Irresponsible assumptions about central bank digital currency (DC/EP) technology

Central bank's irresponsible guess on digital currency (DC/EP) technology

原文作者：任之劼博士，唯链高级区块链技术研究员。 任之劼是荷兰代尔夫特理工大学博士，博士后，主要研究方向包括区块链共识算法，区块链扩容，区块链应用。曾在国际学术会议上发表多篇区块链论文。在知乎上著有深度技术专栏《区块链演义》。曾开设国内第一门区块链共识算法课程。2019 年加入唯链生态担任区块链高级研究员。

The recent news that the people's bank of China's digital currency (DC/EP) is being tested by agbank and will soon be introduced as a pilot has made "digital currency" and "blockchain" popular again. Then there were all kinds of articles about DC/EP on the web. However, in fact, the private test of DC/EP was not officially released, so all the analysis of the technical characteristics of DC/EP was based on some news released by DC/EP since last year, including the interview and article of the director of the digital currency research institute of the central bank, and the patent applied by the digital currency research institute. The official press release didn't follow up on any technical directions or anything like that for DC/EP.

Therefore, since no one has any inside information, I would like to make an irresponsible guess about the technical solution adopted by DC/EP based on the previous information.

First, DC/EP isn't really new to the industry: I've done this kind of analysis once in 17 years, for example, and the general framework is pretty much the same.

Main references for this speculation:

Yao qian: the central bank's digital currency -- the optimization of the monetary system and its issuance design

Yao qian: new central bank digital currency implementation plan based on block chain

Yao qian: technical considerations of the central bank's digital currency

Fan yifei: some considerations about the central bank's digital currency

Chaum, David. "Blind Signatures for Untraceable Payments." Proc Crypto (1982) : 199-203.

George Danezis, Sarah Meiklejohn, "Centrally Banked Cryptocurrencies", 2015

Several articles on the patent of the central bank's digital currency research institute, previous news about the central bank's digital currency (I didn't bother to look for it before), and mu changchun's course on the central bank's digital currency (I didn't take the course, but I learned about it indirectly).

1. Is DC/EP a blockchain/cryptocurrency

Before it actually revealed a important information, is "the research and development of digital currency is not the default technical route, consider block chain or electronic payment", which, if you have been focused on the development of digital currency and central bank official, will know after half sentence is added after the first statement in the block chain is clear that no used. 1. DC/EP does not use block chain at this stage; 2. Block chain technology may be considered in the future. That is consistent with the central bank's consistent message and technical direction.

Therefore, considering that the last message about DC/EP is that currently DC/EP is only for domestic use and does not consider cross-border payments, I can basically infer that:

1. DC/EP is basically a centralized system.

2. It may use some blockchain technology, but not enough to call it "blockchain" or even "private chain". So it's technically not the same as a cryptocurrency like bitcoin. So in other words the core message here is that DC/EP is not Byzantium fault-tolerant.

3. In the future, when DC/EP is relatively mature, it will consider the use of blockchain technology to meet the demand of cross-border payment.

In fact, from an individual's point of view, it is a matter of course that DC/EP does not use blockchain -- first, it is a centralized system from the central bank's repeated messages, and as we all know, centralized systems do not use blockchain technology or Byzantine fault tolerance. I've seen a lot of people in the blockchain industry express disappointment with this, and some bitcoin supporters in particular. In fact, by contrast, I have always been very happy that DC/EP doesn't use blockchain technology -- it shows that the people in the digital currency research institute of the central bank are not in a hurry to rub off on the hot spots of blockchain technology, and they are well aware of the limitations of blockchain technology and what they are doing.

One might say: "why is Libra adopting block chain technology and Byzantine fault tolerance? Does this mean that Libra's technology is more advanced?"

First - Libra does update because the DC/EP and Libra are in completely different stages: both DC/EP are close to landing, and Libra is still in the stage of changing the white paper. In other words, DC/EP started much earlier than Libra, and the technology of blockchain was still immature at that time, so it was right to choose not to use such a theoretically and practically immature technology as blockchain at that time. The Libra project came out only last year, and it would be ridiculous if the technology was still in place when DCEP started.

Second, updating does not mean better or more advanced. It is true that Libra USES HotStuff BFT, which I think is a "final answer" to the blockchain consensus algorithm, and the improved Libra BFT is also a very logical consensus mechanism, but it is only theoretical and has no precedent in practice. At the same time, the development of blockchain consensus algorithm is still not immature and perfect in my opinion, so even now, I don't think it is a good idea for DC/EP to adopt blockchain algorithm, so I think it is very appropriate for DC/EP to position blockchain technology as "possible future technology".

Once again, Libra and DC/EP are positioned very differently -- Libra is a disrupter and a non-sovereign currency, in other words, its vision of competing with sovereign currencies (which has been changed in Libra2.0) is completely invalid if it does not use blockchain technology and has no competitiveness or topic at all. On the contrary, DC/EP is a centralized sovereign currency, which is not necessary for blockchain technology in terms of technology and vision.

2. How is the anonymity of DC/EP realized without block chain

In fact, blockchain technology is not used for anonymity, and there is no saying that DC/EP is completely anonymous. The official word is "controllable anonymity", which is my understanding -- controllable for the central bank, but anonymous for commercial Banks and other users. In other words, the central bank should have a non real-name as similar to the currency of the DC/EP books, when necessary, can be traced and trading of funds, this is "under control", and for other commercial Banks and users, they can verify each transaction security, but does not know the identity of traders, is anonymous.

According to the paper from the central bank's digital currency institute, the basic structure of DC/EP is more like digital currency pioneer David Chaum's Ecash than a cryptocurrency like bitcoin. But instead of using the Ecash's blind signature, the central bank is the trusted centre.

First, DC/EP takes a similar approach to Ecash in terms of currency issuance -- in DC/EP, the central bank is responsible for issuing the currency. It creates a legal digital currency by giving each currency a unique number and then signing it with the central bank's private key along with the face value information (for example, 100 yuan, 10 yuan). The digital currency, in effect, can be viewed as a one-time certificate endorsed by the central bank that anyone can use to get an equivalent amount of cash, or a new certificate.

In fact, of course, as anyone with any knowledge of digital currencies knows, no matter how well they are encrypted, the hardest part is always preventing double payments. At this point, there needs to be a center with all the books on hand to verify that each voucher has been used. So, when someone wants to initiate a transaction, the recipient needs to send the digital currency to the central bank for verification. If the number has not been used before, the central bank cancels the currency and issues a new certificate of equivalent value to the recipient -- a step that is equivalent to checking the currency.

The problem with this scheme is traceability - because the centre will actually know where all the transactions are going, David Chaum came up with a blind signature - a way to "blind" a message so that the centre can sign the Numbers of the notes without knowing it. So, trade center in authentication can only: "well, with my signature on this, it is true, and this number is the first time to see, so there's no use. But I don't remember what time I signed, so I don't know when I will send the voucher is who, I can only confirm it was just my hair."

However, the DC/EP should not use blind signature technology, because there is no requirement for central Banks to be anonymous and to prevent tracking. For the central bank, anonymity is the inability of commercial Banks and institutions to track funds through this ledger or to decode information about the owners of a particular fund

Here's where DC/EP's two-tier operating system comes in.

Simply put, DC/EP is issued and used through two layers. The central bank is responsible for giving money to the subordinate commercial Banks, which use their own system to make corresponding wallet software, and then support users to change money into the digital currency of the central bank. In other words, the central bank has a central ledger of its own, which is responsible for sending money to commercial Banks, which are in turn responsible for sending money to individuals through wallet software integrated into their bank clients. In other words, the use of DC/EP is definitely through commercial Banks.

The central bank is to consider the process of privacy, or rather, part of privacy, at least in the middle of the commercial Banks cannot as the central bank for DC/EP to track, therefore, the communication between the user and DC/EP should be "blind", namely the collectors in yanchao provide data needs to be encrypted, then only the central bank can decrypt. In the middle, the wallet of the commercial bank is only responsible for establishing the communication.

But here's the inevitable:

1. If both sides of the transaction are from the same bank, it is difficult to guarantee the privacy through the client side transaction of the same bank, and the flow of natural capital in the same bank can also be tracked.

2. If the two parties of the transaction are not in the same bank, the inevitable transaction will at least reveal the information of "funds from xx bank".

But, beyond that, the note does have the same anonymous effect as Ecash on the commercial Banks in the middle

Suppose that the user of bank A receives A sum of money from the user of bank B, then, when the user of bank B transfers the money to the user of bank B, bank B will not be able to see that the money is related to the previous sum of money, because the two pieces of money are different and have nothing to do with each other except the face value.

3. Does DC/EP really have nothing to do with blockchain/cryptocurrency?

Not necessarily - one of these schemes has a huge bottleneck, with all transactions going through the central bank.

In this case, the central bank needs to maintain an accounting book, where the historical part of the accounting book records every money that has been spent, and the existing part records every money that has not been spent, or, in other words, every electronic bill that has not been spent.

This ledger, some like the bitcoin-like UTXO ledger, is also known as a token based or value-based ledger, which differs from the traditional account-based ledger, in that instead of recording the balances of individual accounts, transactions are recorded.

The transfer process in the two books is different:

For example, let's say xiao Ming wants to transfer 100 yuan to xiao hong.

In the familiar account model, the process goes like this:

1. Xiaoming sends a transfer request to the system.

2. The system checks whether xiaoming has so much money in his account.

3. If so, deduct the money from xiao Ming's account and add it to xiao hong's account.

And in this value-based ledger:

1. Xiao Ming will send a 100-yuan electronic banknote to xiao hong, who will send a banknote inspection request to the system.

A) the amount of the transaction is indeed 100; B) xiaoming authorized the transaction by electronic signature; C) the money has not been spent.

3. If the check passes, the system destroys the money, then issues a new 100-yuan e-bill to xiaohong, and registers the bill as belonging to xiaohong.

Here, each trading center needs to complete this verification step. It is conceivable that if DC/EP is applied on a large scale, the pressure of this central system will even surpass that of alipay and WeChat. For an institution such as a central bank, it would be wiser to spread the pressure among institutions that already have conditions. We can see this in the conversations and articles of many of the people mentioned above.

Among them, yao qian in the "blockchain-based new central bank digital currency implementation plan" this paragraph mentioned:

Among them, the registration center records the CBDC and the corresponding user's identity, completes the ownership registration, records the flowing water, and completes the registration of the whole process of CBDC's birth, circulation, checking and dying. Its main functional components are divided into distribution and registration, the right to publish, the right to query web applications, distributed ledger services. Issue registration, release, circulation, return process and ownership record of CBDC; Right confirmation release release the ownership information of issue registration asynchronously to CBDC right confirmation distributed ledger after desensitization. The right confirmation inquiry website provides online right inquiry service to the public based on distributed ledger. The distributed ledger service ensures the alignment of CBDC ownership information between the central bank and the commercial bank.

Generally speaking, it can be understood that in the registration center, we built an "online money detector", namely CBDC right confirmation ledger, by taking advantage of the non-tampering and non-forgery features of distributed ledger to provide external inquiry service through the Internet. This kind of design for the current distributed books, in the central bank and commercial Banks both centralized and decentralized dual mode, the application provides a clever idea, on the one hand, the core issue registration books to isolation and protection of the outside world, take advantage of distributed books at the same time, improve equaling query data and system security and reliability; On the other hand, due to the distributed books is only used to provide query access, transaction processing is still done by issuance of registration system, to refine the atomic transaction granularity approach to distributed computing, which trade, so that we can through the way of business design effectively circumvent existing distributed books on transaction processing technology performance bottlenecks. Obviously, such a design gives full play to the technological advantages of blockchain and guarantees the credibility of CBDC's banknote verification, but it does not affect the central bank's overall control over CBDC.

Later, though, the central bank said in its "no preset course" statement that the above is only an idea, not necessarily the final plan, and that the central bank does not preset the digital currency plan. But it also shows that the pressure on a centre to process transactions needs to be Shared in some way. In addition, from other conversations and articles on the significance of central Banks' digital currency (rather than technology), it can be seen that the core meaning of DC/EP is to reduce the issuance cost, optimize the payment system and improve the effectiveness of monetary policy at the macro level. As a result, central Banks want to be able to track the flow of money when necessary, but the aim is statistical rather than real-time. From this point of view, a better design would be for commercial Banks to complete their own transactions and then periodically, say on a daily basis, summarize their books with the central bank. It would be better for the central bank to act only as the issuer and regulator of digital currency, while the commercial bank would be responsible for its day-to-day use - which seems to be better for both commercial Banks and the central bank.

As a result, many people will find -- after all, don't all the nodes need to share an ledger? So, isn't it still block chain? Moreover, this creates even more serious performance bottlenecks -- instead of the central bank recording the entire ledger to process all transactions, each bank now needs to process all transactions in real time, which is not enough to spread the pressure to commercial Banks.

In fact, we don't have to use blockchain in the real sense -- blockchain is a completely untrustworthy scenario, which is not the case in the DC/EP scenario -- in fact, the central bank digital currency RScoin, proposed by university college London in 2015, describes such a structure.

Simply put, we assume that the correspondence between each address and the bank is known to all Banks. Then, let's say xiao Ming in line A wants to pay 100 yuan to xiao hong in line B.

1. Xiao Ming sends A 100-yuan electronic banknote to xiao hong through the address of line A, and xiao hong sends A banknote verification request to the system of line B.

2. At this time, line B system needs to check that a) xiaoming (judged by the address) is the user of line a; B) xiaoming authorized the transaction by electronic signature; C) line A signs to confirm that the money has not been spent and promises to mark it as spent.

3. If the check passes, line B will add the money to xiaohong's designated address and issue a signed receipt to xiaohong.

But here's the problem -- bank B has to trust bank A to actually verify that the money exists and honestly mark it up and not double pay. There are several ways to ensure this:

1. The central bank's arbitration: in the above steps, xiao Ming's client provided A certificate with the signature of line A on it during the transaction. Therefore, if line A has any misconduct, xiao Ming can provide the certificate and appeal to the central bank. Similarly, if B behaves improperly, it can also be found by the central bank through xiao hong's proof.

2. Redundancy can be introduced: at the expense of some efficiency, each address is held responsible by more than one bank (say, three), and a majority of those signatures is then considered legal.

3. Byzantine fault-tolerant algorithm can be introduced: the introduction of redundancy is equivalent to treating multiple Banks as a system where only common errors will occur. We can also consider the case of malicious behavior and adopt Byzantine fault-tolerant algorithm to ensure the consistency of output results.

The implementation of all three is flexible - the first can be used for more credible institutions and the third for less credible ones. Another area of flexibility is the three steps: 1) the paying bank checks to see if the money is available; 2) the paying bank marks the money as spent; 3) the receiving bank accounts for the money, which can be considered separately. In the previous quote by yao qian, we saw such a proposal that the inspection step could be guaranteed to be safe by using Byzantine fault-tolerant algorithm, while the other two steps could be guaranteed to be efficient by not using Byzantine fault-tolerant algorithm.

But the important thing to note here is that even if all Banks are involved in the verification and Byzantine fault tolerance is adopted, this is still not a blockchain system in the traditional sense, which I will explain in detail in the next section.

Then, at the end of each day, the central bank is responsible for compiling all the books, and then reissues the new currency to the addresses of individual users.

But one problem with this approach is privacy and anonymity -- because not every transaction reissues money, and the movement of money in a day can be tracked -- which can lead to privacy breaches. Therefore, the approach that can be adopted is similar to bitcoin's "non-real name" approach. The same account generates different addresses for collection and payment, but only the bank and the central bank can calculate that these addresses belong to the same person, while other institutions can only see that these accounts come from the same bank.

This way, another layer of protection can be added to the anonymity.

4. DC/EP wallet/client

First and foremost - the DC/EP wallet is very different from a cryptocurrency wallet like bitcoin!

The reason is that, as mentioned above, the DC/EP system is not Byzantine fault-tolerant, or even general fault-tolerant. Even if Byzantine fault-tolerant algorithm is adopted in transaction verification and entry, it cannot make the whole system Byzantine fault-tolerant -- that is why it is not a blockchain.

It is possible for this system to have a single point of failure, and that single point is in the wallet/client.

The redundant set and the system above the central bank's arbitration mechanism can guarantee in commercial Banks and central Banks don't go wrong here, for the client, our hypothesis is that there is no reason to deliberate users inconsistencies, such as xiao Ming pay little red scenarios, such as to buy things, xiao Ming want to little red payment prove successful, so he is obligated to give a little red "my 100 yuan is the real one" proves that in providing the proof that the money has been transferred out. And small red received the money, naturally is for the future to use, so she also has the motivation to send the certificate to the address of the corresponding institutions, otherwise she will not spend the money.

But these steps are sure to be integrated into a wallet by the bank. Logically speaking, commercial Banks will not deliberately create inconsistencies, but as long as the software is likely to bug, and in the real network, there may be packet loss or signal suddenly disconnected. In fact, in the above system, if the wallet has a problem, it is easy to cause the inconsistency of the whole accounting book. For example, in let's bank issue legal proof, if using the redundancy mechanism has three bank responsible for verification, if the client only sent one of the legal proof that request, and then broken network, then due to the system bug after reconnection didn't send the request to the other two agencies, or because the broken network operation, to the other two books after reconnection request a different deal, so can cause one of the books and the other two books don't agree. Inconsistencies in situations like this are outside the system design and can lead to consequences that are very difficult to predict and resolve.

Therefore, the security of the wallet is crucial, because unlike traditional cryptocurrencies such as bitcoin, the wallet is part of the security of the entire system. A failure in a bitcoin wallet can lose at most the user's money rather than affect the consistency of the bitcoin ledger, but a failure in a DC/EP wallet could result in errors or paralysis of the entire system's ledger.

Based on this, we have the following speculation:

1, digital DC/EP is RMB, that's it - because, from the technical route of this system is closed, and even into the future is impossible to completely open, namely the purse and books are only open to a handful of authorized institutions to develop, and need a long time test, and the inspection and certification authority, this represents a cannot use DC/EP development Dapp or DeFi this kind of financial innovation, nor is the so-called "programmable currency", This is a far cry from Libra's vision and many similar projects and virtual currencies. At best, developers can only use alipay to invoke the transfer interface provided by the authorization wallet, and cannot directly issue requests or initiate transactions to the ledger.

2, the wallet is probably now the ABC test focus. As the scope of the private test is only limited to agricultural bank, I tend to think that only the central bank and agricultural bank institutions are considered at this stage, and the redundant design of other institutions is not considered in the verification part. Therefore, the probability that DCEP will be able to join in the future is only five rows in the beginning.

3 for anonymity, binding the purse and A bank account, so use A wallet transfer, in fact, no matter from DC/EP logically, or in the way the internal implementation from the bank, should be very similar to using A bank account for transfer, but they are not necessarily bound together, known as the concept of "loose coupling".

4. Technically, DC/EP provides the framework, but what is actually important for the central bank is as follows: 1. The currency is registered in the central ledger, and the commercial bank is only in custody. 2, the internal commercial Banks can use any technique, such as zhang this getting on to build a database, using an existing account or transaction validation completely centralized by endorsement, the credit block chain technology is used to enhance communication between efficiency and security are no problem, the central bank welcome you try all kinds of technology, however, the format of the summary to the general ledger book need consistent, reliable data, and finally the book or the central bank summary shall prevail.

5. Double offline payment

One of the major differences between DC/EP and current alipay/WeChat payment is the dual offline payment. One of the scenarios used as an example is when two phones touch each other to complete the payment under the connected state.

The parts of the two phones that touch each other can be exchanged via bluetooth or NFC, but the core dual-line method, as patented by the digital currency institute, may require trusted hardware.

The specific method is: in the mobile phone that supports TEE (trusted execution environment), the offline payment requires the signature of TEE and checks that no double payment has been made. At this time, the security of double offline payment is actually guaranteed by the manufacturer that supports TEE chip. In this way, first of all, the technical difficulty of destroying TEE is relatively large and the cost is relatively high. Secondly, the phone chip is also real-name. If the security of TEE chip is broken, the device can be held accountable and blacklisted.

But according to the contents of the patent, double offline payment should only be applied to the temporary and small, receive money through online synchronous to the later can be used again - this actually not difficult to understand, because if offline receive money can use have been offline, equivalent to exist two books at the same time, or two digital currency. The security of one online digital currency is guaranteed by the central bank, while the security of the other offline digital currency is guaranteed by various TEE chip makers. This does not follow the logic of DC/EP.

And if will double offline payment only as a temporary method, then the TEE microfinance as an offline only a temporary custody, so even if the security of the TEE is destroyed, also not affect the safety of the DC/EP, central Banks are also require the production of the TEE time to make up for the loss is more easy to calculate.

-- -- -- -- -- -- -- -- -- -- -- -- -- --

However, in fact more than logic does not need the TEE can be achieved, such as change the TEE to a real name and phone binding bank client - this client in the code to write dead above logic, then, if it is found that the client is modified led to data errors, the binding mobile phone write the blacklist and holders of accountability. In fact, the effect of the two is the same, the difference is that TEE is more secure.

However, logically speaking, double offline payment only wallet is achievable - after all, according to the previous design, from the central bank's point of view, all the day's trading in escrow belong to the bank before write general ledger book, itself on both sides of bank credit, and if the wallet software is provided by the bank, so by the wallet software temporarily managed currency is also equivalent to both Banks responsible for the safety of the currency.

6. Basic structure of DC/EP

Through the above analysis, I speculate that the basic structure of DC/EP is as follows:

The central bank has a DC/EP general ledger, which USES a value-based/scrip based structure to record the use and recovery of all issued digital currencies. If it is in circulation, record its current owner and the bank. Every once in a while (such as every day), the central bank collects the books of the day from commercial Banks and other trusted institutions in a centralized manner, and then checks and compares them. The central bank has the right to arbitrate on discrepancies in the books. Then, for all transactions that have not yet been registered with the central bank (the transactions of the day), the central bank registers them in the general ledger book and reclaims the used currency and sends an equal amount of the new, anonymous currency to the corresponding address. Finally, the code for the new currency of the day is issued to all verification agencies.

The central bank also manages an identity database that records the identities of all addresses. The identity database is neither internal nor public, and only allows access to the addresses of people involved in criminal ACTS.

In addition, the central bank manages a query interface for publicly verifying the correctness and integrity of the books (as of the day) and a transaction entry interface for limited traceability of each currency (but to an address that is not a real name), and a certified wallet access only for updating the trading books. Commercial Banks maintain their own books for the day (the general ledger book is not required because the central bank has one) and are required to maintain security and availability. In this case, the ledger may be the ledger of the institution's users, or it may include the ledger of other institutions other than the institution (redundancy design). But in the ledger and in the blue boxes throughout the diagram there is only address information, not user information. The books are anonymous - no one except the parties involved, the Banks involved, and the wallet providers can draw any information from the transaction about the parties or the source of the funds prior to that date.

The central bank authorizes commercial Banks or other trusted institutions to develop wallets, which must be certified and accepted. DC/EP is available and transacted only through the wallet. Here, the DC/EP wallet is an important part of the system and therefore cannot be arbitrarily opened to third parties.

For users, when using DC/EP, the whole DC/EP (the green part) can be viewed as a black box, interacting only with the wallet. But you can also look up the ledger and verify that your transaction records match those provided by the wallet through a third-party call to the query interface. In terms of privacy, the wallet is real-name, so the provider manages the user's information under the supervision of the central bank. Users need to store their personal information in an identity database through their wallet when registering for a wallet.

So, if xiao Ming wants to send A digital currency of 100 yuan to bank B via bank A's wallet and bank A's address (for simplicity, let's consider paying face to face), there is the following transaction process:

1. The wallet used by xiao Ming and xiao hong can communicate with each other through qr code or NFC to obtain a collection address of xiao hong in line B.

2. Xiaoming's wallet selects an address in line A that xiaoming has not spent, and issues the transaction request to the transaction interface.

3. The transaction interface finds the row A corresponding to this address by index (and other rows if there is A redundant design) and then asks them to prove that this currency is available (in the list of current currencies published by the central bank and not marked for use).

4. The above accounting institution returns the available proof through an algorithm a and declares that it has been marked for use.

5. After xiao Ming's wallet receives the return certificate, send the certificate and transaction to xiao hong's wallet.

6. Xiaohong's wallet submits the transaction to the transaction record interface, and then the transaction verification interface finds the corresponding line B of the address according to the index.

7. According to algorithm b, the corresponding institution will add the transaction to the account book of line b (and other lines if there is a redundant design), and return a proof that the transaction has been added to the account book to xiao hong.

8. The deal is done.

Here, if only a single mechanism is involved, algorithm a and algorithm b can be simple reads and writes, and when multiple mechanisms are involved, it can be a general fault-tolerant algorithm or a Byzantine fault-tolerant algorithm.

In the summary of accounts, the central bank has discretion if there is a discrepancy. If the result is a problem, the user can ask the central bank for the lost money through the evidence in hand, and the central bank can investigate the culpable party in the process according to the evidence.

The dual offline payment is not included in this process, but it is easy to see that the logic of the dual offline payment should be embedded in the wallet software and triggered when the network is disconnected. Thus, the transaction is not passed to the ledger, but is simply exchanged through wallet software records, temporarily recorded in hardware, guaranteed by TEE provider (chipmaker) or wallet provider, and then synchronized to the ledger after networking.

In fact, this structure can also correspond to the statement of "one coin, two libraries and three centers" of yao's premise.

The author has proposed the digital currency system of central bank with "one currency, two libraries and three centers". The "one coin", or central bank digital currency, is an encrypted string of Numbers that is guaranteed and signed by the central bank to represent a specific amount of money. "Two base" currency issue Treasury and shall establish subsidiary issue treasuries index word library digital currency commercial bank, the former is the central bank on CBDC private cloud storage CBDC issue fund database, according to the operation management system for the central bank's cash management, which is the commercial bank deposit CBDC database, can be in the data center of commercial Banks in CBDC private cloud, follow the commercial bank cash operation management specification. The "three centers" include the certification center, the registration center and the big data analysis center.

Among them, the registration center records the CBDC and the corresponding user's identity, completes the ownership registration, records the flowing water, and completes the registration of the whole process of CBDC's birth, circulation, checking and dying. Its main functional components are divided into distribution and registration, the right to publish, the right to query web applications, distributed ledger services. Issue registration, release, circulation, return process and ownership record of CBDC; Right confirmation release release the ownership information of issue registration asynchronously to CBDC right confirmation distributed ledger after desensitization. The right confirmation inquiry website provides online right inquiry service to the public based on distributed ledger. The distributed ledger service ensures the alignment of CBDC ownership information between the central bank and the commercial bank.

Generally speaking, it can be understood that in the registration center, we built an "online money detector", namely CBDC right confirmation ledger, by taking advantage of the non-tampering and non-forgery features of distributed ledger to provide external inquiry service through the Internet. This kind of design for the current distributed books, in the central bank and commercial Banks both centralized and decentralized dual mode, the application provides a clever idea, on the one hand, the core issue registration books to isolation and protection of the outside world, take advantage of distributed books at the same time, improve equaling query data and system security and reliability; On the other hand, due to the distributed books is only used to provide query access, transaction processing is still done by issuance of registration system, to refine the atomic transaction granularity approach to distributed computing, which trade, so that we can through the way of business design effectively circumvent existing distributed books on transaction processing technology performance bottlenecks. Obviously, such a design gives full play to the technological advantages of blockchain and guarantees the credibility of CBDC's banknote verification, but it does not affect the central bank's overall control over CBDC.

In particular, this double-ledger inclusive design not only continues the mature stability of traditional technologies, but also leaves room for new distributed ledger technologies, making the two distributed technologies compatible, parallel, complementary, and competitive in the evolution process.

Here, of course, one currency refers to DC/EP, and the two libraries are the books of the central bank and the books of the commercial bank, the former on the private cloud of the central bank, the latter on the private cloud of the central bank, and the latter on the database of the commercial bank. Then, the registration center is the system after the access of the query interface, while the certification center is the system after the access of the transaction entry system. The big data center is a separate interface, which is not drawn on the diagram. The two interfaces can adopt different consensus algorithms, namely algorithm a and algorithm b in the previous trading process. According to yao qian, algorithm a can adopt Byzantine fault-tolerant algorithm to guarantee security, while algorithm b can use general fault-tolerant algorithm to guarantee efficiency.

& have spent 7,

The author's major is consensus algorithm of blockchain rather than non-financial. The understanding of digital currency of the central bank is only limited to news, public information and some other people's interpretation of DC/EP. Therefore, this paper is purely speculation from a technical perspective, mainly including:

1. The anonymity of DC/EP adopts a structure similar to Ecash, but the blind signature is replaced by the center. The anonymity mainly reflects that each "digital currency" is homogeneous and anonymous to all institutions except the user and the central bank.

2. The main architecture of DC/EP is not blockchain, but it may adopt a framework similar to RScoin. Subordinate commercial Banks can use the existing infrastructure to process transactions, or use blockchain or other technologies to achieve inter-bank transfer, as long as the daily summary to the central bank's general ledger is reliable.

3. If the structure similar to RScoin is adopted, in fact, the wallet of DC/EP plays an important role in ensuring the consistency of accounts. Therefore, DC/EP should be a relatively closed and single-function system, that is, it only has the function of digital currency, and the ledger is only open to certified wallets.

4. Dual offline authentication is only used as a temporary means of payment, which can be based on TEE or software only, and then implemented by means of after-the-fact accountability.

5. The earliest DC/EP may be the least viable version, leaving only commercial bank 1 and the wallet from that commercial bank in the figure above.

6. DC/EP has not considered the cross-border payment scenario. If cross-border payment is needed, relevant facilities can be established through its subsidiaries.

Finally, the above is pure speculation, the author is not responsible for any deviation in the implementation, such as reprint please reprint the full text to avoid misinterpretation.

BlockBeats has been authorized by the original author. If you want to reprint BlockBeats, please contact the original author for authorization.

Original address:https://zhuanlan.zhihu.com/p/133817993

Welcome to join the official BlockBeats community:

Telegram Subscription Group: https://t.me/theblockbeats

Telegram Discussion Group: https://t.me/BlockBeats_App

Official Twitter Account: https://twitter.com/BlockBeatsAsia