As banks continue to invest in blockchain, should we talk about the elephant in the room: high transaction volumes?
Yesterday, I was in Café Nero considering whether to order a delicious almond croissant (or whether the extra 400 calories were a bad idea for my upcoming ski competition). As I scrolled through LinkedIn, I noticed an article on ‘blockchain for trading’. The question that came to mind was, why are people not projecting and extrapolating information to verify the feasibility of using blockchain?
I can’t help feeling that banks are investing in blockchain because of hype, fear and impulse rather than considering potential outcomes. They are banking (no pun intended) on blockchain becoming the solution to everything from fraud prevention and supply chain optimisation to processing transactions. Santander recently launched a blockchain-based foreign exchange service and have predicted that blockchain could save up to $20 billion a year in infrastructure costs by 2020.
Blockchain struggles
It might sound like the numbers add up, but let’s look at the facts. Blockchain struggles to produce more than 30 transactions per second, while a global bank creates billions of transactions per day. If we are being conservative, let’s say it’s one billion transactions per day.
1 billion / (12 hours * 3600 seconds = 277,000 seconds). That gives us around 23,000 transactions per second. Blockchain can create 30 transactions per second vs. a difference of 22,970.
To Reduce the blood pressure
To reduce the blood pressure of blockchain believers: Yes, there are excellent use cases for blockchain, but not currently in areas that require high a volume of transactions. We must use projections when assessing solutions that use new technology. I do this on a daily basis, whether it’s considering how technology can process data or deciding whether I should have that almond croissant.
If we look at blockchain and project what we know, it makes sense to apply the principle of linear performance. Linear performance means that run times can be held static even if the transaction volumes increase by 10 or 100 etc. We do this by using this formula on the platform. If an application has a processing time of X and the transaction volume increases by Y%, we then increase the processing capacity by Y% and processing times remain static.
If we take the volume of 30 transactions per second and project that for a bank, blockchain would fail. The size of the transactions would become so large that transporting them between servers and storing transactions would become a huge issue. Blockchain might even fill up the internet. Projecting the limitations tells me that blockchain in no way meets the requirements for high volumes of transactions.
A Platform must Scale
When it comes to creating a banking platform for the next century – whether it ’s for FATBAG (Facebook, Tencent, Baidu, Google, Amazon, Apple) or a legacy bank – the platform must scale. At the moment, the only conclusion I can come to is that blockchain is unsuitable for high volume transactions unless there is a way to apply a meaningful capacity increase.
There have been recent attempts to scale blockchain technology systems, but these have either proven to be short-term fixes or raise further challenges. What we need is a fundamental shift in the way we think about transactions. I believe the banks are investing in blockchain, subconsciously knowing that they are wasting time and money. They know that a major change is needed, but they don’t know how. Learning about blockchain and investing in blockchain are two different things.
At the moment, it feels like banks are doing the equivalent of eating the croissant and hoping to find a way of burning off the calories later.
