Hacker News Reader

Comment by darawk | Hacker News Reader

I think it may be that the decentralized control structure makes organizations who would normally be in competition with JPMorgan more willing to be on board with using something that they developed. For instance, if you are say, Goldman Sachs, would you want to use a private database at JPMorgan for you and your clients funds? Or would you prefer to use a semi-private blockchain, where although it's not open to the public, each participant in the network has equal stature to one another? I think that is the true innovation here, and I do think that it is important. It solves some of the corporate cooperation issues that prevent certain types of value from being created, because nobody wants to let their competitor control the space.

View on HN

This already exists. It’s called the DTCC and everybody uses it.

https://en.m.wikipedia.org/wiki/Depository_Trust_%26_Clearin...

Quoting Wikipedia: "Several companies have sued the DTCC, without success, over delivery failures in their stocks, alleging culpability for naked short selling. [...] Critics blame DTCC, noting that it is the organization in charge of the system where it happens, alleging that DTCC turns a blind eye to the problem, and complaining that the Securities and Exchange Commission (SEC) has not taken sufficient action against naked shorting"

Doesn't sound that trustless to me. With a decentralized setup as in Quorum at least everybody can calculate how much they have to trust the miners.

If your company has miners of their own, you are part of the whole system and not just using a third party service.

Most quorum implementations I have seen don't use mining for consensus. Out of the box you can choose between these consensus algorithms: https://github.com/jpmorganchase/quorum/wiki/Quorum-Consensu...

Thanks for the link. I knew there were other consensus algorithms possible but I didn't know of any specific example.

The main point still stands though, the different consensus algorithm allow you more insight than a simple centralized DB that is just a black box.

I agree very much. I have worked on quorum based blockchain systems for bank consortiums in the past.

And that company is exactly what this tech. replaces. That company does not provide its services for free. This blockchain product does.

> And that company is exactly what this tech. replaces. That company does not provide its services for free. This blockchain product does.

There’s no such thing as “free”. You’re either paying directly or indirectly. The DTCC is owned by its member organizations so it does not have a direct profit incentive anyway.

I highly suggest reading more on the topic and the problems it’s meant to solve before pitching “But blockchain!” as a solution.

It doesn't need to have a profit motive to cost more. It's costs are what cost more.

But there is no reason for it to cost more than a private blockchain solution.

Sure there is, simply by virtue of being the only solution available. All of a sudden, a simple database becomes a sprawling corporation with a multimillionaire board of directors, waging political war among themselves for resources, power and an ever larger number of underlings.

The private sector is not magically superior to a state bureaucracy, without any competition or political pressure to reform it can act even worse.

> a sprawling corporation with a multimillionaire board of directors, waging political war among themselves for resources, power and an ever larger number of underlings.

A blockchain does not get rid of those things.

Erm, yes it does. DTCC has a board of directors. It is a company. This blockchain eliminates that. Literally.

Can you explain? To me, it seems like control is what matters. And the DTCC controls access to its databases, whereas JPM controls access to its blockchain. Whether the parent organization is adding a +1 to a database or a +1 to cryptographic ledger seems inconsequential.

>JPM controls access to its blockchain

Only for the initial group of banks invited to participate. After that it would be up to the consensus algorithm to decide if a prospective member can join.

> DTCC has a board of directors

A large business "has" a lot of things. Why bring up board of directors specifically? I'm sorry but this sounds more like fumbling to grasp a point rather than an actual example of something useful.

I didn't bring it up, the comment you responded to did. You said that it didn't get rid of it, but it does. It gets rid of the entire DTCC company, and all of the salaries that go along with that. All of those costs go to zero.

I'm talking about competition, whatever the technical implementation. That's precisely the point, cost is determined by market structure and incentives, an inferior solution can be much cheaper due to lower margins.

I'm not sure I understand your point. Blockchains do not get rid of competition.

They don't but most competition does lower prices which is his point.

Who's pitching "but blockchain"?

While you show clear ridicule for some sort of "blockchain fad", I don't see anyone pitching a solution here.

I see JPM literally implementing one, and I would assume there is a cost reduction they've calculated.

Typically the cost reduction expected is due to eliminating rent seekers.

Which rarely happens, as new rentseekers replace old rent seekers.

It's not going to be free once they finish paying the army of developers each bank will have to keep on staff to manage it.

You mean like the army of developers required to run an ethereum node?

Or the fact that their coin / Quorum IS Ethereum, so most of the work will be done for free by core devs

You seem to be missing the point. Operating this blockchain is nearly free. Building it out is a capital cost. Once its built, its ongoing expenses should be low.

There's nothing to built... 1. `git clone ethereum jpmorgancoin` 2. profit

You two were actually saying the same thing.

How does this product operate for free? Who's paying for the miners?

You don't mine in a permissioned system like quorum. The CPU load is minimal, and when running the quorum clients, you are basically just checking to make sure everyone in the system agrees on the transactions. If someone tries to cheat (or spam or whatever), you kick them out.

But then we're back to square one, aren't you basically describing a distributed (but still perfetly "trustful") database? Isn't that effectively what this DTCC provides?

What in this allows JPMorgan to provide the same service for free? Is it really technical innovation or is it just a different business model? Wouldn't they be able to do the same thing without using anything cryptocurrency/blockchain related?

I feel like I'm missing something obvious but I don't see what it is.

A blockchain is a distributed non trustful database with a method for resolving disputes built in. Roughly speaking in something like quorum, everyone checks everyone else's work, and the majority consensus is ground truth. DTCC requires a trusted third party to administer the database. Quorum doesn't. Quorum could presumably also replace intra-bank transactions as well (instead of maintaining some other kind of database).

I guess the missing piece that seems to break people's brains; the ledger in the distributed database represents a dollar amount (in the case of quorum's dollar thing) and it has money-like rules, unlike, say, postgres which ultimately relies on some other piece of software to make ledger like transactions.

Unless I am mistaken, DTCC doesn’t provide any trust for internal account balances at JPMC. It only provides trust for the details of transfers. These tokens provide trust before and after transfers. The tokens themselves actually provide no trust during transfers and the transfers are only allowed by trusted parties as a result.

Also, DTCC only handles securities not currency so this doesn’t exist at all.

What actually you should be arguing is that the JP coin and the DTCC need to be combined as they each only provide trust in a certain part of the ecosystem.

Technology can be more than new things; it can also be made of combining old things in new ways, or finding new uses for old things, or even simply new ways of thinking. If we removed "cryptocurrency" and "blockchain" from the above, they'd still have a new settlements process. They're using a distributed ledger to make the process safely automateable and reduce audit costs.

what you're missing, that's obvious, is that blockchain is a solution looking for a problem where none exists. The word was imbued with virtue, and at this point blockchain has come to mean 'virtuous database'

If someone tries to cheat, you suspend their participation if you are empowered to do so, and / or report them to the appropriate regulatory bodies / law enforcement.

Blockchain doesn’t allow anyone to circumvent anything.

Unless it’s a perfectly spherical blockchain operating in a vacuum.

>If someone tries to cheat (or spam or whatever), you kick them out.

Or, if someone doesn't pay for access, you kick them out? Or if someone is your competitor, you kick them out?

It is a technology for trading with other banks.

If you want a chain that doesn't allow transaction censorship, you need something else, like a proof of stake or proof of work blockchain.

Permissioned blockchain generally doesn't have wasteful mining because all the participants are known. So the nodes just end up signing transactions. Double spend attempts can then be prosecuted.

Sorry, there may be small fees paid to the miners. But those fees will presumably be dramatically less than this current company charges.

You are putting a lot of weight on "presumably." Someone has to pay for all that mining. Or, you know, skip the mining and your costs should be lower.

They do skip the mining. Mining defends against Sybil attacks, which you don't have to worry about if you know who all the participants are.

Maybe JP Morgan thinks it's possible to improve on the DTCC. Your link describes some controversy over their handling of naked shorts, for example. Or maybe the new system would be cheaper, or JP Morgan wants to implement new features themselves, without having to convince everybody to trust JP Morgan to track their money.

The thing is that it can take days for a transaction between two different regional banks to clear. Whereas with the solution they're proposing, I'm guessing it'd take much less time than that.

Many people are trying to solve this problem (see: Corda, XRP)

There is no innovation here because JPMorgan controls access to the network thus they centrally control the network. No problems are solved here that couldn't be solved without a blockchain.

We need a new word for a thing that's like blockchain, but not happening in public, and therefore not actually decentralized.

I submit "charm bracelet".

> and therefore not actually decentralized

> The Quorum blockchain has been tested with other major companies including Goldman Sachs and the National Bank of Canada

It is sort of decentralized. but the access is not open to everyone. Most banks aren't comfortable with putting all their eggs in the basket that could end up under someone else's control. Decentralized doesn't really make the bank feel at ease when they back up their coin by trillions of their own money. They'll want some guaranteed form of control.

The purpose is actually to bypass SWIFT and their network with a network that is completely under the bank's control.

Maybe "base that stores data"?

DatabaseChain could be fun as well

D-Chain

Public and decentralized are not necessarily the same thing. Even if just two other banks join in, under an equal footing of ownership of the proof of work, its no longer centralized. I think assuming "all blockchain tech must be used only for situations with anonymous participants" is unimaginative gatekeeping. There is plenty of room to reimagine business process workflows between consenting participants, that doesnt at all hinder the public coin marketplace, or infringe or harm principled stands trying to change the world. These arent coins a single human bank customer would ever hold or use, these are for institutional settlement.

A good example could be a construction project where the general and all the subs collaborate on a database, but no one company owns it.

The smart part of JP Morgan's plan here is scalability. It might be centralized now, but it can be trivially extended into becoming decentralized at the flick of a switch, without a redesign.

Microsoft already has pushed for a non-public-anonymous-participant private blockchain that has throughput of around 1600 transactions per second. https://medium.com/coinmonks/coco-framework-a-game-changer-i...

That's nice to state, but it simply isn't true. JPMorgan controls access to the network, but once you're in the network, you're now on equal footing with them. If you want to, you and the other participants can fork away from JPMorgan and do your own thing. JPMorgan controls the network only in the sense that they initially extend invitations.

Can't they revoke the invitation whenever they feel like?

I suppose they could revoke access to their own node, the same way any Bitcoin miner can. But I don't think they can revoke access to the other nodes.

So JP Morgan would see some state for the accounts and the clients would see something else if they disagreed with JPM?

Doesn't seem like a viable consensus algorithm.

So you can prove to me that my account balance at JPMC is actually $100 when I do a wire or do you take JPMC’s word for it when you see the wire details?

Whether or not the validation will be public could mean you still have to trust JPMC but now you essentially get an API to validate my account balance which you didn’t have before

> you essentially get an API to validate my account balance which you didn’t have before

A blockchain is not needed for this

HTTP wasn’t needed for you to send that comment to hacker news, they could have used other technologies instead. What’s your point?

My point is that the advantages you described (an API) don't justify the complexity of blockchain technology since APIs can be implemented without them.

That’s a purely technical perspective on the situation. One could make the same argument about gRPC versus SOAP versus XML.

What would have been a more constructive response was:

This looks like JPMC is trying to profit off hype by building a new accounting and settlement system on blockchain. Since they control all aspects of their network and ultimately are the only one redeeming tokens for dollars, they could have used something simpler. However, I bet they did it this way because they think forking ethereum is cheaper than building a new system from scratch and also they might get a bump in stock price from the hype. If they built and announced the same system using bespoke tech (that likely solves the problem more effectively), no one would have probably cared.

I think maybe it wasn’t clear that I was responding to your comment about no innovation was made versus that blockchain was required. The innovation is that JPMC actually has an internal accounting and settlement system they can trust, not what database it uses or what language it was written in. Previously they relied on ACH and Wires which they didn’t control.

This explanation would have credibility if the announcement included BofA and Wells. Without multiple major players entering with them, it is only a wish.

I felt it was incredibly self-serving to make the token JPMC only. An arcade is only as good as the games inside. If your token only works at JPMC, that’s a pretty weak banking arcade.

That's not quite right though (and I fault their PR people for not explaining this). The token can be exchanged between any party on the blockchain (which afaik is not JPMC-only). You only need to transact through JPMC to exchange the token for USD.

“Unlike bitcoin, only big institutional clients of J.P. Morgan that have undergone regulatory checks, like corporations, banks and broker-dealers can use the tokens.”

JPMC still controls who is allowed to exchange tokens. Which means that you still have to transact through JPMC to exchange tokens as JPMC controls the blockchain.

If somehow they don’t retain majority control of the blockchain, then you technically don’t have to transact through JPMC but I don’t see that happening.

This is exactly correct. Private blockchains create efficiency between large institutions like JP Morgan and other banks who are constantly moving large volumes of money around by eliminating the traditional processes in place required to move that money. A private blockchain is essentially a shared database (distributed ledger) used by an arbitrary number of private institutions, each of whom has an interest in participating to reap efficiency gains by using a shared database that they can trust, more than they would trust a traditional database owned by a single party. It's not game changing innovation in the same way that crypto futurists think of Bitcoin as a new type of decentralized money, it's incremental efficiency for existing processes. It's more accurate to think about this as blockchain technology for JP Morgan than as cryptocurrency in the sense that Bitcoin is an application of blockchain technology. JPM coin is not like Bitcoin, or ripple for that matter.

The important thing to understand here is that JP Morgan thinks about blockchain technology completely differently than libertarian futurists. Blockchain is not a movement for them - it's a technology that reduces friction for things they are already doing.

> it's a technology that reduces friction for things they are already doing.

And there lies the rub, because as far as I can tell, this particular (and no particular) blockchain implementation does not in fact reduce any friction for things they are already doing.

I know people who have worked in this space (fintech permissioned blockchain). Although this was my initial reaction, they tell a different story. One angle to that is along the lines of "you wouldn't believe how bad the thing they use now is".

This is the aveneue of questioning I'd really like to see more detailed discussion on. [Disclosure: I'm agnostic on the potential of blockchain]

Does TPS increase non-trivially in a private chain vs a public chain? If not, will a private chain's TPS at least substantially outpace the TPS of, say, Ethereum within 3-5 years? That is, will JPM's private chain reach 100, or 1k, 10k, 100k, etc TPS?

Assuming TPS reaches a sufficiently high-level (whatever that could be), would a private chain require fewer sysadmin/engineering/etc. folk, compared to what the DTCC model requires?

Does a semi-private chain (where JPM allows other partners or competitors to plug-in) have a faster on-ramp? That is, would it be faster to plug players in than a traditional solution?

Is there a cost reduction somewhere within the audit process? Other processes?

I have these types of questions, but it is difficult to find much substantive discussion on them. Typically, answers are Yes or No, but with little to no detailed explanation as to why.

I assume they've set up an ethereum network with Proof of Authority instead of Proof of Work or Stake which means this isn't really decentralized since the transactions are being solely mined by a controlled group of trusted nodes. The five-engineer team I work on did the same over the course of the past six months and for us it's just a stop-gap until we can move to Proof of Stake, get some third party nodes on the network and finally consider ourselves decentralized as a blockchain.

That still doesn't need any technology derived from Bitcoin though. git branches and a git merge driver could do it.

It doesn't solve any of those issues. If one bank controls the infrastructure to mine and update, they can hard fork at any time. Why would another bank trust that system? The whole point of "distributed consensus" is to distribute the friggin consensus!

Liquid from Blockstream is actually doing what you're describing, but as fairly and openly is possible. The company itself doesn't have any operating nodes (JPMorgan has), Liquid is open source (https://github.com/Blockstream/liquid) and public, just the operating nodes are selected in a controlled, centralized, but geographically distributed way.

Another difference is that Liquid allows 2-way pegging to BTC, but allows other (for example fiat based) tokens to issued as well (but not part of Blockstream's business model), and the transaction types and values are encrypted.

Reply on news.ycombinator.com