Bitcoin Doesn’t Scale: What’s Next?
A number of things have come to pass since my early 2015 post (below the line) including Bitcoin forking in order to increase transaction speed. This happened suddenly, after what appears to be a reversal of policy by China whose companies control 80% and all 5 of the largest bitcoin “mining pools.”
Chinese companies blocked bitcoin forking for an extended period except Alibaba’s ANTpool. Suddenly all five approved the XT and then Classic fork and voila. What happened? We can only speculate, however it appears the Chinese government decided it was better to control Bitcoin than attempt to block it. So yes, Bitcoin is a “distributed” governance model but make no mistake that Bitcoin governance stops in Beijing.
To be sure, further forking to significantly increase transaction throughput is necessary with “segwit” currently being the leading candidate. There are proprietary workarounds such as Coinbase’s Thunder however these are not universal. There is only one voter that matters in the real contest.
___________________________ ORIGINAL POST BELOW _______________
(February 2015) That’s a pretty bold statement which I’ll explain here, along with some thoughts about evolutionary paths for Bitcoin and other digital currencies.
Bitcoin’s technology is based on what is known as a “blockchain” which is a record (think ledger) of every Bitcoin ever “mined” (created) and every transaction and owner thereof in numerical form. Got that? The blockchain encapsulates every single transaction to every single Bitcoin from and to every single owner and repository since the invention of Bitcoin.
Bitcoin’s blockchain is not a centralized database like credit cards or bank accounts. Copies of the blockchain are kept by every Bitcoin repository or exchange — there is no “central master record” because a core part of the fundamental distributed tenant of Bitcoin was to avoid a “central bank.”
A year ago, the blockchain was approximately 13GB. Today it is 26GB. The number of Bitcoin account holders has quadrupled to about 6 million during that time. Yet there are only 86,000 daily Bitcoin transactions (5.6 transactions per second) and that is up just 50% from a year ago. In other words people are largely buying or selling Bitcoins as an investment but rarely buying stuff with them. The real challenge is that, without fundamental changes to its blockchain technology, Bitcoin has a hard limit of 7 transactions per second due to technology issues outlined below below.
But Bitcoin is posited to be an easily useful currency –so what would happen if Bitcoin did become a currency that transacted on the level of cash or credit cards?
Each Bitcoin transaction adds about 200 bytes (0.2k bytes) to the blockchain: sender & receiver account identities, and the unique ID of every single unique BTC (and fraction thereof) involved in that transaction. So today, 2.6M Bitcoin transactions per month x 0.2k bytes = +520 MBytes per month added to the blockchain.
Mining adds significantly more to the blockchain as it becomes more complex. Each minute about 20kBytes is added to the blockchain by mining = 864MB per month. Close enough to round up to 1GB per month. This matches the current ~1.5GB per month blockchain growth rate noted above.
Now let’s look at what the block chain growth rate would be if BTC transacted like credit cards.
The average monthly credit card transaction volume is 10,000X greater than Bitcoin at 20 billion transactions per month. 20B transactions per month x 0.2kBytes per transaction would cause the blockchain to grow 4 Terabytes (TB) per month. The blockchain would reach 50TB in just over a year which is impossible to nimbly move and amend by every merchant and account keeper. Even running at just 1% of credit card volumes would create a block chain of 500GB in a year. This is the reason that BTC is presently hard-limited at 7 transactions per second.
What does this mean? It suggests BTC, in its present distributed incarnation, cannot be a mass payment mechanism because you can’t process a transaction without the buying and selling accounts having access to the entire realtime blockchain. For example, I acquired a few Bitcoin three years ago and they have sat in the same account since. If I bought something with these Bitcoins, the transaction would need to reach back through three years of blockchain records to find my ownership and BTC numbers in order to authenticate my ownership and the deal. This is why the blockchain must be read and updated for each transaction.
So that drives us to the notion of whether scalability requires some kind of central repository (or a few repositories) with the “master” blockchain. That starts a real chain of questions because, by definition, Bitcoin has no central governance. Who would keep that (those) master record(s)? Would it be one or more of the “big 5” mining cooperatives? Will efforts such as the Coinbase’s new regulated exchange (which can, of course, support any digital currency) help, (disclosure I am an account holder but not an investor)? Or do efforts like these turn Bitcoin into a pseudo-central banked oligopoly of the kind that Bitcoin was designed expressly to countermand?
It remains to be seen whether a later generation digital currency that is designed to scale better, supersedes Bitcoin through market forces or whether Bitcoin’s efforts to evolve stay ahead of the curve. A risk is a Bitcoin blockchain “fork” will fail technically or fail to be supported. What then? One of the contenders is Ripple aka XRP (formerly known as OpenCoin), the second largest digital currency by value and the largest in the foreign exchange trade (disclosure I also own XRP but am not an investor in the company) due to its speed and fluidity in exchange (Bitcoin transactions take minutes due to the unwieldy blockchain.). It uses a consensus ledger managed by a small group of authorized exchanges, and as a fixed total # of Ripples were issued on day one, there is no mining load on the ledger. Fed officials have commented positively on Ripple’s emergence and its refinement of Bitcoin roots.
In the meantime, does this relegate Bitcoin to be a reserve investment instrument like gold which appears to be its main use / value today as transactions are growing more slowly than the number of Bitcoin accounts. Fiatleak dramatically shows us (in realtime) that Bitcoin purchases have already become concentrated to a very few countries. And would such a digital currency have any value because it lacks real utility unlike other reserves like gold which is physically beautiful and genuinely useful as an electric conductor that does not oxidize.
There are few articles on scalability of Bitcoin which should be important to any digital currency user or investor. This, from always excellent O’Reilly is among the best. Regardless, solutions must quickly move beyond white papers if Bitcoin is to scale to its market potential. (Mahalo to Ryan Ozawa for contributing to these thoughts.)
Let the discussion begin.