Pan Chao: The clearing system and blockchain network behind JP Morgan Chase

The issue of the implementation of the stable currency by JPMorgan Chase has sparked a wide-ranging discussion in the circle. Among them, the "JPM Coin Trilogy" by Maker Dao economic researcher and head of China, Pan Chao, is the most detailed and profound. Recently, we have the privilege of inviting Pan Chao to be the "carbon talk" offline salon for the value of the carbon chain. In this salon, Pan Chao analyzed the difference between JP Morgan Chase's stable currency and our common stable currency (such as USDT), and what kind of consensus mechanism JP Morgan Chase might adopt. What are the main performance issues? The following is the full text of the carbon chain value:

01 The birth of Morgan coins

What is Morgan Dollar? In the simplest terms, a Morgan coin equals one dollar and looks like the normal stable currency. But they are not targeting the same market. USDT, USDC, and GUSD are retail stable coins that can be used by general users. Investors can also buy on the exchange, do daily transactions, or exchange with legal currency. But the object of Morgan Coin is not a regular user, but a wholesale bank. Ordinary people can't get in touch, and there won't be any exchanges that can get this currency. Morgan coins are used for real-time large-value settlement, which can be called clearing vouchers.

JPMorgan’s member banks deposit dollars in a designated account and generate Morgan coins in 1:1. The Morgan currency can then be returned back to get a one-to-one dollar. This seems to be the same as the generation and redemption of the USDT, but it is not. As I just said, Morgan Coin is used for real-time large-value settlement.

Why does a large amount of real-time fund settlement require such a thing? First, you need to understand how funds are transferred between banks.

– When there is only one bank, assuming a 10 dollar transaction between the two, then directly record Alice to reduce 10 dollars, Bob increase 10 dollars.

– If there are two banks, for example, I am at ICBC, if you are at CCB, if I want to pay you ten dollars, I need to tell ICBC that I am going to give you ten dollars. ICBC must reach an agreement with CCB to establish a relationship account. In this relationship account, both ICBC and CCB will deposit a little money in this account beforehand. When the two parties trade, there is no need to settle a settlement between the two parties. And you can settle directly at this bank level.

But such a model can also cause problems. Relationship accounts are more convenient when there are only two banks, but if the number of banks increases, a country may have tens of thousands of banks, then if such an account is established between every two banks, it will Is an exponential relationship.

So how to solve such a problem? There are two modes in clearing the settlement, one is called real-time big settlement, and the other is called delayed settlement. For delayed settlement, different banks are only responsible for billing and do not need to settle immediately. Just remember, for example, how much A owes B, how much B owes C, and then look at the net at the end of the day. For example, when we use WeChat to make a payment, it is only a payment link, and it will not be settled at the bank level until after a period of time.

But there is a problem with delayed settlement. The transaction can be withdrawn in the delayed settlement. If the other party pays a very large amount, it often requires real-time settlement, which is the finality of the transaction. Because the cost of revoking funds cannot be afforded. Banks also run the risk of bankruptcy and cannot place a relationship account in every bank.

There is a solution, this way is to return to the original banking world, relying on a very large public ledger. What this book does is responsible for large-scale real-time settlement, which is one of the reasons why there is a central bank. The central bank acts as an authority and everyone will believe this book. JP Morgan Chase first played the role of the central bank in the United States, which explains why Morgan coins will appear. When it comes to large-value payments, Bank A and Bank B use the JP Morgan Chase network as a big pool. Bank A and Bank B can put their deposits in a large pool and then generate a 1:1 Morgan currency. For instant billing.

02 What is the difference between the birth of Morgan coins and the blockchain of Morgan?

So why does this network need to be based on a blockchain?

The reason why JPMorgan used the blockchain is not because of transparency, decentralization, etc. The characteristics of the traditional public chain are the Bug for JP Morgan Chase.

So what are the characteristics of the blockchain books required for traditional financial institutions?

The first point is privacy. Financial institutions need a high degree of privacy. It is not possible to disclose all the books to the public, otherwise it will lead to run-off risks and liquidity risks. At the same time, the bank does not want to let the rival bank see the situation on its own account, but also protect the user's own property information. This is basically impossible in the public chain, especially the Ethereum network, because as long as you open any Ethereum browser, you can see all the information of the account, the balance, and the object of each transaction. If traditional financial institutions use blockchain research and development, they must solve the privacy problem.

The second point is the need for high-performance blockchain. Although the large amount of money transfer is not so urgent, it is difficult to meet the national or large-scale financial application.

The third point is final. The final thing is that the deal was sent to you from my account, then I can't change the deal, just like cash. I will give you one hundred dollars. If you take this one hundred dollars, it means the end of the transaction. Instead of transactions that can be withdrawn, such as in WeChat transfers and Paypal transactions, they can be withdrawn. Many blockchains are not final because of the probability of bifurcation. When the ETA event was attacked, a fork and rollback was carried out. The hacker attack and subsequent transactions were not recognized, and traditional financial institutions could not accept it.

03 What is Quorum? What consensus does Morgan Coin use?

What is the blockchain chosen by JPMorgan Chase Bank? JPMorgan Chase chose Quorum.

Quorum can be said to be a clone of Ethereum, an access system, more similar to a coalition chain. Quorum provides privacy protection through chain-chaining, while flexibly supporting multiple consensus mechanisms, and relatively high performance and speed. And anti-forking, can achieve the finality of the transaction.

Quorum is updated according to the Ethereum client Geth. This means that the contracts used in Ethereum can be directly ported to Quorum, including some development protocols (Truffle), which can be deployed directly. Ethereum is currently the most robust public chain, and has also experienced many years of attacks and anti-attacks, can be said to be Bulletproofed.

So what is the difference between Quorum and Ethereum?

First of all, there is a difference in the transaction. A transaction can choose public or private transactions on Quorum. For public trading, the deal was done on the Quorum main chain. The nodes on this main chain can see the transaction information, such as the balance of the account, who the initiator is, who the receiver is, and the logic of Ethereum. If you choose a private transaction, you will get the chain and use a separate server.

Only the hash value of the encrypted data is stored in the Quorum main chain, and the data of the private transaction is stored under the chain and shared by the management engine (Tessera and Constellation) at the node. Only the parties to the transaction (and the regulatory authorities) can see the details of the transaction, and the non-relevant party cannot obtain the transaction details.

But there are also some problems, there may be a single point of failure, and some details , including the regulatory level. For example, the central bank or a regulatory agency was not on the private list at first, but he suddenly felt that there was a problem with the transaction and wanted to check the transaction. However, because it is impossible to add the account directly, it is impossible to see the previous transaction information, which is not ideal from the regulatory level. For the traditional institution of JPMorgan Chase, his first thought was the regulatory friendliness. At the same time, this solution cannot solve the double-flower problem by the consensus mechanism on the blockchain.

Quorum works with the Zcash team to provide a chain of privacy schemes that connect private contracts and the main chain. The business logic of the contract is agreed within the private contract, then liquidated on the main chain, and z-token is used as a bridge to protect privacy in a zero-knowledge proof. The z-contract contract generates a token asset z-tokens that is 1:1 from the main chain asset. Z-contract also runs on the main chain, except that its assets are Shielded Assets that can hide transaction information (sender, receiver, amount of assets, etc.).

Quorum neither uses PoW nor PoS, because both PoW and PoS belong to Nakamoto Consensus, and anyone can join and quit nodes at any time. This consensus mechanism selects the biller through some fair "voting" between nodes. Since nodes have no identity, they are free to create and distrust each other, so voting resources must be scarce. Under the PoW mechanism, the lack of resources is the physical power, and under PoS, this resource is an economic right. No need for access brings decentralized checks and balances, but “competitive” accounting between nodes inevitably sacrifices speed and efficiency.

And for a coalition chain or private platform, financial institutions, does not allow anyone to go to the public chain as a node to book, so Quorum will not use the Nakamoto Consensus global consensus mechanism.

To solve these problems, Quorum tried to use the high-performance fault-tolerant distributed system Byzantine Fault Tolerant (BFT), such as PoA and IBFT; and the fault-distributed distributed system Crash Fault Tolerant (CFT), such as RAFT.

The full name of the PoA is the Proof of Authority. PoA is based on a set of nodes with identities, which are billed in turn. In other words, each node is backed by its own identity and authority.

Each block can be confirmed by only one signature, because only these nodes do not need to compete with other nodes for time, and the efficiency of the centralized server is comparable. There may be a problem that the power of one node is too large. In order to solve such a problem that the power of a single node is too large, each node must be billed at intervals. For example, if there are four nodes on this side, you must wait for two nodes before you can make the next accounting. This way is to limit the rights of a single node. Quorum originally used PoA, and later no longer used PoA because of the finality (which can be forked). JPMorgan Chase handles $6 trillion a day. If you split the fork and refer to the impact of bitcoin cash forks on the entire system, you can understand why JP Morgan does not allow forks because it would cause many transactions to be invalid. .

In contrast, the other two consensus mechanisms supported by Quorum, RAFT and IBFT, are anti-forked. RAFT is actually a widely used traditional distributed consistency protocol, which is applied to container cluster management systems such as Kubernetes and Docker Swarm. RAFT is very effective for faulty, trusted nodes, closed federations that require faster block times and finality. Compared with Ethereum, RAFT also has its own nodes. Relative to any node in Ethereum, the node of RAFT can be divided into Leader, Follower and temporary Candidate. The Leader is the only node responsible for the production block. The Follower listens to the Leader's "heartbeat" and collects the blocks passed by the Leader. The purpose of accepting the heartbeat is to resist system failure. If the Follower does not receive the heartbeat sent by the Leader during its period, the new node continues to be a blocker as the Leader.

When a new transaction is generated, the Leader will not immediately log to the chain. Instead, after receiving all the confirmation receipts from the Follower, it will record and broadcast an executed message, and then all Followers who receive the execution message will block the block. Record on the local chain to avoid forks and ensure finality. The default block time interval under the RAFT mechanism is 50 ms, and the block is only generated when a transaction occurs, which greatly saves storage space. But the RAFT mechanism also has shortcomings, and the prerequisite for it to work is that all nodes are honest. Although RAFT can accommodate a single point of failure, it is not fault tolerant and cannot prevent nodes from doing evil and tampering with historical data.

How to prevent nodes from doing evil while ensuring efficiency and finality? Quorum supports the Byzantine fault tolerance mechanism in Istanbul.

Different from RAFT's complete belief in Leader, IBFT's premise is to accommodate 1/3 dishonest nodes, and through multiple rounds of verifiers, they will reach each other and get out of the block. Since there is only one node at each block height responsible for the block, there is no risk of forking. The account book cannot be changed. To attempt to modify the history record, you need to obtain the private key of all backup nodes and the master node. Compared to PoW, IBFT has no competition mechanism and is faster to get out. However, the disadvantages of IBFT are also obvious. The structure of multiple verification phases increases the number of messages and the number of nodes exponentially. Therefore, IBFT cannot have too many nodes, and is usually used as an enterprise-level and government network.

What kind of consensus mechanism does Morgan Money use? Although JPMorgan’s information on the Quorum network is very public, it does not publish a consensus mechanism on the Morgan currency.

Inferred from the above, Morgan Coin will not choose the PoA mechanism, because it is impossible to achieve the finality, then only one choice will be made in IBFT and RAFT. If all the nodes are partners of Morgan Communications, RAFT is the best choice; if you only trust some member banks, IBFT is a viable pilot.

At least in the early stages, the object of Morgan Coin is a relatively closed alliance. JPMorgan’s clearing network can greatly increase the network effect between banks and provide a more secure compliance information exchange agreement for cross-border payments. The launch of Morgan Coin is a huge boost to Ethereum, including blockchain applications. From a technical perspective, the Quorum network, which is compatible with Ethereum contracts, has the opportunity to interoperate with existing public blockchains and build bridges between centralized and peer-to-peer currency systems.

(Speaker: Pan Chao, Head of Maker Dao China; Editor: Hydrogen 3; Source: Carbon Chain Value.)