While the Core developers' misplaced priorities and theymos's censorship have both stonewalled Bitcoin's development for the past year, much of the fault for Bitcoin's problems lies at the feet of miners. Given that there are so few mining pools in control of the Bitcoin network, any one of the large Chinese mining pools could easily decide to move forward with an expansion of the network's blocksize and others would undoubtedly follow. Despite numerous conferences, agreements, and promises, no miner has stepped up to do so.
Miners are not incentivized to take any action to deal with the blocksize problem because their profit motives are not aligned with the actions of normal Bitcoin users. Miners earn largely the same amount of money regardless of whether users use the network or not, and their income stream would take on more risks than benefits if the Bitcoin network forked. Businesses, on the other hand, are suffering greatly from the capacity impasse and the benefits of a fork greatly outweigh the risk of the network failing (since many businesses cannot use a crippled Bitcoin network anyway.) Miners are likely to continue to delay unless the status quo threatens their revenue, and users are currently powerless to stop them.
This post presents a proposal to shift the power over the decision of whether to fork the network from miners to users and businesses. Since polls have repeatedly shown that users are largely in favor of increasing the network's blocksize, I believe the Bitcoin network would be made more capable. If a blocksize increase does not result from this action, that can also be considered a positive result; it would then be clear that users and businesses believe that Bitcoin's blocksize should remain fixed, in which case developers can devote effort to other Bitcoin features and businesses that require capacity can use other coins.
Creating a crisis
Unfortunately, action rarely happens until a crisis occurs. Guardrails are installed on a highway after someone drives off the road. Runway lights are upgraded when a plane overshoots and crashes. Weapons are banned after shootings. Sometimes, bad things need to happen to demonstrate to people that a problem exists. When that recognition occurs, people often act quickly and decisively. For example, it often takes a decade for a highway reconstruction project to succeed, but the I-35W bridge in Minneapolis was demolished, designed, constructed, and reopened with more lanes in just one year after the previous bridge collapsed.
While Bitcoin has a severe capacity problem, in the absence of a crisis few have recognized or care about how users and businesses are slowly adjusting their plans to use other coins or private blockchains. Bitcoin needs a crisis to be created to force action, and I suggest that a crisis which harms Bitcoin investors and miners in the short-term can be created by a relatively small number of people.
The client
A Bitcoin client should be created that is identical to a version of Bitcoin Core in all respects except for two changes:
- At a certain block, the client accepts Bitcoin blocks that are merge mined, in addition to normal blocks mined through the regular method.
- At this block, the blocksize limit is removed, all blocks are allowed to be of uncapped size.
Until the designated block, sending bitcoins using either client spends money on both blockchains. When the designated block occurs, the chains permanently diverge. Smart miners will then mine both chains, because they can supplement their income by selling coins from one or both chains. Pools which do not offer merge mining will lose customers, because the customers will move to pools where payouts are better. Every bitcoin user will have money on both chains, which means that millions of people have an immediate interest in the unlimited blocksize fork.
The crisis
The divergence should be timed to occur as soon as possible, after giving people a reasonable amount of notice to upgrade. The purpose of the timing is two-fold: first, to come to a resolution and end the uncertainty as soon as possible, and second, to drive the price of bitcoin down in the short term. Miners will only respond if their livelihood is threatened, and a decline in the price of bitcoin is the only way to force them to pay attention to this problem and take action.
It is even possible that miners may disapprove of this proposal so much that they decide to take preemptive action and adopt a large blocksize increase to prevent the fork from occurring.
A theoretical rollout of the new client could involve the following schedule:
- Day 1: announcement of impending client release date
- Day 8: client is released
- Day 15: exchanges open markets in merge-mined chain
- Day 29: merge-mining and unlimited blocksize activates, and chain forks
The most significant advantage of this idea is that it can be executed by very few people. Widespread miner consensus is not required, nor would it be sought. The only necessary people are a group of developers to modify the Bitcoin client to add the two changes, a major exchange to offer trading on the new fork, and someone to conduct marketing and advertising to ensure that miners are aware of what is happening. The software development is time-consuming but does not require a technological breakthrough. At least one exchange has an incentive to open an initial market in this coin because of the huge profit available. The most difficult part is determining how to spread the word to people who only read forums controlled by theymos.
Switching from miner control to user control
In addition to forcing action on the issue by creating a panic, the other primary objective of the proposal is to switch the authority over whether the blocksize is increased or not from miners to users. In all of the other current proposals for increasing the blocksize, like Bitcoin Classic and BIP101, a 75% or 95% consensus of miners is required to execute a hard fork.
Miners are not the people who should be making this decision. People who build businesses around and who spend bitcoin on everyday transactions should be deciding the future of the currency. On a philosophical level, when someone provides a service in an economy, they are paid for their work by the buyer. Buyers who are unhappy with the service provided can demand changes or withhold payment. The seller works for the buyer, not the other way around.
In this case, miners are providing the service of securing the bitcoin network for users, who are paying transaction fees to them for their work. Miners should be compensated for their work, but the Bitcoin network does not exist to serve them. It exists to serve the needs of users, who expect to get what they are paying for - a reliable and capable network. It isn't wrong to demand that they provide that service, and that they re-invest a portion of their earnings to buy more bandwidth to provide the level of service demanded.
The fact that hashrate increased after the halving event demonstrates that miners are earning immense profits. Since people are still turning equipment on, the profit margin was at least 50% before the halving! The argument of many miners that they cannot afford the resources to support a more capable network was completely discredited by their behavior after the halving occurred.
How the chains would be valued
After the split, the limited and unlimited blockchains would be valued differently. Theoretically, the total current value of bitcoin would be divided between the chains, and the value of each would represent the probability that that particular chain will succeed and become the canonical chain. In reality, the panic caused by the fork would likely cause the total value of both chains to decline, and that value would not be regained for some time.
It is tempting to think that one of the chains will lose all value and be discontinued, but that is unlikely to occur. Since there is no reason to stop mining the old 1MB blockchain and there is no event where difficulty will dramatically decrease, if the switch is successful the old chain will still retain a very small value, and it will still be able to be used by people who cannot upgrade their software or who want to avoid the main chain, just as other altcoins are used now. However, it is possible that Peercoin, another widely mined non-merge mined SHA256 coin, could supersede the old chain for this purpose.
During the time when both chains have value, people who own bitcoins can buy and sell based on their views of future value of the chains. Every user would, at the outset, possess the same amount of currency on both chains. Investors who do nothing retain the value of the current total price of bitcoin - regardless of which chain succeeds, and even if the transition drags out indefinitely.
Discovering whether the blocksize limit is truly needed
The proposal to merge mine blocks of undefined size allows it to be determined whether, in the real world, a blocksize limit is necessary at all. None of the current experiments can definitively answer this question, because they can't predict user behavior, only technical capabilities. In addition, it isn't possible to simply create an altcoin without a blocksize limit, because it wouldn't have enough usage to gather meaningful data. Most of the altcoin's blocks would be empty, and the large blocks would consist of meaningless test transactions.
However, splitting the chains would allow the question of whether an unlimited blocksize is dangerous to be answered once and for all. The initial stages of the chain fork would likely see the unbounded chain valued significantly lower than the limited chain. If a failure were to occur because of excessive traffic, it is most likely to occur shortly after the split, which would minimize losses.
Potential problems
This section examines a few issues that could result in the failure of the proposal.
Not getting the fork listed on an exchange
Coins have no value unless they are listed on an exchange. The prerequisite for this proposal's success is for the unlimited version to be listed on an exchange, and for it to be tradeable beginning on the day that the fork occurs. If the fork were not listed on an exchange, then the fork would have no value and nobody would merge-mine it.
Getting coins listed on exchanges can be accomplished in two ways. One path is to convince a major exchange, like Poloniex, to list the coin by pitching that they could become the first to what could possibly become a hugely profitable market. Poloniex began trading Ethereum and remains the market leader because Coinbase waited too long to relax its Bitcoin-only business model. An exchange may be willing to add the fork simply because the reward is so high.
A second method of getting listed on an exchange is paying a fee for listing, like can be done with any altcoin at exchanges like Cryptopia. However, these fees can be as high as one to three bitcoins, and it's not clear that anyone would be willing to pay such an expensive amount.
I don't think it matters if the exchange that initially lists the fork is small. There is either support for an unlimited blocksize or there is not. If support exists, that exchange will be overwhelmed with profit, and other exchanges will quickly list the fork too.
Finding someone willing to lead the effort
Since the code for these changes is simple (for the blocksize), or has already been deployed and is widely used (the merge-mining), even I could make those changes alone if necessary. Testers would be required, but surely five or ten testers could be found to ensure that the changes didn't introduce any regressions.
The most significant manpower problem is finding someone who would be willing to lead the effort in the long-term. Some users may make their usage and investment decisions in the larger chain because they want to see a well-organized team in place once the Core is removed from control. A well-spoken, charismatic leader would also assist in explaining the merits of the new chain to users and in spreading the word to miners to make sure the network gains a significant proportion of the old chain's hashrate quickly.
Finding this leader is going to be difficult because the position isn't paid, and the people who are the most qualified need to earn money with their time. In addition, the natural and most experienced leaders (like Gavin Andresen) have specifically declined leadership opportunities in the past. Others are afraid of the confrontational nature of this approach and would not want to ruin relationships.
Withstanding attacks
Supporters of the new fork are likely to be subject to DDoS attacks. While it is possible to plug security holes, there isn't a solution to DDoS attacks. In the past, criminals have flooded nodes that supported BIP101 and Bitcoin Classic with enormous amounts of traffic, knocking entire networks offline.
Until the new fork becomes supported by enough users, its nodes could be shut down by these attacks. Businesses accepting the new chain could also be subject to attacks. Because of the potential for attacks, the success or failure of the system is likely to be decided within the first few days. There will be an initial period of weakness where price is low, node count is limited, and the low volume in exchange fees does not provide enough revenue to the exchange to justify the cost of the extra bandwidth required to withstand the attacks.
Conclusion
The Bitcoin network can be forked by introducing a client that allows for merge mining and the removal of the blocksize cap. The fork would remove the decision of whether to keep or discard the blocksize limit from the hands of developers and miners, and would place it in the hands of businesses, investors, and users. Once the fork is deployed, users will be able to freely demonstrate their support by buying and spending coins in the unlimited fork. As the old fork becomes more and more congested, transactions will move over to the new fork, and price action will determine which fork becomes the canonical chain. A side effect is that the introduction of the fork will initially drive down the total Bitcoin value and demonstrate to miners that their revenue is in jeopardy unless they take action, possibly causing them to remove the blocksize cap immediately to avoid this outcome.
To get started, the primary resources required are the money to get the fork listed on an exchange (or the contacts to convince an exchange to do it for free) and a marketer to publicize the client. A longer-term need is for a leader who will take charge and continue Bitcoin development when the current Core leaders are ostracized.
In an ideal world, Bitcoin would upgrade through a consensus of miners and users. Unfortunately, years of discussion have demonstrated that a cordial solution will not be able to be achieved. It is time for users to take action and determine whether there is demand or not for a blocksize increase. If demand exists, the availability of an unlimited chain will see users voting with their money and their feet as the limited chain becomes congested and expensive. If users are happy with the status quo, then the unlimited fork will receive no attention and die. Either way, the issue will have been determined once and for all, and miners will have little say in the decision.