Tired of the endless debate over capacity expansion? Listen to what Satoshi Nakamoto said

The debate over Bitcoin’s scaling didn’t start last year, or in 2013 or 2011. The scaling issue existed when Bitcoin was born, even before Satoshi Nakamoto officially launched the cryptocurrency. The debate has never stopped since then.

Sometimes debate leads to progress, and sometimes it doesn’t. In the Bitcoin scaling debate, we found that the focus of the debate has not changed over the years.

2008: James Donald

Most people know that the dispute over Bitcoin’s block capacity is not groundless, but few people know that it all started in November 2008. At that time, Satoshi Nakamoto first disclosed his project to his circle of cryptography friends.

In fact, after Nakamoto proposed the concept of Bitcoin, the first response he received came from James A. Donald, a Canadian libertarian and cypherpunk. He said: "It doesn't seem to scale."

I understand your original intention of proposing this concept. We need such a system very much, but it seems that it cannot scale according to the demand... In order to verify and prevent double spending in time, a transaction must contain a large number of past transactions, so everyone must have a large number of past transactions or recently verified transactions. If millions of people make transactions at the same time, it will consume a lot of bandwidth.

Satoshi Nakamoto’s response is still being quoted by many on-chain scaling supporters:

Before the network was overloaded, users could secure transactions by simply using Simplified Payment Verification (SPV) to check for double spends. In this case, the entire chain would need block headers, or about an additional 12KB per day.

Only miners need to run full nodes, and Satoshi seemed to be fine with the emergence of large server farms.

First, most users will run network nodes, but as the network scales, more and more server farms with professional hardware will appear, and node operation will be more handed over to professionals. A server farm only needs to run one network node, and the rest of the LAN can connect to this node.

In addition, Satoshi Nakamoto also believed that the demand for bandwidth was not as serious as Donald thought.

The bandwidth problem is not as big as you think. A typical transaction is about 400 bytes, and the error correction code (ECC) is compact. Each transaction must be broadcast twice, assuming each transaction is 1KB in size. In fiscal 2008, Visa processed 37 billion transactions, an average of 100 million per day. This large volume of transactions should consume 100GB of bandwidth, which is the size of 12 DVDs or two HD movies, or about $18 of bandwidth at current prices.

It may take several years for the Internet to really expand to this extent. By then, posting two HD movies on the Internet will not be a big deal.

Generally speaking, most people stop citing this part of history here. Many people use this passage to prove that Satoshi Nakamoto had a plan to expand the blockchain, and believe that he was very confident that Bitcoin would reach the level of Visa after the expansion. But Donald's response was very interesting. He fully elaborated his point of view, and more importantly, he also anticipated the views of most of the opponents of on-chain expansion today.

First, he explained the relationship between the government and economic entities, and said that he was not willing to break away from this relationship and go to large server farms to mine Bitcoin:

If a group of companies are issuing new coins, this model may be less vulnerable to attack than a single company. But governments often choose to attack financial networks. As I write this, the attack is still ongoing and many financial companies have gone bankrupt.

For Donald, the only viable solution is to run as small a mining node as possible:

I think that as a currency issuer, as a small currency issuer, we should reduce data and bandwidth requirements as much as possible, and the model described by this protocol is a waste. The smaller the data storage and bandwidth requirements of the currency issuer, the stronger the system's resistance will be, and it will not be easily attacked by the government.

In Donald's opinion, there is no comparison between Bitcoin and bank cards:

You have to go where there are no bank cards.

Currently, file sharing is essentially binary unit sharing... File sharing costs are extremely low, and each client can process several transactions per second.

In his opinion, there is only one way to achieve Bitcoin file sharing:

…Bitcoin should have a bookkeeping layer that supports the most transactions with the least space and provides anonymity. It is a bookkeeping currency based on Chaumian currency (digital cash).

Donald's idea is now known as a payment channel. Today, the Lightning Network, as the second protocol layer of Bitcoin, has built a payment channel network that can send countless transactions in almost real time.

Based on his vision for the Lightning Network, Donald believes that Bitcoin is a settlement system, not a payment system:

We can build a privacy layer on top of a digital gold coin based on ledger coins and JOM coins, just like the US banking system separated ledger coins and paper money (issued by banks) for gold coins before 1915. Indeed, we urgently need to reduce transaction costs in this way and support micro-transactions due to bandwidth control, file sharing and non-whitelisted communication charges.

From the above quotations, it can be roughly seen that the seeds of the block capacity dispute had taken root long before Satoshi Nakamoto sent the first Bitcoin transaction to Hal Finney.

2010: Hal Finney

In December 2010, another veteran cypherpunk and cryptographer, Hal Finney, expressed his views on Bitcoin banks:

In fact, Bitcoin banks are necessary. Such banks can issue their own digital cash currency, which can be exchanged for Bitcoin. Because Bitcoin itself does not have the scalability advantage, it cannot broadcast every financial transaction in the world to everyone, and the same is true for blockchain. A lightweight and efficient second-layer payment system must be developed. Similarly, the time required for Bitcoin transactions also means that it is impractical to become a medium for large-value transactions.

Hal Finney does not consider Bitcoin to be a payment system, but rather a currency. Like all currencies, the payment process can be achieved through an intermediary, which he calls a bank.

Bitcoin banks can solve these problems. They function like all other banks before the nationalization of currency. Different banks can adopt different policies: some are more aggressive, some are more conservative; some can hold reserves, while others are 100% based on Bitcoin; interest rates vary; some banks' funds can even be traded at low prices.

Bitcoin, a network where miners confirm transactions on the blockchain, can serve as an anchor for individual Bitcoin banks:

I think this is the ultimate fate of Bitcoin - deposit reserves, a reserve currency for banks to issue digital cash. Most Bitcoin transactions are executed by banks, making network transfer settlements. Private Bitcoin transactions will become very rare, just like the amount of Bitcoin purchases today.

Hal Finney's idea received little but sharp response. A bitcointalk user commented:

This sounds like a fractional reserve system, and isn’t Bitcoin claiming to be able to avoid this fiat currency trap?

But at that time, people could never have imagined that this topic would trigger the endless debate it does today.

2013: Gavin Andresen, Peter Todd, Mike Hearn, Gregory Maxwell

At the end of January 2013, the issue of capacity expansion became a hot topic on BitcoinTalk. At the time, some people believed that limiting the block size to 1MB was problematic and suggested expanding the capacity through a hard fork.

Some people suggest increasing the block capacity as soon as possible:

We must seriously consider expanding the block size before it exceeds the limit. Currently, most blocks are over 300KB. Bitcoin is growing rapidly, and there is no doubt that it will soon break the 1MB limit.

The 1MB block limit is a more practical issue, involving users' expectations and visions of Bitcoin: in a 1MB block capacity, only 7 transactions can be processed per second (in fact, only 3 or 4 transactions are processed per second in most cases). This situation is completely inconsistent with Bitcoin's title of "world currency".

However, this post about capacity expansion was met with strong opposition as soon as it was published. User da2ce7 believes that increasing the block capacity through hard forks touches on an essential problem:

This topic has been discussed countless times. 1MB is a strict upper limit set by the protocol, and changing this upper limit is as difficult as changing the total amount of currency issued...which is to say, basically impossible.

Some Bitcoin investors actually do not accept the cap set by the protocol.

Even if most people agree to change the upper limit, some users, including me, will definitely not recognize this chain.

Gregory Maxwell, a thought leader of the anti-expansion group, also expressed his views on this topic. First, he believes that the 1MB cap can provide economic incentives for miners in the long run:

From the current situation, it is unreasonable to increase the block capacity, because once the block capacity is increased, the miner subsidy may be reduced and the Bitcoin system may no longer be secure.

The value of Bitcoin lies in its scarcity. One important manifestation of Bitcoin's scarcity is its limited supply, and another is the block capacity limit; only when the block capacity is limited can transaction fees exist, and the fees can ensure the security of the blockchain as a mining subsidy. At present, there is no other way to ensure the long-term development of Bitcoin except for miners, unless we accept corporate monopoly or regulatory intervention (both of which are not feasible because they do not value the decentralized nature of Bitcoin).

But he also had other concerns:

After the block capacity is infinitely expanded, Bitcoin may not be able to guarantee its decentralization: only a small number of large banks that voluntarily participate will be able to verify transactions.

Since 2013, there has been a rumor in the community that large blocks will jeopardize important features of Bitcoin such as decentralization.

That year, Bitcoin's capacity was a soft limit. Miners could raise specific limits through a soft fork. Bitcoin transaction volume did increase that year, and developers were discussing a soft fork.

At that time, Gavin Andresen began to consider how to solve the problem of block capacity limitation in the long term. He suggested adopting a floating block strategy to ensure that the block size changes with demand. Peter Todd objected after listening to his suggestion:

Block size expansion has been a hot topic recently. However, few people have considered that the abuse of miners’ incentive policies after the block size increases is likely to lead to the centralization of the network.

Once the block capacity increases infinitely, miners will continue to increase the block capacity and eliminate weaker nodes. This is also the main reason why many people oppose Bitcoin Unlimited today.

Regardless, miners, especially large miners, are able to profit from large blocks.

As Todd said, the end of the story is "centralization of the mining industry."

At this point, Andresen also joined the debate. His point of view can be said to be a complete victory for both sides:

I strongly suggest that we should not position Bitcoin as a 'high intensity currency' system, after all, Bitcoin can only process 7 transactions per second at present.

I agree with Stephen Pair (BitPay CEO) - it is likely to be a highly centralized system.

One side is concerned about the centralization of nodes and miners, while the other side is concerned about the centralization of transactions. Andresen explained his point of view:

I think we should put users first. What do users want? They want low fees and fast confirmation. Let's move in this direction. After all, users are the value of Bitcoin.

We already know what was discussed next. On one side were Andresen, Mike Hearn, and others, and on the other side were Todd, Maxwell, and others. Like many other users, Pieter Wuille seemed to remain neutral. Most users seemed to support block size expansion, but the specific plan, size, and time were not yet agreed upon.

Perhaps this user has captured the essence of the scaling debate:

The battle over capacity expansion is about whether Bitcoin is a payment network or a store of value.

It’s an old, long, and never-ending story: Which is more important? A decentralized store of value? A payment system for users? Should nodes or transactions remain decentralized?

The topic of scaling only lasts for a few pages in this article. But the debate is far from over; issues within the community have arisen and will not stop for a while, only increase.