Discussions regarding the block size are often burdened by complex technical and engineering details, but fail to emphasize the key driver of Bitcoin’s future success. The critical driver of Bitcoin’s future success is its continued utility. If Bitcoin should for any reason become useless, it will become irrelevant! If it were to become inaccessible, or if it were to perform poorly in its main functions (1) assuring the security of the cryptographically scarce counterfeit resistant digital token bitcoin or (2) fail to securely facilitate the movement of these tokens from one address to another in an efficient, reliable, predictable and affordable manner – Bitcoin & bitcoin would become irrelevant.
The original maximum allowable block size was 32MB1. In the early days the average block size was only a small fraction of 1MB2 and there was a legitimate concern that a malicious actor may arbitrarily generate a block of 32MB to deliberately harm the nascent Bitcoin network. The cost to do so would have been zero. A prophylactic limit of 1MB was introduced. Shortly after discussions related to increasing the block size beyond 1MB began3. Today we are seeing the continuation of that conversation, but with much higher stakes4. It appears that demand for transactions will soon exceed Bitcoin’s current capacity for transactions. As a result, that conversation is instead an intense and heated debate.
The fact that Bitcoin has survived seven years is a remarkable achievement5. It continues to evolve in many ways6. The use of bitcoin addresses7 and the demand to transact8 has steadily grown over the past seven years. Demand for transactions is approaching the theoretical limit of what a 1MB maximum block size can provide. This is a moment to celebrate!!
Diverse approaches to overcoming the transactional bandwidth constraints that a block size limited to 1MB impose have been proposed. These have been countered with speculative predictions about why and how a particular proposal will lead to the failure of Bitcoin.
I think there is great benefit, in this discussion and as we speculate about the impact of our decisions, to consider what is the critical scarce economic resource we are considering with a perspective on maintaining the utility of the Bitcoin so that it does not become irrelevant.
I am confident that it is so, but I must make the assumption that current participants in the Bitcoin ecosystem are not malicious, are pleased with Bitcoin’s progress thus far, and wish to preserve Bitcoin’s utility to a broad audience long into the future. On February 10th in an Open Letter titled: A Call for Consensus9, the Bitcoin Roundtable stated “We see the need for a modest block size increase in order to move the Bitcoin project forward” and more importantly “Our shared goal is the success of bitcoin.” The first statement is self-evident if we are to meet the growing demand for transactions. The second statement; all of us need to embrace.
The scarce economic resource that we must expand is space within blocks for transactions
First we must realize that the Bitcoin ecosystem is in reality a transportation system, and therefore this is a problem of logistics. Confirmation of delivery or arrival, like one would receive from FedEx, is inclusion of a transaction in a block which is successfully and durably appended to the longest chain. This process is “messy”, requires consensus and is at the heart of the current debate. Increasing capacity in a way which remains accessible to all participants is the issue, not the actual size of the block alone. As demand for transaction delivery increases in relation to the theoretical limit, different participants will have different concerns and responses.
All participants are ultimately concerned about the capacity and performance of the Bitcoin network. Orphan and propagation risk interfere with delivery. Network topology, fees and network performance influence which participants have access. Each local node’s individual performance will impact their ability fully participate in the network. How cost efficiently the aggregate network can meet demand and append transactions onto the longest chain will determine Bitcoin’s appeal to merchants (recipients), miners (processors), and individual Bitcoin (spenders/senders/holders) users equally.
bitcoin or more precisely transactions are the parcels or passengers. Blocks are the vehicles analogous to passenger aircraft or buses. The origin of the travelling bitcoin are geographically dispersed, published at the direction of various types of individuals expecting to pay a fee. Miners load the vehicles (buses) with parcels (transactions) and are only paid when the loaded transactions are durably appended to the longest chain. There is only one destination that counts; the blockchain. When buses do not arrive it is because they lose their way or arrive too late behind a competing bus which arrives before the gate closes (i.e. wins the race). Those that arrive late have wasted their resources and must quickly change focus to the next race. The prize is granted only to the bus operators that deliver the parcels carried by their buses to the blockchain on time. The miners operate the buses and the bus terminals and have strong opinions about the configuration of the buses, the road network, border crossing, location of the bus terminals and generally accessibility (fees) to those who seek to purchase space on their buses. The space on the vehicles is actually auctioned at the discretion of the miners to those that wish to send transactions to the blockchain. I hope the similarity to the transportation system is apparent.
This analogy disregards the block rewards of today (25 XBT) and considers only the miner fees hence this analogy may be more valid as block rewards decrease and miner fees become more important. It also only considers valid blocks, or shall I say buses tagged with a completed proof of work after the buses have been loaded full with parcels (transactions). Proof of work is mostly an entirely separate issue.
Space on the vehicles is auctioned, but miners remain able to choose which parcels they carry, how full the buses are. However the blockchain, the bus operators and those purchasing space on the buses must agree on the specific configuration of the buses and their size. The road network and bus terminals and the blockchain must be able to accommodate the buses carrying the transactions to the blockchain. Mismatches between the vehicles and the underlying network’s performance results in propagation failures perhaps unevenly where geographic proximity or the lack of borders may infer an unfair advantage to some and disadvantage to other miners (bus & terminal operators). Larger blocks may travel more slowly or experience difficulty on bad roads and routes. For example, miners based in China, as a result of larger blocks (buses), may experience challenges in filling and delivering their buses to the blockchain due to geographic distance or borders (the great firewall of China).
If we are able to accurately predict the future, we can expect demand for transactions to approach and exceed the current theoretical limit of the Bitcoin network. The above analogy provides a model where different manipulations may be ‘mentally’ simulated. Let us try to iterate through the possibilities.
For the sake of simplicity, I omitted one very special class of participant. These are the full nodes which propagate/distribute information to connected nodes. They need not be miners, but they do transmit information about broadcast transactions for inclusion into blocks and propagate information about successful blocks with a complete proof of work and also make the contents of the longest known chain available to any connected node that needs or wants to know it in detail. The motivation for these participants who may mine, but who do not is not clear and likely weak. This is evidenced by the worrisome decline in full nodes that we have experienced as the local demands for maintaining the full blockchain has increased. It seems only the truly interested or altruistic will continue to run full nodes without expecting a monetary return for their allocated resources. Nonetheless, these participants are critical to the performance of the network even if they do not fill or operate the buses and bus terminals. They are also subject to the condition and performance of the road network as they fulfill this critical function of passing along information quickly.
I will examine four scenarios assuming the current state of the Bitcoin network. (1) demand below capacity, (2) demand at capacity, (3) demand just exceeding capacity and (4) demand grossly beyond capacity.
(1) In aggregate, those that broadcast transactions for inclusion into a block will find available space for their transactions. Transactions will confirm without delay if a competitive fee is included with the transactions. Miners will continue to compete for transactions and transaction fees and will continue to prioritize high fee transactions and as many of the rest as possible to the extent that they can produce blocks quickly enough and that they propagate quickly enough to have a fair chance at appending their block to the blockchain. Fees remain low’er’ since pricing power is low for the miners since competition remains strong. If the miners are unable to cover their costs and earn a return, they must either lower their cost base, cease operations, or develop advantages that permit them to win the race more predictably, earn the mining fees and today the block reward. The senders and receivers of the bitcoin are generally pleased because they can expect transactions to confirm without delay, for reasonable cost. If perceived as useful, they may increase demand for transactions. The miners, however, continue to rely on scale for economic survival.
(2) Demand = Capacity: This is the sweet spot. The miners have just enough pricing power to raise prices to meet their cost and earn a return, while competition is adequate to hinder broad price gouging. The true/natural price to include a transaction in the blockchain will be reflected here. In other words, the buses are filled to capacity, There are enough bus operators. Lines are short. Parcels are delivered on time. Prices are reasonable and the earnings of miners are reasonable. Truly free markets tend to revert to this condition as inefficient carriers do not survive, new entrants (miners) are attracted by the potential earnings, only meaningful transactions are worth paying for (little spam), and services (space for your transaction on the bus) are accessible to those that need or want them at a fair price.
(3) Where demand exceeds capacity, miners will have increased pricing power. Miners will be in the position to select the most lucrative transactions to include in their blocks leaving the least lucrative without immediate services. Those unserviced parcels (transactions) will linger in the mempool until there is empty space in a subsequent block at a time when demand is transiently below capacity. If the miners expect high demand to return, they may continue to exclude these lingering non-lucrative transactions in an attempt to maximize their local performance and increase their chance of winning the next block with less work (fewer transactions/smaller block). Naturally, this would frustrate those needing or wanting space in a block. Those broadcasting transactions would either abandon the network reducing demand or increase the included transaction fee to prioritize their transactions. The line for the bus would get longer, the price would be higher, but some would abandon the bus network reducing demand. On the other hand, miners may push to increase capacity because there are additional potential transactions fees to earn (unsatisfied demand). As long as those who need or want to send bitcoin can send them without delay, the utility of the Bitcoin Network would be preserved.
(4) Demand that grossly exceeds capacity would be seriously detrimental to the Bitcoin network. Transaction costs would be high due to the miners very high pricing power. The lines to get on the bus would be long and inclusion into a block would be unreliable without high fees. High fees and delayed delivery would destroy the utility of the Bitcoin network. If that were to become apparent to all the participants, they would all eventually abandon the network. Users would prefer networks that work like visa and paypal rather than Bitcoin. Demand for a seat on the bus would collapse. The blocks would again have free space at a reasonable price if Bitcoin were to survive this demand collapse. If demand were not to return, the miners would also abandon Bitcoin seriously harming the security of the network with the reduction in hashing power. This scenario may be unsurvivable.
The above framework should also allow us to mentally model what different manipulations/events affecting the network might result in such as Larger blocks, SegWit, etc. Mining centralization could be reflected as a reduction in the number of bus operators clustered in a specific geographic area. It could provide insight on whether Bitcoin would remain useful for the transfer of funds as various levers within the system are manipulated.