Bai Shuo: From the technical point of view, DeFi feasibility, 3 major challenges, 8 big mistakes, how to crack | 2019 Global Blockchain (Hangzhou) Summit Forum

On May 17, the 2019 Global Blockchain (Hangzhou) Summit Forum, hosted by the Hangzhou Financial Office and hosted by Babbitt, was held in Hangzhou.

Bai Shuo, director of the Chinaledger Technical Committee and former chief engineer of the Shanghai Stock Exchange, shared the theme of "Defi's Advancement from a Technical Perspective". He said that DeFi is a decentralized format of finance, and its application prospects are very broad, but There are currently three major technical challenges. DeFi's propulsion environment has been greatly improved compared with a year ago, but there are still 8 typical misunderstandings that are hindering us. If we break these misunderstandings, it will help us to promote such a format. Bai Shuo believes that with the huge wave of digitization, DeFi is unstoppable, because everything needs to be digitized, and the digital world cannot be lack of trust and value. It is not allowed to use the blockchain to support this matter.


The following is a compilation of Bai Shuo's shared content, organized by Babbitt.

DeFi is a hot thing this year. Its connotation is a trust path that is hoped to be supported and redefined by decentralized technology. So, is it feasible to use it in the traditional financial industry, or to start a new business in a new business? Today we will do some discussion.

1. Three manifestations of finance

First of all, what are the current forms of financial applications based on blockchain technology? It mainly includes the currency side, the asset side and the financial side.

1. Currency application


There are three types of currency end applications: the first is virtual currency (wild); the second is the token (the generation) that anchors the legal currency; the third is the legal digital currency (true).

Wild: Virtual currency has experienced ten years of development, with climaxes and lows. As a means of payment, it is mainly used for comprehensive e-commerce websites, e-commerce portals for merchants, third-party payment platforms, and electronic black markets; as a means of exchange, there are also some performances, mainly used for digital currency exchanges; Private equity markets and equity crowdfunding are currently in need of resolution of legality issues.

Generation: The digital currency bound to the legal currency, which is what we usually call the business circle currency, has several characteristics: there is no independent "exchange rate", it can be used as an intermediate accounting unit, which is somewhat similar to "clearing reserve", its Imports and remittances are circulated within a certain range. The most important problem to be solved in the circulation process is the problem of liquidity.

Real: Statutory digital currency, some big countries with better technical strength are staring, generally the central bank is dominant, once it is introduced, it has statutory effect, online circulation, the impact of the introduction of legal digital currency includes controlling counterfeit currency, anti-money laundering, and convenient withdrawal. , convenient for super hair and so on.

From the perspective of positioning, the legal digital currency will be in parallel with the cash, and it is a private sector. It is completely different from an intermediary service such as a bank. Under such a position, is the legal digital currency a threat or an opportunity for business? I think it is actually an opportunity. It allows banks to expand beyond the traditional business to digital cash and digital accounting.

What are the selling points of legal digital currency, peer-to-peer experience? But this is enough for mobile payments. In fact, the legal digital currency will not give us a better experience; my money is my master? That is not necessarily true. From the perspective of anti-money laundering, exchange control, or other regulatory perspectives, there are always people who want to be your master; programmable? Yes, if our legal currency is programmable and the imagination is infinite, I think this is the only opportunity and the place to fight.

2. Asset-side application


Asset-side applications can be divided into equity registration and transfer. The stable currency that is now said is a manifestation of bonds in a sense. The traditional exchange trading mechanism is classified into two modes: order-driven and quotation-driven. Order-driven trading mode has an absolute advantage in traditional trading. There is no possibility at all, but the quotation driving can be done to a certain extent. Yes, the score is the last form.

3. Quasi-financial application


Quasi-financial applications include letters of credit, order pledges, and supply chain financing. The digitization process itself will involve credit and trust, and it involves the re-digitization of value. This is unstoppable. It is nothing more than a question of how to do it.

In addition, there are many non-financial applications, including notarization, elections, charity, and environmental protection, which are reflected in the aspects of depositary (notarization and government affairs), evidence collection (product traceability), and property rights (Internet of Things).

In summary, there are some positive signs in the blockchain field. There are currency changes in the currency circle and chain changes in the chain. The nearest circle is also actively exploring how their business can be expanded under the new technology framework.

In this context, let's take a look at the technical challenges and how to deal with them.

Second, technical challenges and solutions

Basically, the technical challenges facing the blockchain industry have not escaped the three areas I have previously summarized: performance, cross-chain, privacy.

Technical Challenge 1: Performance


In the dimension of performance, there are two axes, which are batch flow and slice parallel. If the two are simultaneously wound, it can reach a considerable length, which is 10,000. Going up to the 100,000, 10 or even 10 million levels, the water is relatively large, and the second is that there is no standardized definition of caliber, which has no meaning in the past.


As shown in the figure, full verification, full storage, and full chain are the traditional practices. Later, various optimization measures include: in the verification aspect, only store and verify related party transactions; in terms of storage, fragmentation and snapshot There are also different evolutions in the out-of-chain configuration.


There are also three new trends in the evolution of the blockchain consensus mechanism: from the most to the less, from the chain to the graph, from the determination to the random.

From the most to the least, from four digits to three digits to two digits, there is a trend of centripetal aggregation; from the chain to the graph, there is a trend of topology optimization; from determination to random, because of "less" The goal is to centralize and target homogenization. As long as you control the platform, the consequences are very serious, which is why it is so important.

In the process of more or less, we have seen that the right to consensus has also decreased, and the danger has existed. It is now necessary to distinguish between the potential consensus power and the actual consensus power, so that the potential consensus power becomes larger, the actual consensus power remains small, the actual consensus power determines the final efficiency, and the potential consensus power determines the security, the random algorithm. Therefore, it is mainly for this reason that it is valued.


This is the logic of their evolution: from more to less, reducing the consensus communication link, reducing the complexity of consensus; shortening the length of the evidence chain from chain to graph, reducing storage overhead; from determining to random, expanding the scope of confusion, reducing the consensus node being attacked Probability.

Technical Challenge 2: Cross-chain


The second technical challenge, cross-chain. The lofty ideals that we are pursuing in the mature form of the Internet, if we want to go over, the mountains and the hills will certainly not work. We must connect and interoperate, that is, we must take the road of cross-chain.

A relatively low-level form of cross-chain is a side chain, and a further development is a public side chain. The final form may be a circulation chain. It is not important where you are issued. In the end, the real circulation is not in the same place as your distribution. But in this case how to ensure the conservation of value and avoid all kinds of attacks, this is the key to cross-chain.

Technical Challenge 3: Privacy


Privacy, this is also a big proposition. It is because with the development of the Internet, data is too quickly concentrated on the hand of a few data oligarchy, causing the public to control their own data and loss of privacy.

This involves a variety of techniques, which are collectively called preventive data sharing, divided into several types: back-to-back comparison (point-to-point), back-to-back query (point-to-batch), back-to-back filtering (batch-to-point) Back to back convergence (batch to batch).


Back-to-back comparison, it is as if Zhuge Liang and Zhou Yu wrote a word in the palm of their hand, and then compared it, everyone wrote fire. When technically implemented, it is verified by cryptography. After a certain conversion, everyone can compare.


Back-to-back query, this is water, fire, wood, there is fire, there is fire, I want to check the corresponding fire is there, if it is, what is the corresponding data, in the form of credit data cooperation, use this A back-to-back query is very important. It is necessary to prevent all kinds of institutional data from being aggregated into a super center when it comes to credit information. Personal credit may be feasible, but for corporate credit, its data comes from many different aspects. It does have a lot of difficulty.


Back-to-back filtration is reversed, that is to say, there is so much water, wood, and fire coming over. After my filtration device, only the fire is left.


Back-to-back aggregation, which is often used in artificial intelligence. The reason why artificial intelligence is so powerful now is that data can be aggregated. The reason why data oligarchy is so powerful is because of the artificial intelligence generated after data aggregation. a synthetic effect.

Then, we are thinking, can we achieve artificial intelligence without sacrificing personal privacy and sacrificing the data rights of small units, small league economic entities, and legal entities? This is a big proposition, we call such a type back-to-back aggregation of data.

This kind of aggregation involves a variety of mathematical operations. After we have undergone a certain cryptographic transformation, we can continue to use it as long as it can be witnessed. For example, addition, after the transformation, the relevant horizontal equation is still not the original balance. Keeping addition is the so-called homomorphism. If you still maintain a kind of nonlinearity, you can use the neural network to do it. With the neural network, you can do the experiment of the neural network.

We know that a millionaire's problem is an inequality operation, or whether the inequality relationship can be verified in the case of encryption. In our bookkeeping, it is that you can't overdraw, the balance doesn't have that much money, you want to spend so much money, how to endorse in the case of encryption.

There is also a geometric inclusion relationship, which involves point-to-point intersections. An advertising company and a mobile phone company need to cooperate with customers. With the mobile phone number as the medium, they can extract the intersection, but they don't reveal themselves. It can be said that such a scenario is broad enough. If such a technical means of privacy protection can be provided, we can do more in the field of commercial applications.

This can be done with machine cryptography or non-machine cryptography to build some security protocols with privacy protection. Then, in non-machine cryptography, you can also have a similar confrontation network to learn. By adding these tools to the blockchain, you can do integration. For example, many hospitals have medical images, which are important assets that cannot be taken out. However, if the sample of medical images is not adequately trained, it may not be effective. So how can we get more samples together, and at the same time save their assets in their own domain, not by the opponents, this problem can also be solved by means of confrontation with the network.

Three, eight common misunderstandings

There are a few misunderstandings that we often encounter when we create a new DeFi format:


The first misunderstanding is that decentralization is the essential attribute of the blockchain. We must first look at the premise of decentralization, whether to be an athlete or a referee. This is doubtful. The second is the original intention. The original intention of a system is to be more fair, but the result of evolution is monopoly. This is what we have to think about. In addition, if there are scenes that can accommodate a certain degree of centralization and do a better job, why not? Finally, in the case of emergency, it must be handled by centralization.


The second misunderstanding is that the blockchain efficiency is bound to be low. We have already said the issue of consensus just now. Now there are some new breakthroughs, not necessarily inferior.


The third misunderstanding is that the blockchain cannot be without coins. This is not the case. There are some intersections here. These intersections can do a lot of articles and use legal means to do some legal things.


The fourth misunderstanding is that the blockchain can be not open source. Is it necessary to hide your own code to be competitive? no. Open source in the development phase, testing phase, there are hackers will polish your blockchain security better, but closed source, its security may not be reassuring. We have seen that some company codes are not open source, but the hype of their own team is counterproductive in the blockchain field.


The fifth misunderstanding is that the blockchain can be a product unicorn. I personally think that it is unlikely to produce a unicorn.


The sixth misunderstanding is that the blockchain is a simple move of the business carrier. If it is just a simple move, there is no need to do blockchain. There must be a blurring of the enterprise boundary, optimization of the trust path, the process can be dismantled, a certain incentive is embedded, and the governance means is fully digitized. At this time, the block Chain is a good choice.


The seventh misunderstanding, I believe that the blockchain can not be hit by quantum computing.


The eighth misunderstanding is that the blockchain is open and transparent and cannot protect privacy. This has already been said.

Fourth, from the distribution of the three powers, DeFi promotion ideas


From the perspective of the distribution of the three powers, it has the right of accounting, supervision and use. There are three strange things based on this. First, the accounting rights overlap with the supervision rights. Second, the owner of the right has no feeling for the blockchain. It is not normal, it is an extreme situation. Third, there is a lack of governance, including operation and maintenance control, error correction, business logic adjustment, and protection of data assets.


What we need most is how to configure the priority situation, that is, the combination of the three powers, that is to say, it is two words, self-certification. We must not only do things, but also be able to prove ourselves. There are several different levels of self-certification, such as drying the sun, drying the relationship, drying the books, and drying the source code. We hope that there will be more self-certifying ecology.


To sum up: DeFi is a decentralized format of finance, and its application prospects are very broad, but technical challenges still exist. The promotion of the environment has been greatly improved compared to a year ago, but the promotion ideas need to be changed. At present, there are some inherent misunderstandings that hinder us. If we break these misunderstandings, it will help us to further promote the emergence of such a format. Along with the huge wave of digitization, I think this is unstoppable. Everything needs to be digitized. In the digital world, there is no lack of trust and value. Blockchain is used to support this matter.