7 Misunderstandings of Bitcoin’s Scaling Problem
If you prefer to listen, an audio version is available Here.
I did, for quite some time, fail to see the real scaling hurdle that faces Bitcoin. It was all about the blocksize and how many transactions we could get on-chain. Blocksize just jumps out at you, before anything else, as the obvious and most direct corollary to how much we can accomplish with the blockchain. So how could someone argue that such a thing may be one of the least important parameters for Bitcoin’s success?
Being a very early adopter of this technology I’ve gotten used to it operating a certain way. I’ve become comfortable with what a transaction is, how long I have to wait, and am generally familiar with all of the software and means of interaction. Someone coming along and saying my interaction with it is going to have to change throws up a red flag for me. And it did, when I first heard the suggestion that blocks should be kept small. How the hell am I supposed to buy stuff on NewEgg or Amazon if transaction fees are high? Bitcoin becomes unusable doesn’t it?
So, whenever I get worried, nervous, or unsure of what comes next, I read. I consume anything and everything I can get my digital hands on. I promise you I have consumed and assessed nearly every possible argument for and against a blocksize increase, as is easily discoverable on the internet. And after hundreds of nuanced, fascinating, heavy-handed, belligerent, thought-provoking, or outright idiotic arguments, I have firmly come down on the side of keeping the blocksize as small as we can manage. Before you get angry and start deciding the reasons why I’m stupid (or which bank is funding my article), just hear me out.
During my journey from infinite blocksize to “maybe we can get away with no direct blocksize increase at all,” I went through a series of critical realizations. Here they are:
1 • What size block solves high fees?
When diving into this debate you will inevitably run into the comparison to transactions per second (TPS) handled by VISA. It seems very logical at first. People shop and make transactions the same way they always have at a predictable, albeit increasing, rate. You buy a magazine, you go out shopping, you get gas, and you buy your morning coffee. These are the normal situations where a transaction occurs, and Bitcoin, if successful, would have to handle all of that activity by some method. The obvious conclusion would be that we need a blocksize that can accommodate transactions for every individual who wants to use it… right?
Here’s where something was nagging at me. Think about just the handful of revolutionary use cases Bitcoin can deliver at this extremely early stage. Development teams are, at this very moment, building decentralized cryptocurrency exchanges, IOT networks for devices to trade and share resources, smart contract ownership models, trustless gambling and prediction networks, social media with built-in selling and tipping, and networks for exchanging computational resources in real time. These are only the first and most obvious use cases for the Bitcoin technology. We are still probably a decade away from the Google, Amazon, and Facebook game changers in the development timeline.
From that perspective, using VISA as a barometer for transactions, is analogous to predicting today’s email traffic by counting the number of faxes sent in the 80s. We are talking millions or billions of transactions, not thousands. If Bitcoin is to be the revolution it promises to be, we are talking about scaling to orders of magnitude higher throughput than VISA.
There is no feasible way to create a widely distributed, agile, and censorship resistant network where every single node has to store the entire history of this enormous transaction potential. No sustainable blocksize will ever accomodate every user’s transactions on-chain, and certainly not while requiring only a $0.01 fee to do so. Even VISA, with massive central servers, charges the merchants a $0.20-$0.50 fee per transaction.
The scaling problem is an exponential one, and simply can’t be solved with a linear solution. A 100 MB blocksize isn’t a drop in the bucket of the scaling problem, it is a drop in the ocean.
2 • Is there any way to safely get 1000x capacity?
As I hope I illustrated with #1, our scaling challenge isn’t trivial. If a bigger blocksize is our only hope of accommodating this scale, then we have already lost. The only way Bitcoin can be a universal and widely used currency is if we build secure layers on top of it. If we can get 1000x the capacity right now through layered protocols like lightning and sidechains, without any consensus breaking blocksize increase, then there is hope that a decentralized currency AND payment system is possible on a single, highly layered chain. But, if we can only get 2–3X, then raising the blocksize is just delaying the obvious truth that this technology can’t scale to a global, high speed payment system while sustaining independence and decentralization. If that is the case, I say a small censorship resistant network (providing a secure monetary backbone to traditional services) solves a far more difficult and historically significant problem (and therefore is vastly more valuable), than yet another Paypal that merely replaces the merchant paying the transaction fee, with the consumer paying the fee.
3 • Are Lightning/Sidechains just another trade off?
In their dogmatism for larger blocks, many “hard forkers” have foolishly taken to fundamentally attacking some of the most exciting innovations in the cryptocurrency space. Sidechains and layered solutions like the lightning network are highly desirable regardless of scaling advantages. The true promise of these layers are new features, instant and cheap transactions, and advanced smart contracts. If we want distributed business models and complex agreements enforced by the blockchain, then we need specifically designed sidechains and layered protocols that take the underlying security of Bitcoin and extend it to a unique sets of rules or network topologies.
I am often led to believe (by some hard fork supporters) that the handful of core developers associated with Blockstream are in opposition to larger blocks simply and only because it would somehow eliminate the use of sidechains or layered scaling like the lightning network (aka, make Blockstream development work unnecessary). I find this claim rather absurd with even a rudimentary exploration of Bitcoin’s scaling problem, or a brief examination of the potential of layered protocols and sidechains.
4 • Is the Bitcoin blockchain really a better payment network?
As I explored in my article “Bitcoin is valued for fast and cheap payments,” Bitcoin still has some major hurdles to overcome as a payment network. And I ask the critical question of what Bitcoin’s value proposition really is.
What I have slowly come to realize, is that Bitcoin is a rather poor payment network, but it is a revolutionary breakthrough as an arbiter. Bitcoin transactions, even at the very beginning, have always been unreliable and slow in comparison to a credit card or Paypal transaction (see the above article for a personal recount). Even very early on, when fees were $0.05, it has always been more expensive for the user than nearly every major centralized alternative. On-chain transactions have never been well suited for in-store purchases, and only work reasonably well for online purchases.
According to those who think consensus is achieved and secured through mining power alone, the entire security of the Bitcoin network rests with… 6 companies.
On the other hand, as a dispute arbiter, the Bitcoin blockchain represents a veritable revolution compared with current alternatives. So far, no institution in the world has been capable of challenging the outcome of a contract or transaction executed on the Bitcoin blockchain. For any system that can be adapted to a programmable contract (essentially all business/trade contracts could eventually be enforced this way), there is no service-based immutability, trust, or independence that can come close to matching what is achieved by a global and decentralized blockchain. However, its security is entirely dependent on the resiliency and distribution of rule-enforcing nodes. Which leads me to my next realization…
5 • Bitcoin’s resilience is dependent upon the distribution of validators.
If miners were the only players who secured Bitcoin consensus (as some believe), Bitcoin would be no significant improvement over the legacy banking system. Basic peers would hold no power over the monetary rules that they followed. All value and consensus would be with a very small group of highly visible, well-funded participants.
Currently, 6 mining institutions represent the hashing majority (nearly 60%) of the network. According to those who think consensus is achieved and secured through mining power alone, the entire security of the Bitcoin network rests with… 6 companies. IMO, such a system would be as arbitrary as the Federal Reserve, easily controlled, and could in no way be considered “decentralized.”
The reality is that a large part of Bitcoin’s consensus is more akin to a pre-established relationship between miners and validators. Nodes establish which is the valid chain, and miner incentives work to extend the chain that has the most economic reward (i.e. whichever is most valuable). The Bitcoin Cash fork is an excellent experiment that proved these principles in practice. Constant shifts in mining profitability (thanks to the EDA) resulted in immediate correlating shifts in the hash power. If a majority of nodes and economic players stake a claim to a specific rule set, all they have to do is wait, as those rules allow a substantial reward to any miner who creates new blocks while adhering to them.
What we have witnessed with this technology, is that the degree of mining power working with nodes only affects the frequency of new blocks in the short term, the difficulty adjustment in the mid-term, and has very little long term effect. If miners change their rules, they simply create a separate ledger which all full nodes will ignore by default, unless they also specifically alter their software to match it.
To put it simply: A Bitcoin miner cannot decide which network is more valuable anymore than a miner of precious metals decides which one will have a higher market price.
6 • Is large block latency and resource cost really negligible?
What many don’t realize unless you look very closely at the current Bitcoin network, is that 1MB blocks are already resulting in security trade-offs that begin to compromise the robustness of the validation process.
Right now, many large miners use direct channels and rely only on the block header as a means of accepting valid blocks. Because of just 1MB of data, miners compromise validation (in albeit a seemingly minor way) in favor of expediency and begin mining the next block before validating the previous one.
This already puts large, institutional miners at an advantage if they can communicate directly with other large players. In this way, they both get a head start from faster relay, and don’t have to perform a full validation of new blocks. Now, maybe this is only a 1% cost advantage, maybe it’s 5%, but every single time we raise the network resource costs, the gap is worsened. Wal-Mart may only have a few percent advantage in their economy of scale distribution, but that small advantage puts tens of thousands of competitors out of business.
An agile network with low resource costs, with 1,000,000 validators across the globe incontestably securing the network’s monetary principles, and with quickly deployable nodes is one that could isolate the manipulation of money for political gain to the pages of history.
A network where nodes require significant resource costs, where nodes are inconvenient and slow to deploy, and where miners are considered the only important enforcers, will give no significant resistance to the political desire to control the world’s money… and worse, it still won’t have low fees.
7 • How will scaling affect Bitcoin privacy?
Lastly, on-chain transactions are bad for privacy. Each centralized service in the Bitcoin ecosystem is a point of data that can connect an entire history of transactions and balances to their owners and associates. Many are not aware of this, but the company Chainalysis has partnered with the IRS to identify and connect all bitcoin users with their addresses. The company claims to have reliable data on the owners of over 25% of all bitcoin balances on the network. Along with a number of other companies, like Jeff Garzik’s Bloq, they are built upon selling the connections between Bitcoin balances and their real world owners.
The more on-chain transactions, the easier it is to establish these connections and the more information there is to sell. The fact that Garzik is the same person who has been one of the loudest voices in the “big blocks immediately” campaign should throw some heavy salt on trusting his intentions.
Off-chain and layered solutions offer far greater opportunities for privacy, which is critical to protecting innocent people from oppressive governments and centralized data collection. Bitcoin is a global, decentralized currency, the only barometer for success is its security from the worst governments, not the best.
The bigger-as-needed block argument, by my consideration, tends to follow a “strongest link holds the chain together” mentality. As long as we get more users, a small hit to resiliency, number of nodes, speed of node deployment, network latency, miner decentralization, and bandwidth/storage costs to validate are all acceptable. Yet a hit to each of these characteristics, even a small one, could have not-so-insignificant consequences in the long term. A 10MB blocksize today would barely graze the real scaling problem, but in 3 short years would have us securing a blockchain nearly 2TB in size and growing at a far faster rate.
However, the small block mentality recognizes that a weakest of any of these characteristics is the best possible measure of Bitcoin’s security. Bitcoin must be secure in every way to be secure at all.
I agree wholeheartedly that, in general, high fees are undesirable. But as a regular and long time user of Bitcoin, I am fully willing to pay higher fees to know with the highest certainty, that Bitcoin remains the world’s most secure and trustless monetary system. Too many underestimate the profound effect that even a small change in the foundations of an economy can have. Bitcoin has the potential to unlock a torrent of productivity and freedom for the entire world at a time when it is desperately needed. I’m not willing to risk losing that over a $2 transaction fee. So, spin up that full node you have been procrastinating, and join the revolution.