Understanding the scalability trilemma requires grasping the basic premise of blockchain technology.
Georgios Konstantopoulos eloquently defines a blockchain as
“…a database that can be shared between a group of non-trusting individuals, without needing a central party to maintain the state of the database.”
Why would we want to get rid of a central party, some may ask. Doesn’t that bring stability and consistency? Yes, however with a central party also comes centralized power: the power that is often used by it (or appropriated by someone else, such as a hacker) to take advantage of others through deception, fraud, etc. You know the saying: power corrupts.
Among the younger generation, the shift toward blockchain technology is propelled by an idealistic passion to remove centralized power. It can be seen as the originator of countless injustices – on all scales – over many years. In its purest form, blockchain technology even eliminates the possibility of corruption by enabling the creation of immutable public records. This promise of blockchain is premised on its ability to achieve a high level of security and decentralization.
However, several projects claiming to employ blockchain are somewhat lacking in decentralization. This reality is more a matter of degree rather than kind. The argument regarding how much decentralization is required is analogous to that of political systems, in that the more decentralized power that exists, the less the likelihood of corruption. Likewise, the more decentralized a political system becomes, the slower it is at making decisions. Due to the subjective nature of this topic, there will always be purists and pragmatists at each end of the spectrum arguing their cases. There usually has to be some trade-off to allow the system to function; this is the essence of the scalability trilemma.
The Scalability Trilemma
One of the factors inhibiting the widespread adoption of blockchain technology is its failure to scale (handle the throughput requirements for a large set of users). Being early blockchains, Bitcoin and Ethereum have always focused on the core aspects of security and decentralization. So as to ensure security, all full nodes on these blockchains must reach consensus before they accept a transaction. This requirement has resulted in a very minimal rate of transactions per second (TPS): around 7 for Bitcoin and 15 for Ethereum. To address the issue, Ethereum developers are working on scaling solutions that they believe will provide scalability without sacrificing security and decentralization (see “sharding” & “Casper”). Developers are also working on layer 2 scaling solutions such as the Lightning Network for Bitcoin.
The three aspects of scalability, security and decentralization form what Ethereum founder Vitalik Buterin labeled “the scalability trilemma.” All three are required, yet only two can be obtained concurrently. At presently constituted, if you want a fast blockchain, you have to be willing to give up some decentralization. Ripple and EOS are two projects which provide higher transaction speeds by sacrificing decentralization. As a result, these blockchains have minimal censorship resistance: for example, powerful entities are capable of freezing the funds of any use (Ripple, EOS). This goes against the very definition of
How Fast is Fast Enough?
Crypto enthusiasts are quick to draw comparisons with VISA, which has achieved 11,000 TPS in real-world conditions and over 24,000 TPS in stress tests. Though on average, the VISA network processes only around 2,000 TPS.
Moreover, VISA only processes transactions that occur within the VISA payment network. It doesn’t include ordinary bank-to-bank money transfers, PayPal transfers, etc. For any blockchain project aspiring to become the undisputed king of global money transfer, it will need to process not only VISA transactions but various other transaction types as well. And even if it does obtain that magic TPS number, the increase in network speed will not impact the consumer experience. Instead, it will be user support, accessibility, and marketing that determines widespread adoption.
Developers from many projects are hard at work attempting to solve the scalability dilemma, and some newer projects have claimed that they already have. It’s still very early days though, and no solution has been tested widely under real-world conditions.
Perhaps the most likely response to this problem is multiple fit-for-purpose blockchains. That