Cai Wei-de: Viewing blockchain from a systemic perspective

On August 21, the "China Blockchain Industry Conference" co-organized by the Global Shared Finance 100 Forum and Contemporary Financier magazine was held at the Beijing National Convention Center. Cai Weide, director of the Digital Society and Blockchain Laboratory of the Beijing University of Aeronautics and Astronautics, attended and participated in the roundtable forum.

The following is the full text of the speech:

Cai Wei De: Some people say that a serious problem with blockchain is the scalability issue. There are many solutions for both public and private blockchains. On the one hand, they rely on algorithms, and on the other hand, they rely on system engineering. But in fact, it is not that simple.

First of all, we need to know whether the chain we are building is a private chain or a public chain. Some people say that a private chain is controlled by everyone. On the contrary, a so-called private chain is a blockchain that cannot be tampered with by both insiders and outsiders. Therefore, insiders can only reach a consensus if they have specially licensed nodes, but not just anyone can change it. No one, no unit, and no mechanism can tamper with it. Therefore, many people on the Internet have misunderstood this point.

I have always thought that "decentralization" is a big misunderstanding. To truly solve the consensus problem, the first thing to do is algorithm, and the second is engineering. Now many people talk about blockchain, and the first thing they talk about is the concept of blockchain, such as Bitcoin and Ethereum. The first time I saw blockchain was as a system. Does this system conform to traditional software engineering? It violates many things in traditional system engineering.

Many people say that we are using blockchain to do finance. First of all, confidentiality is very important. Public blockchain has its value and is for enterprises. Therefore, no financial institution abroad is willing to try it. The first one to try it was The DAO, which eventually suffered a huge loss.

There are some similar vague concepts in blockchain, such as smart contracts. Is putting software on the blockchain a smart contract? I talked to lawyers and they said you are just running code on the chain. So for things like this, if we look closely, you have to reach a consensus. How do you reach a consensus?

First, let's look at what the demand is. For example, if a financial unit wants to meet the demand, how many networks does it have? What kind of technology do you want to use? One by one, it needs to be done through system engineering. Whether it is Ethereum or Bitcoin, they have made very big mistakes in system engineering. From the perspective of general system engineering, this is a wrong design, which violates some very obvious system engineering errors.

When we talk about blockchain today, we must first understand the technology and the problem clearly, and look at it from a systemic perspective, so that this industry can develop rapidly. If it is consensus, you are doing public blockchain, it has special restrictions, and some algorithms can be given priority. Foreign technologies have some innovations in consensus, and we have some innovations ourselves. In the final analysis, the first is the computing capacity, the second is the communication capacity, and the third is how many nodes there are. This is the problem.

Now the internal structure of the computer must also be changed to optimize, rather than just taking an algorithm. The algorithm is just the beginning, and the hardware and communication must be combined to achieve the goal. In my laboratory, my hardware and software are not exactly the same as the traditional ones. It is no longer just an algorithm, but a system, at least with computing, database, communication, storage, and it is becoming more and more complex. Therefore, the future blockchain is very different from the blockchain we see now. Our algorithm is basically the same as the traditional algorithm, but the entire architecture design is completely different.