Bitcoin: Too big to scale?
With about 11.000 mining nodes Bitcoin is the largest blockchain network out there. However, is it really scalable? What does it mean for a blockchain to scale? And what are approaches to improve Bitcoin’s scalability? That’s what this blog post will dive into.
The scalability of a blockchain platform can be understood as the extent to which the network can add more participants without a loss in performance. Performance here refers to the number of transactions processed per second impacted by the latency between transactions and each block size.
Scalability can further be divided into performance- and node-scalability.
(1) Performance scalability refers to the latency of the network, that is, to the number of transactions per second.
(2) Node scalability refers to the number of nodes the network can add.
With respect to node scalability, the Bitcoin protocol is perceived to scale relatively well as it currently features thousands of mining nodes.
With respect to performance scalability however, the Bitcoin protocol is very limited. The average throughput amounts to 3–4 transactions / second. Block creation and validation take on average 10 minutes per block and thus lead to relatively long waiting times.
But once I waited 10 minutes for my transaction to be included in a block, I can be sure that it is valid, right?
Unfortunately, not quite. In the case two or more miners validate a block almost simultaneously, they both send their blocks to other nodes in the network.
As a result, different nodes in the network end up holding a different version of the blockchain. This means that the original blockchain is forked and two versions exist in parallel.
This remains the case until the length of one of the chains supersedes the other one significantly. As a rule in the Bitcoin protocol, the longest chain is then accepted by all nodes in the network and the other one orphaned to re-establish consensus.
Therefore, in order to be sure with high certainty that a transaction is really validated in a block, it is suggested to wait for 6 more block confirmations. Consequently, the time to validate a transaction with high certainty can result in waiting times of several hours.
Think about what this means for an everyday payment with Bitcoin? If you pay your groceries at the local store with bitcoin, the store owner would need to wait several hours to be sure the transaction is valid and he actually received the payment.
What limits Bitcoin’s scalability?
Factors that limit Bitcoin’s scalability are network bandwidth, network size, and storage requirements. As every node needs to hold a copy of the full ledger, participating and verifying transactions in the network becomes increasingly computational expensive with the increasing amount of transactions.
Different approaches exist to solve issues related to scalability. One is the implementation of a hierarchical network structure in which supernodes (miners), full nodes (exchanges), and wallet nodes (wallets or thin clients) each download the blockchain in different detail. This approach could help to handle an increased network load from higher transaction rates.
Furthermore, approaches to increase the block size from the current artificial size of 1 MB exist. This is what Bitcoin Cash did. It increased the block size to 8 MB and can thus handle 8 times Bitcoin’s transaction volume as Bitcoin. Keeping in mind that Bitcoin currently handles about 4 transactions/sec, this means 32 tx/sec.
So why don’t we just increase the block size?
An increase in block size to about 500 MB would result in an increased transaction rate of up to 2000 tx/sec. However there is a high risk for doing so: Larger blocks would take longer to propagate through the network and consequently would increase the probability of forks in the network when different validators send blocks through the network that conflict each other. As a consequence of an increased fork rate, an attack on the network becomes more likely.
Another approach to scale the Bitcoin blockchain aims at shortening the block generation interval. E.g. instead of creating blocks every 10 minutes, these approaches aim to cut the time down to every 1 minute or even seconds.
Both approaches, increasing the block size and shortening the block generation interval, come at the cost of increasing the blockchain fork rate and consequently lower Bitcoin’s security. Higher vulnerability to hacks are the consequence.
As a result, performance scalability remains an issue in the Bitcoin blockchain and limits Bitcoin’s ability to become an alternative, global banking model. For comparison: Visa processes on average 2.000 tx/sec, Bitcoin 3–4 tx/sec.
Further approaches to scale Bitcoin
As indicated with Bitcoin Cash, there are different ideas on how to improve scalability. Apart from increasing the block size, a promising idea is the use of so called “off-chain” solutions. In an off-chain system, participants can handle high numbers of transactions in non-blockchains systems. For instance, they could use messaging networks such as WhatsApp to send money. The final result of these transactions is then settled onto the blockchain. An example for such a solution is Blockstream’s Lightning network.
A consequence of such solutions are a reduced transaction volume needed to be processed on the blockchain. Less transactions would help reduce the need to increase block sizes. It also would make settling many small payment amounts on the blockchain possible, because high transaction fees for each transaction can be avoided when settling off-chain.
Therefore, such technology is very important for IoT projects on the blockchain, when for example your fridge pays the neighbor’s solar panel system a tiny amount (e.g. Euro Cents) for power consumption.
As seen recently with Bitcoin Cash and Segwit2x, two approaches that among other features aimed to increase the block size, making decisions about which proposals to implement can be very cumbersome in a decentralized blockchain network. Consequently, Bitcoin’s governance plays an important role for the network to improve scalability in the future.
Thus, it is important to ask who makes what decisions in a network with a market cap of roughly $156 Billion USD (November 2017)? To find out, see here.
To see an overview of all articles, go here.
