Let’s solve these four major problems before talking about blockchain

The recent popularity of "blockchain" has not diminished. Whether in the financial application field or in the IT field, there are many fantasies and expectations, but many details are not clear and it is a bit foggy. At the beginning of 2016, Caixin Weekly interviewed Zhou Xiaochuan, the governor of the People's Bank of China, and gave a rational perspective on the attitude of decision-makers towards "blockchain", saying that we need to see this thing clearly and maintain a "curiosity".

From my understanding, blockchain is a "consensus" implementation technology. Through blockchain, all transactions between the Internet can be recorded, and blockchain users can witness the realization of "consensus", and the information content on the chain is "unchangeable". This "unchangeability" increases the cost of malicious tampering of content through the existence of multiple copies in the system. In the case of the Bitcoin system, all the bright spots in the figure below represent a set of ledgers with consistent content. Therefore, when all records are made public, the problem of "two tables are unmeasurable" in real life is solved. The reason why the two tables are unmeasurable is that there is no center, and the observers have no way of knowing the different values ​​of the two tables. But the center is not necessary. When there are multiple tables and most pointers point to a time, the minority obeys the majority, and the observers know it well. Therefore, one of the problems that blockchain needs to solve is "the minority obeys the majority". The existence of the minority may be an error in data generation or malicious tampering. In other words, if you want to tamper successfully, you must change 51% of the copies in the system [i.e. 51% attack], from the minority to the majority. It is conceivable that the cost of tampering is huge in terms of technical difficulty, time consumption, and personnel use. At the same time, another key combination of blockchain is the combination with smart contracts. The contract is executed through the program, because the former has achieved the "original flavor" of the contract content, and the latter becomes an "observer" to use machine execution to eliminate the breach of contract and execution interruption of both parties. From the earliest payment to the high development of financial derivatives, financial activities are all signed and executed one by one. At the same time, most of the various financial innovations are also to ensure the execution of contracts and default compensation. If the contract is flattened and the execution is thorough, the whole process will be more concise and efficient, which is a premise that blockchain will also be valued by the financial field. In addition, blockchain is a constantly "expanding" recording system that can record all transactions from the beginning to the future operation of the system. This mechanism also facilitates the tracking, monitoring and transfer of asset transaction information by third parties, especially financial regulators.

Figure 1 Blockchain copies in the Bitcoin system Description: Each bright spot is a terminal in the Bitcoin system, and each terminal maintains a ledger (copy)

Two types of data storage modes

If the contract is the content, when the content is digitized, the blockchain becomes the carrier of information storage. Therefore, combined with Melanie Swan's opinion, there are two modes for data storage. One is that the data is stored in the block, and the other is to compress the data into a digital sequence, which is the dimensionality reduction hashing process [see Melanie Swan's "Blueprint and Introduction to the New Economy of Blockchain" page 21]. The former is easy to implement, but the consistency problem is prominent. The latter only synchronizes the digital sequence and solves the consistency problem by reducing the synchronization amount, but the dimensionality reduction of the hashing process becomes the key.

Consistency issues

The consistency problem is a difficult problem to solve in distributed storage systems. Usually, this link plays a role in disaster recovery and backup, and synchronization operations are mostly performed in "idle time" as a background operation. Blockchain uses "front-end" synchronization operations. When a block is generated, the update of the block chain requires all copies to complete the replication of consistent content. If the content is stored in the block, the amount of data updated is the updated content multiplied by the number of copies, plus the transmission distance and bandwidth between copies. The entire operation is time-consuming, and the overall efficiency is reflected in the longest synchronization time in the system, as shown in Figure 2.

Figure 2 Blockchain architecture

Therefore, this approach will affect the response to high-frequency applications, such as many transaction scenarios. Therefore, the solution to the consistency problem is to control the size of the updated content and to limit the number of copies. The Bitcoin system uses the former, while the private chain uses the latter.

Dimensionality reduction hashing problem

The consistency problem is indeed the key to the real application of blockchain technology in the future. How to overcome the "demon"? Mei Wanni Swan gave his idea. If the amount of synchronized data is only a 64-bit or 128-bit digital series, then the consistency is at least controllable. In this mode, any digital asset data (such as files, pictures, video recordings, monitoring records, software, etc.) of the blockchain service does not need to be stored on the blockchain in the end, and the updated content is only the corresponding series. Therefore, how to reduce the dimension is the problem in this mode, and the content of the consensus is natural semantics. When natural semantics transitions to digital expressions recognized by programs, multi-dimensional natural semantics and metadata need to complete a one-to-one relationship mapping. This process requires information dimensionality reduction, which will use hashing, encoding and other technologies. For example, a 9G genome file needs to be reduced to a 1-dimensional binary series to balance the negative externalities of the blockchain synchronization front, and ultimately enable the execution algorithm of the smart contract to identify the "authenticity" of the object. One thing that needs to be reminded is that the hashing process in this way is not an address mapping (the address is the storage address of digital asset data, such as the storage address of genome files. To distinguish the hash address in Bitcoin, its essence is to clarify the ownership of Bitcoin, similar to the A and B parties of the transaction, which is the content of the transaction). If the address information is stored in the blockchain

Figure 2 Dimensionality reduction + hashing of blockchain

First, it violates the content consensus mechanism, and second, it greatly reduces the security of the system. Once an illegal intruder obtains any copy, he can find the storage space of the contract content and tamper with it. The process of dimensionality reduction hashing is similar to "compression", which requires compression based on specific information such as data content, file characteristics, and timestamps. In this way, the execution algorithm of the smart contract can determine the "authenticity" of the specified content. Therefore, this process is complex and arduous, and the workload is huge.

Smart Contract Connection

The core of smart contracts is to use program algorithms to replace people to execute contracts. From this concept, the execution of contracts can be understood as "Alipay". When a buyer and a seller reach a transaction on Taobao, the payment of money and goods will be executed according to the Alipay program, and both parties are more in line with the implementation of the program. Therefore, the premise of smart contracts must be that the signed transaction cannot be tampered with before execution. Blockchain has therefore become the natural cornerstone of smart contracts in a decentralized environment. This mechanism ensures that the "original" contract can be obtained before the smart contract executes the program.

Figure 3 “Triggering” a smart contract

There are usually two natural solutions to the "triggering" of smart contracts, Mode A and Mode B, as shown in Figure 3. The former drives the execution of smart contracts by forming trigger events through the content of the data itself, so it is necessary to verify whether the data content and the hash value match before execution. If they do not match, it means that the original data has been tampered with and a disaster recovery mechanism needs to be matched later. If Mode B is adopted, the smart contract needs to interpret the hash value. If the dimensionality reduction hash is "compressed", then the connection with the smart contract needs to be "decompressed", and it needs to be interpreted from the characteristic values ​​of the digital series to realize the "event triggering" of the execution algorithm. Therefore, the application of blockchain and smart contracts needs to standardize the content of digital assets, and realize the customization of asset registration, title confirmation, transfer and other functions through standard component configuration.

Blockchain search

If the solution to the consistency problem is the practice of blockchain "write" operation, then the retrieval of blockchain content is a "read" process. Therefore, for the needs of applications, blockchain should not be used for the sake of blockchain. On the basis of achieving broad consensus on the content, we can choose the right way to optimize the traversal on the blockchain. Centralized retrieval or the compound retrieval function of the blockchain itself are both reference methods. Vechain, developed in China, adopts the latter method, hashing the address information and content information at the same time. However, this solution reduces the security of the system because there are too many copies in the blockchain system. If an illegal intruder obtains any copy, he will get a complete address table of the system, and the intruder can easily find the storage space of the original data for "tampering". Therefore, it should be emphasized that the "immutability" of the blockchain only refers to the content stored on the blockchain (the content circled in red in Figure 4, the serial number in the figure explains the whole process) is difficult to tamper with, not other data. At the same time, compound address information also sets obstacles for the load balancing of future systems from another perspective. Since the blockchain cannot be "slimmed down", the storage information is expanding day by day, which will gradually reduce the "read" efficiency. Will the client need to load the complete blockchain in the future? After all, the proportion of information involved for a single user is very low, and establishing a blockchain server in the region is not an expected direction. In this way, the "decentralized" blockchain will transition to a "non-centralized" blockchain, and the blockchain server will retain the complete chain information.

Figure 4 Applications similar to VeChain

Conclusion

The infrastructure of traditional finance is centralized. The aggregation of information, the content of contracts, the endorsement of credit, and other operations in the transaction process must be centered on financial intermediaries. By endorsing the contract reached by both parties, it is assumed that all transaction information carried by the intermediary is correct. Therefore, the traditional centralized service mode, the system itself is not fault-tolerant. Once the intermediary itself has moral risks or the intermediary is maliciously modified by hackers, there will be loopholes in the performance of the transaction contract, and the participants will suffer property losses. This further explains the motivation for the repeated attacks on the servers of financial institutions in recent years. The existence of blockchain allows the system to be fault-tolerant, and through "decentralized consensus" to achieve a unified digital expression of the system, it brings supply and demand closer. Continuous attention is itself a "rebellion" against the status quo of traditional finance. The essence of finance is the flow of value, and demand, supply and intermediaries have become the three links to achieve flow. At this stage, it is difficult for traditional financial intermediaries to achieve direct satisfaction of supply and demand. Blockchain strengthens supply and demand and weakens intermediaries to optimize the matching of supply and demand efficiency. This is also a technical realization of the current supply-side reform.

When the mind surpasses the footsteps, when the "legends" in the market begin to run wild, I am only here to highlight some areas for further exploration in the future. Due to the limited space, I can only list some conceivable details in a few words. Only when the pace of practice keeps up can we bring the vision closer and go further.