Credits: info@ornaat.nl

A shortcoming of blockchain: the scalability trilemma

Sander van Laar
3 min readApr 16, 2019

--

If you work in IT, then you have probably heard of the CAP theorem: for distributed systems it is impossible to satisfy all three aspects of Consistency, Availability and Partition tolerance at the same time.

Since blockchains are distributed systems, this theorem also applies to blockchain, however blockchains have to deal with a more complex property, the so-called “Scalability Trilemma”, introduced by Vitalik Buterin.

The Scalability Trilemma

Distributed Ledger Technology (DLT), which contains blockchain technology, has three characteristics: scalability, security and decentralization. The Scalability Trilemma is defined as: it is impossible to develop a blockchain solution that simultaneously meets all three properties scalability, security and decentralization.

In 2008, Satoshi Nakamoto published a white paper describing Bitcoin (and thus the first blockchain). This white paper describes an innovative combination of existing technologies:

  • Peer-to-peer networks (as torrents): scalability, decentralization
  • Public-key cryptography (as in SMTP): security
  • Consensus on globally distributed systems: decentralization

However, this blockchain and other newer generations of DLT all have to deal with the Scalability trilemma.

The Scalability trilemma consists of the following three aspects.

Credits: info@ornaat.nl
  • Scalability: the efficiency, the speed (in transactions per second) at which transactions are executed on the blockchain,
  • Security: the correctness of the blockchain, the ability to guarantee the integrity of the blockchain, including protection against malicious attacks,
  • Decentralization: the number of participants to validate the transactions in the blockchain, using consensus. In a fully decentralized system all…

--

--

Sander van Laar

Blockchain specialist, entrepreneur and consultant. Former university lecturer in computer sciences and data analytics.