My thoughts after reading "It's time to lift the block size blockade"

Preface: I just read "Bitcoin Jesus Roger Ver: It's time to lift the block size blockade" and was very touched, so I wrote this article to express my point of view. I have always supported the "Bitcoin Roundtable reached a consensus on expansion" reached on February 21, 2016. I often mentioned and called on everyone to agree to support the code development of the hard fork expansion in July 2016, which was completed and released for use; in July 2017, if it can get the broad consensus support of the community, the hard fork will take effect. That is, I support the 2MB expansion very much, and I am only against the hard fork that was hastily carried out with 75% computing power support for some versions due to insufficient consensus and greater risks.

1. Many Options Theory

Roger Ver started by comparing Bitcoin block space to Starbucks coffee, and further elaborated on it into the "theory of many options."

In the coffee market, there are many coffee suppliers. There are also many other beverage options, such as cola, water, tea, etc. Consumers will compare the prices of coffee from different suppliers and coffee substitutes. Then, they will decide the quantity to buy based on their personal preferences.

The same applies to the Bitcoin block space. As consumers, we also have many choices, including Bitcoin, Litecoin, Ripple, Ethereum, Dogecoin, etc. We also have non-blockchain options, including banks, credit cards, Alipay, WeChat Pay, PayPal, cash...

Based on the above similarity, Roger Ver made his important point: " Starbucks would make a mistake by not providing enough coffee to its customers, and Bitcoin miners would also make a mistake by not providing enough block space to meet user demand. "

【 Maya Comments 】

First of all, I want to know whether "Starbucks does not provide customers with enough coffee" is necessarily wrong? I don't know much about coffee, but when it comes to coffee shop brands, I think most people, like me, will think of "Starbucks". From this perspective, Starbucks is a successful company. It is difficult to say whether "not providing customers with enough coffee" is the key to its success or a stumbling block to greater success. You can think about it.

All decisions have their pros and cons. The disadvantage of "not providing customers with enough coffee" is obvious, but are there no advantages at all? Starbucks' management can make the brand so big, they can't all be fools. I guess this is a decision made after fully considering the pros and cons. It is possible that if it had positioned itself as unlimited from the beginning, there might not be Starbucks today because the coffee shop was too crowded and the atmosphere was bad, or for other reasons.

I also found a problem. Roger Ver said that Bitcoin needs to meet user needs. From the list of alternatives, there is no gold. It can be seen that it only refers to the user's demand for Bitcoin consumption. You should know that Bitcoin has other needs besides consumption and circulation, such as the need of Bitcoin fans to store their own wealth. And what is needed to store wealth is Bitcoin's decentralization, Bitcoin's upper limit is constant, and Bitcoin's network is stable. As for whether the transaction fee is high or the transaction confirmation speed is fast, Bitcoin fans who use Bitcoin for value storage purposes do not care much. Other competing coins, bank cash, and even gold cannot replace Bitcoin to better meet the wealth storage needs of this part of Bitcoin fans. If Bitcoin nodes are reduced in large numbers and centralized, then it will no longer be able to meet the wealth storage needs of Bitcoin fans. The selling pressure of such fans after determining that it can no longer be used for value storage will be much more serious than the loss of some users due to the inability to meet the circulation consumption needs. Therefore, I personally think that the priority should be to meet the demand for wealth storage.

2. Pros and Cons Analysis

Roger Ver proposed four negative effects of scaling the main chain and reviewed them:

1) Additional bandwidth is required to relay transactions in the network.

Roger Ver believes that the least efficient miners and nodes will be abandoned by the network, just like the least efficient producers in any industry.

2) Additional CPU time is required to verify these transactions.

Roger Ver believes that it will take some additional time for every full node on the network to verify the payment.

3) Additional storage space is required to store these transactions.

Roger Ver believes that computer storage can also keep up with the deployment of 100 MB blocks, or about 5TB of transaction data per year. In the near future, 1GB or even 1TB blocks will not be a problem from a storage perspective.

4) Due to the above three issues, there may be fewer nodes in the network.

Roger Ver believes that this is one of the biggest risks facing Bitcoin. If the above three problems become very serious, fewer and fewer people will be willing to run full nodes, and Bitcoin will lose some of its anti-censorship capabilities. However, he believes that if the user base of Bitcoin expands to hundreds of millions or even billions, then even if the proportion of people running full nodes will be smaller, the total number of full nodes will increase.

【 Maya Comments 】

First of all, I would like to praise Roger Ver's level. As a person who spends most of his time studying economics, it is very good that he can understand the technical level to this extent. However, as he said, "Let cryptographers do cryptographic work and let economists think about economics." The technical level involved in this expansion issue is really not as simple as he understands. For example, the problem of isolated blocks mentioned by him is actually relative. If the isolation rate of miners increases, there will be no impact on their income.

1) You don’t have to worry too much about bandwidth speed

There is a technology called "thin blocks" now, and the core code is intended to support this technology. With this technology, no miner or node will be abandoned simply because of network speed problems. This technology simply means that the entire block is not transmitted, but only the block header and the list of transaction IDs packaged into the block are transmitted. After the transmission is sent to the next node, the node will extract the specific transactions from its mempool according to the transaction ID to package them. If some transactions are missing, they will ask other nodes for them. Therefore, with the "thin block" technology, blocks of any size will only require very little data to be transmitted after packaging. It can be approximated that the instantaneous transmission of new blocks can be spread out within ten minutes. Think about how many MB of data can be transmitted in ten minutes even in a very poor network environment. Therefore, there is no need to worry too much about bandwidth speed issues.

2) The CPU time to verify the transaction is negligible

Note that the large amount of computing power that the mining machine needs to perform is to find the hash result that meets the difficulty, not to spend a lot of time verifying the transaction. As an ordinary full node, it is not necessary to perform a large amount of hashing. It is enough to just hash the block header once to check whether it meets the difficulty. In addition, it is necessary to check whether the transaction records packed into the block meet the total output is not greater than the total input, as well as the calculation verification of the Merkle Root. These calculations are sufficient for ordinary computers. Even if the computer is configured with a very low CPU speed, it is completely stress-free and negligible to verify a block in a few minutes during normal operation. It is just that when the new wallet is synchronizing the block, there are a large number of blocks to be verified, which may be a little stressful.

3) The total storage space of the blockchain is a big problem.

I don't quite agree with Roger Ver on this. He thinks this is not a problem, but from what I understand, most people are unwilling to run the core wallet. The reason many people give is that it takes up too much hard disk space. Now the Bitcoin core wallet takes up more than 77GB of space after synchronization. I agree that hard drives of tens of TB or hundreds of TB may become popular in the future, but you have to understand that many people are satisfied with just enough hard drives and will not easily replace them. In addition, many people like to have a low usage of hard drives. A 77GB large program that takes up space that could have been used to store more than a hundred movies will be psychologically rejected.

In addition, the large total storage space of the blockchain brings a direct problem, that is, it takes a long time for new users to synchronize the core wallet. The download speed will be the bottleneck of synchronization, and the total block size divided by the network speed is the synchronization time. Now it increases by 1MB every ten minutes, and the total size increases by 365*24*6/1024=51.3GB per year. It is still acceptable. If the block size reaches 10MB, it will be 513GB per year. The synchronization time required at that time is even more unacceptable to ordinary users, and they will choose other wallets. As a result, the number of full nodes of personal core wallets will decrease. Note that although the cropping mode or light wallet can save space, they are not complete wallets and cannot check all historical data. The historical data of the blockchain is difficult to compress, and the speed of generating new blockchain data can only be controlled by controlling the size of new blocks.

4) The biggest risk is that there may be fewer individual nodes in the network.

I agree with Roger Ver that the possible reduction of nodes is one of the biggest risks facing Bitcoin. Note that it is not due to the above three problems. I think it is mainly the third problem. The first two are not a big problem. Even though hard disk technology is developing rapidly, the update speed of personal computer hard disks may not be much faster. The current hard disk space is enough. How many people would think of upgrading the hard disk space specifically for Bitcoin? The possible result is that some personal nodes will stop running.

Roger Ver believes that as long as the total number of full nodes increases, node centralization will be alleviated. I don't quite agree with this view. I think it has little to do with the total number of nodes. The key is to see what kind of nodes they are, personal nodes, cloud nodes, or company nodes. The large number of personal nodes is the key to node decentralization. Otherwise, under super-large blocks, personal nodes will gradually shut down due to excessive network speed, CPU, and hard disk usage. In the end, only some large coin companies will be running, or they will be put on cloud servers, and there will be almost no personal full nodes, which will be very scary at that time. If hackers hack into a cloud server and a coin company with more nodes, then with long-term accumulation and cooperation, a hacker team may be able to control most of the nodes, thus becoming the hidden center of Bitcoin and hard forking at any time. On the contrary, if there are many personal nodes scattered around the world, and they are turned on and off from time to time, it will be almost impossible to hack and control a large number of changing personal nodes, thus achieving node decentralization.

Node decentralization is the foundation of Bitcoin, and it cannot be overemphasized. It maintains the last and most basic counterattack capability in case of a computing power concentration attack. Otherwise, if almost all nodes are operated by the mining industry or currency-related companies, how can we rely on node voting to fight against attacks by joint control of computing power? The centralization of computing power under the premise of node decentralization is not terrible, and it is even the trend of mining development. However, node centralization must not be allowed. When almost no individual is capable of running a full node, the entire Bitcoin building will be shaken when nodes are centralized.

3. Should we increase the block size?

Roger Ver said this: "Many Bitcoin Core members told me directly that the reason why the Bitcoin block size has not been increased so far is not a technical problem. They do not allow it, but only for economic and social reasons."

I don’t know how much “a lot” is here. Then Roger Ver put forward his own point of view:

“Artificially limiting block size does not bring any benefits. It will only cause potential Bitcoin users to use other things. Now, it is time for cryptographers to do cryptographic work and economists to consider economics.” It is obvious that he considered economics as an “economist” and believed that block space should not be artificially limited like Starbucks deliberately limited the amount of coffee produced.

Finally, Roger Ver called for: "It's time to end the block size blockade." That's the title.

【 Maya Comments 】

I think Core is an open team. Not only people who understand code can join this team, but people who understand economics can also join. The views of some people in it cannot represent the views of all Core members.

The technical level expansion is indeed not a problem. The simplest way is to change 1 to 2. In order to reach a full consensus, you can make a 90% block to determine the number of blocks, and then start the code. In theory, this will ensure a safe hard fork. The technical level implementation is very simple. In fact, a certain version has already been implemented. Although the consensus is too low and there may be vulnerabilities, the vulnerabilities that appear can theoretically be fixed.

I think the most important thing is not the implementation problem, but the technical route problem, which is the long-term impact on the technical level. As analyzed above, if the upper limit is completely removed, then there is a possibility of super-large blocks appearing technically, and then the total block size will increase rapidly, raising the threshold for individuals to run core full nodes, and even personal full nodes will disappear completely, resulting in the disappearance of the decentralized nature of the nodes. Hackers can control cloud servers and coin company servers with multiple nodes, thereby controlling Bitcoin from the technical bottom layer, and then the decentralized nature of Bitcoin disappears, and the value of stored wealth is gone. Limiting the upper limit of block size can avoid these technical chain reactions, thereby ensuring the decentralization of Bitcoin nodes.

In fact, Roger Ver should have proposed a suggestion in economic terms, starting from the field of economics that he is familiar with and good at: "We need to improve Bitcoin's transaction processing capabilities and reduce transaction fees." Instead of directly proposing a suggestion to remove the upper limit of the block size in the field of technical code cryptography. In fact, the economic target effect is not only achieved by the method of "removing the upper limit of the block size". The "main chain + lightning + side chain" solution can also achieve the effect of improving Bitcoin's transaction processing capabilities and reducing transaction fees. In addition, the more mature lightning network can have a transaction processing capacity that exceeds Visa, with confirmation in seconds, and the handling fee can be very low, which is an effect that is difficult to achieve by simply removing the upper limit of the block. For details, please pay attention to the article on June 9 of the WeChat public account Bite Jiang to introduce the principle of lightning. As Roger Ver himself said, those who understand economics design economics, and those who understand code design code, and they should not interfere too much in each other's internal affairs. Roger Ver is very knowledgeable about economics and has proposed that the transaction processing capabilities need to be improved and the economic needs of reducing transaction fees can be met. However, at the code level, whether to directly expand the capacity through a hard fork to improve the transaction processing capabilities, or to achieve this through a soft fork to build a "main chain + lightning + side chain" to improve the transaction processing capabilities, this is what people who understand the code know best how to choose to do.

I know that many people may think that I am against capacity expansion after this article is published, so I said at the beginning that I support the 2MB hard fork capacity expansion. Because although I know that hard fork capacity expansion is risky and unnecessary from a technical perspective, it is necessary to understand from the community's perspective. I also suggest that the people in the Core team think more about it. Not many people can understand the harm of hard fork capacity expansion at the code level as you do and it is unnecessary. Many people in the community do not understand the code, especially some people who understand economics but fall into the misunderstanding of capacity expansion. They must hard fork capacity expansion. In order to unite the largest community, it is okay to hard fork once. As long as more than 90% consensus is achieved before launching the hard fork, and pay more attention to the moment of fork, and prepare emergency code. The problem should not be big, and it should be done.

Because the " Hong Kong Roundtable Capacity Expansion Consensus " is a consensus that balances the needs of all parties. It is hoped that most people in the Core team can fully realize the importance of this consensus, release the hard fork capacity expansion code in July in a timely manner, and set it to be launched in July 2017 when the computing power vote exceeds 90%. That is to say, it is completely carried out according to the Hong Kong consensus. At least the developers who signed it should fully lobby other Core team members. The difference between 1MB and 2MB is not big, and it will not have a big impact on node centralization. It is just that the annual increase of 51.3GB has increased to 102.6GB, which is completely acceptable. Some Core team members really can't figure it out, or still think it is unnecessary, so just treat it as a compromise with miners and the community. After all, Bitcoin belongs to everyone, not just you. If you insist on your own way, then it is possible that major mining pools, major trading platforms, and major currency applications will collectively refuse to update any new version updates of the Core team as a punishment for the Core team ignoring the suggestions of the entire community and openly violating the "Hong Kong Roundtable Capacity Expansion Consensus". They could even disband the Core, fork the code, and regroup with members who support releasing the hard fork code. Everyone would agree with this new team that follows community advice and run the new team's version update code.

Finally, I would like to make an appeal: "The block size limit was set by Satoshi Nakamoto. Many of his settings were initially considered wrong or unnecessary, but later proved to be correct. Please do not underestimate Satoshi Nakamoto's wisdom. There is no problem with the block size increasing from 1MB to 2MB, but the upper limit must not be completely removed."