Trust Models

Describes different systems that require reliance on others by plotting across two axes—how many people need to behave correctly out of how many for the system to work.

Examples:

  • 1 of 1 is the traditional centralized model—you rely on a single entity to behave well e.g. Google
  • N of N every actor needs to act as expected for things to work
  • N/2 of N is blockchain where the majority of miners are honest
  • 1 of N is a load balancer with liveness checking—only 1 of many need to be operating as expected to work
  • Few of N there are many actors and as a small fixed number act as expected
  • 0 of N there is no reliance like checking hashes yourself for validity

See also:

In order for an economy to prosper, there must be a high degree of trust in the system. For instance, without trust that someone will uphold their side of a contract, only limited kinds of transactions can occur. If there is trust that a contract will be enforced then there is higher assurance the other party will make good, making it possible for more complex arrangements and entities outside the system can more easily participate.