too complex. Plus, how would they buy bts? They would need to use an external exchange for that LOL.

They could share a multisignature BTC, LTC, DOGE, etc accounts with other exchanges to prevent that... But again, I just don't see exchanges willing to do that. The only ones who might be willing to do this is either an exchange that sees the potential in this or is just a  very small one.  Then we would work our way to the top till we have many small exchanges...

There might be another way, but atm the only way I see it is having them sharing all wallets with multi-signature. However only small exchanges would accept that. No one will want other exchanges to give them permission to use their funds... but then again, it's not THEIR funds, it's not the exchange's funds. It's USER'S funds, so they should do this in good faith..

This is the only option I see atm. Otherwise we can't prevent an exchange from going away with user's funds. We need to turn the game around. Make the exchanges WANT to join us and not the other way around. Because we are the best option. That will make them "give us" something in exchange, which they really aren't because funds are not theirs, are from their users. Imo this argument makes sense, but exchanges can be just stubborn and refuse it. We just have to go with people who trust us, see potential in this project and are willing to provide the best services for their users.

This scheme would make it even easier to audit! For example if there are 1569 OPENBTC issued on the DEX then with a simple click they can check the multi account address and  check there are 1569 BTCs there. For the sake of transparency.



Edit: We also need that bug that rounds up the order values not letting you match other orders. This is essential. No exchange can operate or join us if we dont fix that.

[Empirical1.2]

We would either need them to use BitAssets or a common UIA like OPENBTC for example. Other than that it doesn't really change anything.

Now what if your reserve currency tokens were good in many other exchanges and you got transaction fees from everyone using your currency?  Transfers between member exchanges is a simple matter of exchanging IOU tokens, bypassing all the slow non-real-time blockchains out there.  And giving your customers another reason to do all their business with you.

In the current model the customer has to trust the centralized exchange but the exchange doesn't have to trust anybody.
They only credit a customer with an IOU once they've confirmed the actual alt-coins have been received in their wallet.

While I'm sure I've misunderstood something, does your paragraph above mean an exchange would have to redeem IOUs from other member exchanges?

If so that would create a situation where a Cryptsy default would fall on other member exchanges or worse a rogue member exchange could issue IOUs out of thin air and redeem them on other member exchanges?

[Stan]

My explanation of "why" was dismissed as "Don't use big marketing paragraph when you speak to developer".

That's what I wrote above "you drawn a huge pie", yes, it looks like a drawn pie, not a real one. 

Here's another way of explaining it in terms of our growing understanding of this particular exchange's Way of Thinking. I'd like to get your opinions on it before I forward it to him...

Since you are first and foremost a technical developer, why not work with us to design the world's first exchange network?
Help us optimize the design until it's too good to resist using yourself.  Here's a sample of what we've got so far: 

My question (as if I'm an exchange) would be: if you can't help us, why would we help you? Or say, they are less motivated if we say we need their help to build something, we should (be able to) say we already have something and if they join they'll be benefited (Of course they need some efforts to be able to join).

From a technical perspective, a decentralized exchange has access to the orders of all member exchanges and the ability to execute atomic trades via a single neutral trading engine resident on a shared blockchain.  Here’s the mother lode of  “simple” technical data:  http://docs.bitshares.org/bitshares/user/dex.html

As you pointed out, an exchange can emulate this using a bot talking via APIs to the trading engines of many exchanges, but race conditions and conflicts are inevitable.  On a common decentralized blockchain platform like BitShares, trades are finalized in one atomic 3-second block, no matter which exchange initiated the offer.  No race conditions are possible.

When you all place your orders under the control of the same neutral exchange engine, you get instant atomic arbitrage across all asset pairs. 

Then it gets interesting.

As an exchange you routinely issue your users what amounts to IOU-USD or IOU-BTC or IOU-DOGE “virtual-assets” to trade with on your exchange.  You take their real USD and BTC and DOGE at the door (just like always) and let them trade with your exchanges "IOU" assets like poker chips until it’s time to check out.  They trust your reputation that you will always reliably redeem their poker chips when it is time to cash out.  So your IOU assets are backed by your reputation. 

These paragraphs make sense. But maybe not so attractive.

Why not make your reputation a globally tradable commodity?  As it grows, more people prefer to hold your asset as their reserve currency.  And you earn fees every time they use your global reserve currency.

Now what if your reserve currency tokens were good in many other exchanges and you got transaction fees from everyone using your currency?  Transfers between member exchanges is a simple matter of exchanging IOU tokens, bypassing all the slow non-real-time blockchains out there.  And giving your customers another reason to do all their business with you.

These make less sense for me. It's not an exchange's core business.

As a customer, my assets can take advantage of arbitrage opportunities between member exchanges in a twinkling of an eye.  People taking the slow bot route get left in the dust or find that their assets are sitting on the wrong exchange when they need to move quickly.  Not so on the common real-time trading platform powered by BitShares.

Something wrong here. Tell these to users but not exchanges.
Cross-exchange business (or arbitrage opportunities) is not which an exchange would focus on, it's more attractive to traders.
With a shared order book there will be less arbitrage opportunities.

Thinking as an innovative developer, you have a unique opportunity to help work out the details of the world’s first network of exchanges. 

Right now, there are just two exchanges involved so far - CCEDK’s OpenLedger and BitShares DEX.  Several other small exchanges have expressed intent to join but are not yet members.  You could be the second mature exchange partner working with OpenLedger and BitShares to optimize the network interoperability to make the benefits to member exchanges more compelling and obvious. 

When we have fully optimized the design to where CCEDK and C-DEX are both happy with the increased synergistic services we can each offer our own customers, then the three of us can go rolling up all the small exchanges as new partners.  This will increase our collective apparent size, making membership more attractive to even bigger exchanges.

Like a rolling stone.

So, since you think like a developer looking for a new edge, why not work with us and help invent the only exchange network the world will ever need?  By the time the Big Exchanges wake up, they'll be asking to join us.

Yes "an innovative developer" may be attracted by these words, but a "conservative" developer would refuse them. I don't know whether you know you're talking to whom.


Here I listed some brief Q&A's, wish we can improve them together:

Questions from an exchange and answers to them:
1. I want more volume / users / profit / influence
--> shared order book

2. I don't want to send users elsewhere
--> self designed/hosted GUI (Website and/or mobile app)

3. I want more security
--> user data and trading data on blockchain, so don't need to worry about database-hack
--> (unavoidable) if host a GUI, need to make sure it's secure.

4. I want less risk
--> you don't need to issue IOUs which need to be guaranteed by yourself *ALONE*, other member exchanges will share your risk
--> (unavoidable) with BitShares you still have to manage the equities deposited by your customer

5. I want lower cost
--> less maintenance cost on market engine, databases if use BitShares blockchain and market engine

6. How to migrate to BitShares (required efforts/changes on my system)

7. The cost of migration?

8. Risks of migration?

9. Difference / Pros and cons between self-issued assets and shared assets?

Excellent input.  Thanks! 

[abit]

My explanation of "why" was dismissed as "Don't use big marketing paragraph when you speak to developer".

That's what I wrote above "you drawn a huge pie", yes, it looks like a drawn pie, not a real one. 

Here's another way of explaining it in terms of our growing understanding of this particular exchange's Way of Thinking. I'd like to get your opinions on it before I forward it to him...

Since you are first and foremost a technical developer, why not work with us to design the world's first exchange network?
Help us optimize the design until it's too good to resist using yourself.  Here's a sample of what we've got so far: 

My question (as if I'm an exchange) would be: if you can't help us, why would we help you? Or say, they are less motivated if we say we need their help to build something, we should (be able to) say we already have something and if they join they'll be benefited (Of course they need some efforts to be able to join).

From a technical perspective, a decentralized exchange has access to the orders of all member exchanges and the ability to execute atomic trades via a single neutral trading engine resident on a shared blockchain.  Here’s the mother lode of  “simple” technical data:  http://docs.bitshares.org/bitshares/user/dex.html

As you pointed out, an exchange can emulate this using a bot talking via APIs to the trading engines of many exchanges, but race conditions and conflicts are inevitable.  On a common decentralized blockchain platform like BitShares, trades are finalized in one atomic 3-second block, no matter which exchange initiated the offer.  No race conditions are possible.

When you all place your orders under the control of the same neutral exchange engine, you get instant atomic arbitrage across all asset pairs. 

Then it gets interesting.

As an exchange you routinely issue your users what amounts to IOU-USD or IOU-BTC or IOU-DOGE “virtual-assets” to trade with on your exchange.  You take their real USD and BTC and DOGE at the door (just like always) and let them trade with your exchanges "IOU" assets like poker chips until it’s time to check out.  They trust your reputation that you will always reliably redeem their poker chips when it is time to cash out.  So your IOU assets are backed by your reputation. 

These paragraphs make sense. But maybe not so attractive.

Why not make your reputation a globally tradable commodity?  As it grows, more people prefer to hold your asset as their reserve currency.  And you earn fees every time they use your global reserve currency.

Now what if your reserve currency tokens were good in many other exchanges and you got transaction fees from everyone using your currency?  Transfers between member exchanges is a simple matter of exchanging IOU tokens, bypassing all the slow non-real-time blockchains out there.  And giving your customers another reason to do all their business with you.

These make less sense for me. It's not an exchange's core business.

As a customer, my assets can take advantage of arbitrage opportunities between member exchanges in a twinkling of an eye.  People taking the slow bot route get left in the dust or find that their assets are sitting on the wrong exchange when they need to move quickly.  Not so on the common real-time trading platform powered by BitShares.

Something wrong here. Tell these to users but not exchanges.
Cross-exchange business (or arbitrage opportunities) is not which an exchange would focus on, it's more attractive to traders.
With a shared order book there will be less arbitrage opportunities.

Thinking as an innovative developer, you have a unique opportunity to help work out the details of the world’s first network of exchanges. 

Right now, there are just two exchanges involved so far - CCEDK’s OpenLedger and BitShares DEX.  Several other small exchanges have expressed intent to join but are not yet members.  You could be the second mature exchange partner working with OpenLedger and BitShares to optimize the network interoperability to make the benefits to member exchanges more compelling and obvious. 

When we have fully optimized the design to where CCEDK and C-DEX are both happy with the increased synergistic services we can each offer our own customers, then the three of us can go rolling up all the small exchanges as new partners.  This will increase our collective apparent size, making membership more attractive to even bigger exchanges.

Like a rolling stone.

So, since you think like a developer looking for a new edge, why not work with us and help invent the only exchange network the world will ever need?  By the time the Big Exchanges wake up, they'll be asking to join us.

Yes "an innovative developer" may be attracted by these words, but a "conservative" developer would refuse them. I don't know whether you know you're talking to whom.


Here I listed some brief Q&A's, wish we can improve them together:

Questions from an exchange and answers to them:
1. I want more volume / users / profit / influence
--> shared order book

2. I don't want to send users elsewhere
--> self designed/hosted GUI (Website and/or mobile app)

3. I want more security
--> user data and trading data on blockchain, so don't need to worry about database-hack
--> (unavoidable) if host a GUI, need to make sure it's secure.

4. I want less risk
--> you don't need to issue IOUs which need to be guaranteed by yourself *ALONE*, other member exchanges will share your risk
--> (unavoidable) with BitShares you still have to manage the equities deposited by your customer

5. I want lower cost
--> less maintenance cost on market engine, databases if use BitShares blockchain and market engine

6. How to migrate to BitShares (required efforts/changes on my system)

7. The cost of migration?

8. Risks of migration?

9. Difference / Pros and cons between self-issued assets and shared assets?

[Stan]

Here's another way of explaining it in terms of our growing understanding of this particular exchange's Way of Thinking. I'd like to get your opinions on it before I forward it to him...

Since you are first and foremost a technical developer, why not work with us to design the world's first exchange network?
Help us optimize the design until it's too good to resist using yourself.  Here's a sample of what we've got so far: 

From a technical perspective, a decentralized exchange has access to the orders of all member exchanges and the ability to execute atomic trades via a single neutral trading engine resident on a shared blockchain.  Here’s the mother lode of  “simple” technical data:  http://docs.bitshares.org/bitshares/user/dex.html

As you pointed out, an exchange can emulate this using a bot talking via APIs to the trading engines of many exchanges, but race conditions and conflicts are inevitable.  On a common decentralized blockchain platform like BitShares, trades are finalized in one atomic 3-second block, no matter which exchange initiated the offer.  No race conditions are possible.

When you all place your orders under the control of the same neutral exchange engine, you get instant atomic arbitrage across all asset pairs. 

Then it gets interesting.

As an exchange you routinely issue your users what amounts to IOU-USD or IOU-BTC or IOU-DOGE “virtual-assets” to trade with on your exchange.  You take their real USD and BTC and DOGE at the door (just like always) and let them trade with your exchanges "IOU" assets like poker chips until it’s time to check out.  They trust your reputation that you will always reliably redeem their poker chips when it is time to cash out.  So your IOU assets are backed by your reputation. 

Why not make your reputation a globally tradable commodity?  As it grows, more people prefer to hold your asset as their reserve currency.  And you earn fees every time they use your global reserve currency.

Now what if your reserve currency tokens were good in many other exchanges and you got transaction fees from everyone using your currency?  Transfers between member exchanges is a simple matter of exchanging IOU tokens, bypassing all the slow non-real-time blockchains out there.  And giving your customers another reason to do all their business with you.

As a customer, my assets can take advantage of arbitrage opportunities between member exchanges in a twinkling of an eye.  People taking the slow bot route get left in the dust or find that their assets are sitting on the wrong exchange when they need to move quickly.  Not so on the common real-time trading platform powered by BitShares.

Thinking as an innovative developer, you have a unique opportunity to help work out the details of the world’s first network of exchanges. 

Right now, there are just two exchanges involved so far - CCEDK’s OpenLedger and BitShares DEX.  Several other small exchanges have expressed intent to join but are not yet members.  You could be the second mature exchange partner working with OpenLedger and BitShares to optimize the network interoperability to make the benefits to member exchanges more compelling and obvious. 

When we have fully optimized the design to where CCEDK and C-DEX are both happy with the increased synergistic services we can each offer our own customers, then the three of us can go rolling up all the small exchanges as new partners.  This will increase our collective apparent size, making membership more attractive to even bigger exchanges.

Like a rolling stone.

So, since you think like a developer looking for a new edge, why not work with us and help invent the only exchange network the world will ever need?  By the time the Big Exchanges wake up, they'll be asking to join us.

[Stan]

thanks @Stan and I understand you're busy. Would you consider chatting with C-CEX on Skype? He even came back to me telling that no one has contacted him yet (I told him I would try to get someone who could explain things better). I can pm you or anyone else who would give this a go, his skype account.

I had a nice chat with him today. 
I have now been assigned more homework.

That said, I love how some of you guys have become proactive, rather than waiting on CNX as the huge bottleneck.  BitShares was always intended to be something decentralized that everybody works on.  It would appear that we had to take a big step back to make room for others to step up.  I'm looking for more ways to encourage that.

Nice! Did you manage to explain things better where I failed?
btw don't know if you saw it but a brazillian exchange claiming 100k users might approach you too, I provided your email

Nope.  You did just as good as I did. 

My explanation of "why" was dismissed as "Don't use big marketing paragraph when you speak to developer".

So, like you, I had to ask for a reprieve to go write a simple+technical summary of how it works.

It was graciously granted.

[Akado]

thanks @Stan and I understand you're busy. Would you consider chatting with C-CEX on Skype? He even came back to me telling that no one has contacted him yet (I told him I would try to get someone who could explain things better). I can pm you or anyone else who would give this a go, his skype account.

I had a nice chat with him today. 
I have now been assigned more homework.

That said, I love how some of you guys have become proactive, rather than waiting on CNX as the huge bottleneck.  BitShares was always intended to be something decentralized that everybody works on.  It would appear that we had to take a big step back to make room for others to step up.  I'm looking for more ways to encourage that.

Nice! Did you manage to explain things better where I failed?
btw don't know if you saw it but a brazillian exchange claiming 100k users might approach you too, I provided your email

[Stan]

thanks @Stan and I understand you're busy. Would you consider chatting with C-CEX on Skype? He even came back to me telling that no one has contacted him yet (I told him I would try to get someone who could explain things better). I can pm you or anyone else who would give this a go, his skype account.

I had a nice chat with him today. 
I have now been assigned more homework.

That said, I love how some of you guys have become proactive, rather than waiting on CNX as the huge bottleneck.  BitShares was always intended to be something decentralized that everybody works on.  It would appear that we had to take a big step back to make room for others to step up.  I'm looking for more ways to encourage that.

[Bhuz]

@Akado  I just want to second others opinion that you did a very good job.  Thanks for taking the initiative.

Absolutely! 

[Akado]

@Akado Imo it's better if you can talk to ccedk and metaexchange and blocktrades directly if you have chance , with some prepared information and questions (perhaps Stan has already tried but it doesn't matter). Perhaps they haven't much time to spend on the forum, so missed the discussion or unable to get the essentials of the discussion.

Thank you.

I think that's a good idea. I'm not convinced they would turn to OPENASSETS, they have their reasons. However it will be an opportunity to know those same reasons so in the future when we contact other exchanges and services and they use those arguments, we already have solid counter arguments against them 

@Akado  I just want to second others opinion that you did a very good job.  Thanks for taking the initiative.

@puppies thank you, I really appreciate it!

[puppies]

@Akado  I just want to second others opinion that you did a very good job.  Thanks for taking the initiative.

[abit]

@Akado Imo it's better if you can talk to ccedk and metaexchange and blocktrades directly if you have chance , with some prepared information and questions (perhaps Stan has already tried but it doesn't matter). Perhaps they haven't much time to spend on the forum, so missed the discussion or unable to get the essentials of the discussion.

Thank you.

[Akado]

I think OPENASSETS have a chance. Obviously every exchange decides what it's best for their business model. But from a user pov shared orderbooks could be very interesting. If we at some point get bigger I'm sure we could get smaller exchanges to use them.

So why not blocktrades and metaexchange join OPENASSETS? They are competing for the first bite. I think it's best if we start with this, try to let them combined as soon as possible.

Well they could use OPENASSETS but I guess that's up to them and their business model. I really think since they're bridges they should adopt it as it gives people a more easier opportunity to come in and out of the system, it makes liquidity easier imo.

Right now you need to go to metaex and get metabtc on the DEX then if you want to trade against more stuff you might need to trade to bts and then to openassets or something like that. If all used the same assets, it would be easier and be better for the order books. But they have their own reasons..

@Shentist could you fill us in? I'm sure you have your own strategy, but on the other side, dont you agree it would be easier to get in and out of the system like this?

In the end it's up to each business, I'm sure we'll have some who prefer to issue their UIA, other might prefer to use OPENASSET. Everyone will have their reasons, we just try to convince new ones joining to share the same asset if that's something we thing benefits us.

I think the secret is getting the smaller fish. We get some more volume first and then we get them. They're the ones who have the reasons to join us. Like Stan said. If we see this from a small exchange pov, they're joining a network and instantaneously get a bigger amount of volume and they can stay competitive with other bigger exchanges. So instead of having 10 small exchanges working separately with $2k volume each, you can have those same "small" exchanges have $20k volume instead.

@ccedk would you be willing to share permission with other exchanges if they used OPENASSETS? Assuming it's feasible of course. I guess they would  want that as some kind of insurance so they don't get played.

The idea with the bond markets would create a very interesting dynamic, the only problem is they can't be sure there are enough people providing enough collateral at all times, which they need otherwise they risk not being able to keep up with deposits. If one user decides to deposit 100 BTC image an exchange saying, "sorry we can't accept it, that's too much for us." That would be terrible