Public opinion is a very interesting thing. From the end of last year to the beginning of this year, we have heard so many praises about blockchain technology. Now, we have heard people’s total negation of this technology. Recently, I have seen different articles in different places, received a lot of hits and been forwarded by people. These hot articles are a theme, emphasizing the futility of blockchain technology, and listing all kinds of scams related to blockchain technology.

In fact, this is normal. As early as the 1990s, Gartner, a market research company, summarized a model called “the hype cycle of technology cycle” to summarize how a new technology matures and evolves in a variety of hot speculation, cold devaluation and people’s hopes and disappointments.

What's great about blockchain

The basic cycle of Gartner technology cycle curve reveals the cycle experienced by new technologies: the period of birth – the peak of high expectations – the bottom stage of the bubble – the bright period of steady climb – Image Credit: )

What's great about blockchain

The latest state of technology cycle curve updated by Gartner in 2018 shows that block chain technology is moving from the “high expectations peak” to the “bottom stage of the bubble” (Image Credit: )

For a person with non senior technical background, it is even more difficult to clarify the essence of technical value from the miscellaneous market performance in the early stage of technology development. The more basic level of technological development, the more serious information asymmetry will be caused, and the easier it is to be used to construct scams. The more in-depth exploration is needed to see its essence through the phenomenon. The same situation occurred in the early days of the Internet.

As a PhD in computer science, it took me nearly ten years to see the clue. I have seen what blockchain technology brings in essence and what existing technologies, infrastructure and their combination cannot achieve. On the 10th anniversary of Nakamoto’s paper on bitcoin, I want to sort out my experience in the field of blockchain and why I believe that blockchain technology is a great technology, and share it with you through this article.

The first decade

I remember that in 2009, when I was in the Microsoft Research Institute for Asia (MSRA) to rush papers and vote for ACM siggraph annual meeting (a top meeting in the field of computer graphics), I inadvertently saw Nakamoto’s paper on bitcoin. I thought it was very interesting. Amway went to dig mines with my friends. Of course, I forgot it a few weeks later.

At that time, I didn’t start to engage in the research on distributed system, didn’t realize how great the premissionless consensus protocol occurred in the whole network, dumped the Byzantine fault tolerance (BFT) algorithm for a few blocks, didn’t realize that this thing would change its financial trajectory in a few years, and didn’t realize that this work was the beginning of a new world.

I am very glad that I have experienced these 10 years, although most of the time, I am just a spectator.

Many friends asked me, what technology is blockchain? Is it reliable? Here’s the conclusion. For the first half of the question, my answer is: blockchain is a unique technology and realizes an unprecedented capability, which cannot be realized by the combination of existing technologies. For the second half of the question, I don’t make any investment suggestions here, but my answer is yes: blockchain technology must have a practical implementation scenario, which can meet the needs that could not be met before, and bring benefit improvement from an unprecedented perspective. However, I must point out that in my opinion, the maximum probability of the reliable scenario of blockchain technology is not what you see now, nor is it an improvement on the existing solutions, but a new business model and new profit model under a new paradigm. It is relatively easy to evaluate the correctness of a technology itself, but it is even more difficult to evaluate the social impact and economic value of a technology. The more technologies that bring fundamental changes, the more difficult it is to figure out their development trajectory. Blockchain is an example.

In 2013, I led the design and development of a GPU cluster based machine learning computing platform at Microsoft headquarters. I used this framework to dig Wright coins (LTC) in my home for some time. Now, it was my first time to “get rid of poverty” and took off the hat of “no coiner”. Then, I also entered the pit exchange and entered as a new leek. Then, after being cut several times, we learned from the pain and finally opened the professional branch line of quantitative trading. In this process, the recognition of technology did not help me to have any effective understanding of the currency price of cryptocurrency. What is the value of a technology without practical implementation? Just use quantitative procedures to explore short-term price trends. At that time, I didn’t understand what the blockchain brought. In 2017, I came to the Innovation workshop as eir (resident entrepreneur) and investor. I read countless blockchain business plans, various fancy white papers, from bottom technology to top application, various flawed technical architectures, self talking economic models and absurd application scenarios. Fortunately, in this year, we invested in bitland and performed my role as an investor. Mining machinery business does make money, and the exchange also makes money, but all this is based on the future expectation of digital currency, rather than any practical business of current blockchain.

How far can this long-lasting drum pass flower go? If there is no large-scale and practical business on the top, I believe there will always be a day to end. However, who knows, just like the pyramid selling of Ethereum in 2016, it has not collapsed, but was continued by the larger ICO in 2017. However, during this period, DAPP on the market gave me some tips.

Core technology value

I believe that blockchain technology has brought something special, something unprecedented.

In 2018, a game appeared in July, with many players, which once again seriously blocked the Ethereum network. Fomo3d, a typical fund tray game, players buy keys at an increasing price. If no one buys keys for a period of time, the game stops. The last key buyer wins half of the deposited funds, and the other half buys back all keys at an average price.

This is a well-designed Ponzi game. At the end of the first round, it divided up ether worth $1.8 million. What is interesting is not the game itself, but the way the game is executed. This fund tray has no value package, but only describes the rules naked. The project party is also completely anonymous, has no background to check, and is ready to run every minute. But when the game opened without trust endorsement, countless players believed it and rushed in. Then everything happened as scheduled, and the previously agreed rules were strictly implemented without any accidents. Usually, Ponzi is always connected with a scam, but this game is a “Ponzi”, but it is not a “scam”.

This is the core value (perhaps one) of blockchain technology I want to say. It is an unprecedented ability that can not be realized by any information technology before:

Within the scope of formalization of all business information, this computing system can prove its innocence.

Back to the game mentioned earlier, it is precisely because it runs on Ethereum that the blockchain technology realizes that once the rules are formulated and deployed, they cannot be modified, and even the operators and developers cannot modify them. As long as Ethereum is not destroyed by 51% attack, these preset rules will be strictly implemented without anyone’s interference and control. The rules will be deployed in an open source manner, and anyone can view and be sure of every detail of the rules. Even if the original developers suddenly evaporate, it will not affect the normal operation of the game. Such a game system can prove its innocence. Everything will happen in the preset way without any accident. All players don’t need to trust the game developers to do no evil, and don’t worry about believing that there will be any fishiness in the operation process of the game. If the information service system can prove its innocence, the business it carries can win the trust of everyone. In fact, this situation does not establish trust, but the operation of the business does not need trust. At the same time, you can also simply agree with the value conveyed by such a business. There is no need for various pre trust building processes, which saves a lot of time and energy. This is the value of consensus.

However, blockchain technology is not omnipotent, and there are preconditions for its self proven operation:

1. Relevant business information must be formalized and reflected in the system. Therefore, the storage, payment and issuance process of bitcoin is guaranteed and does not need trust. However, the price of bitcoin to legal currency is not within the system. The system itself cannot guarantee this thing and completely depends on external trust (or belief)

2. The physical computing equipment of the system must be decentralized, and there may be barriers in the process of participating in the system, but there must be no authorized controller. Compared with the traditional payment access, bitcoin, as a payment system, no access party needs to apply, sign an agreement with any party, and will not be discriminated by the rate or worried about being revoked one day.

3. The business logic code carried by the system must be open source, and it can be proved that the code running in the actual system is the one whose source code is open. Combined with the previous article, this means that the operation of the system will no longer belong to either party. Everyone in this system has no special advantage over others beyond the established rules, including developers.

Turing machine separated from specific physical computing device

In 1936, Alan Turing proposed the Turing machine, a highly abstract computing model, which opened a chapter in the era of computer science. In 1945, von Neumann gave the von Neumann architecture that can be realized in engineering based on Turing machine. Until now, most computers from CPU to GPU, Intel to arm follow this architecture, and improve its scale, speed and performance power consumption ratio to an unimaginable level by semiconductor technology. However, the basic principle is always the structure shown in the figure below.

What's great about blockchain

If we look at blockchain from the perspective of computer architecture, it is an extension of von Neumann architecture.

In the blockchain, “input” is an unordered and unconfirmed transaction, “output” is an ordered and confirmed transaction, and “memory” is the status of the account book. The CPU executes the hardcoded transaction logic in the blockchain node software, or the smart contract deployed by a third party.

What's great about blockchain

I show blockchain from this perspective in order to introduce the working principle of blockchain in the most concise way, not to say that blockchain itself is also a general-purpose computer or a world-wide computer.

Von Neumann architecture was originally given under a completely centralized setting. Today’s single computer, mainframe, cluster and cloud services are essentially such a model. Their concern is to complete the calculation efficiently, and the state is updated correctly to obtain the expected output. The input, execution logic and original state of the whole system are controlled by a trusted single party.

The blockchain is given under a decentralized setting, and the whole computing system is not controlled by any party for a long time. Blockchain is concerned with how the calculation is completed, whether the system input is legal, and whether the execution logic is tampered with. As for the actual amount of calculation carried by it, all historical transactions in this decade can be calculated in minutes with a single machine. In the existing von Neumann architecture computing system, the computing process is always completed on a fixed physical computing device, and the computing device has a single controller. It could be a cell phone, a computer, or even a data center. For the controller of this system, as long as he is well protected against hackers, he can trust this computing system. Others have no basis for trust, and the controller of the system cannot prove to the third party that the system is credible, even if it is open source, that is, using SGX technology is not helpful.

This is the problem that blockchain has overcome. The blockchain further expands the computing architecture of von Neumann architecture and separates it from specific physical computing devices, so as to fundamentally avoid the computing process being controlled by a single controller, so that everyone can trust the computing system. Whether it is consensus mechanisms such as proof of work (POW), proof of entitlement (POS) or Byzantine fault tolerance (BFT), the actual steps of the above calculation process are completed on different physical bases (nodes) around the world. The essence of different consensus algorithms is to give a scheme to select these physical bases in turn to ensure their fairness and security. At the same time, at the data transmission level (submitting input and obtaining output), through the existing point-to-point network technology, the communication does not have a fixed access point and does not depend on a specific IP address. Thus, the computing system can complete the computing process reliably under the supervision of everyone, which is difficult to be tampered with and prevented.

In order to get away from specific physical computing devices, blockchain pays a significant price in performance. To complete the calculation steps intermittently on different nodes, each node needs to be prepared with the context and input data required for calculation. After a calculation step is completed, each node needs to obtain the latest output data and update the context. It involves a lot of redundant information transmission and storage and corresponding calculation. Of course, there are many excellent teams to improve this design and improve its performance and capacity, but no matter how the design is optimized, these redundant communication, storage and computing can not be completely avoided. However, these costs are worth it. It gives the von Neumann architecture a new computing essence: a computing method that can be separated from a specific physical basis and a computing process that can be trusted by any third party.

The computational nature of blockchain

What's great about blockchain

By stripping the application of digital currency in blockchain technology, the figure above shows the essential calculation model of blockchain. The biggest difference from the conventional von Neumann architecture is that it is separated from specific physical computing devices. Once the code is deployed, the state is initialized, and the whole network begins to work, the computing system will run in a way that is not controlled by any controller and will never stop. The system continuously obtains externally submitted operations (OPS) attempting to change the system state, confirms these operations in turn, outputs these legal operations in an orderly manner, and synchronously executes these operations to change the system state.


Account books are not the essence of blockchain. The status maintained by the system can be anything other than account books, of course, not databases. It can be a social relationship map, but it can be the relevance and ranking of web pages, as well as everyone’s credit score and history. Its rules may not be transactions, but any information that can be formalized and digitized. Transaction tamperability is not the essence of blockchain. What the system cannot tamper with are the established rules (codes) and the initial state that cannot be tampered with (such as the handed down block of bitcoin). If the upper level business wants to, we can include the contents of reversing or canceling the transaction in the rules. Orderly chained block packaging is not the essence of blockchain. There are many ways to output data orderly. Batch packaging of blocks is only a compromise between performance and transaction confirmation delay in the current system. Theoretically, there is no problem with one by one, as long as it is orderly. The order of the output transaction may not be a strictly global complete single linear order. For example, a DAG structure is also a possible implementation of the blockchain, as long as the incomplete order does not confuse the state reconstruction. Similarly, the sharding technology based on asynchronous consensus does not have a global order, but it does not prevent it from becoming a possible solution for high-performance blockchain. Of course, we can still call it “blockchain”.

The world computer is not the essence of blockchain. The core appeal of blockchain is to focus on the credibility of the computing process, not to obtain the computing results. This is different from the essence of cloud computing in order to complete the calculation efficiently and obtain the results. Cloud computing cannot prove to a third party that its computing process is credible, and the blockchain cannot be as efficient and low-cost as cloud computing for computing tasks.

In short, the computing essence of blockchain is a von Neumann architecture that can be separated from specific computing devices. It is a trusted computing paradigm that can be proved to third parties. The purpose is to make the computing process authentic and credible, support the self verification of business logic on the upper application, and bring computable trust to the world.

Responsible editor: CT

Leave a Reply

Your email address will not be published. Required fields are marked *