Actor Model

The Actor Model is a mathematical model of concurrent computation, first described by Carl Hewitt in 1973, in which the universal primitive is the actor. An actor is a lightweight, isolated unit that holds its own private state, communicates only by sending and receiving asynchronous messages, and can create more actors in response. Because nothing is shared and everything is a message, the model sidesteps whole categories of locking and race-condition problems that plague thread-and-lock concurrency.

• Isolated private state - No actor reaches into another’s memory; the only way to affect an actor is to send it a message.
• Asynchronous message passing - Actors process one message at a time from a mailbox, decoupling senders from receivers in time and space.
• Location transparency - An actor address is the same whether the actor is in-process or on another machine, which makes the model a natural fit for distributed systems.
• Supervision and resilience - Actor hierarchies let failures be contained and restarted (“let it crash”), a pattern Erlang/OTP and Akka made famous.

In API operations the Actor Model shows up wherever a system has to handle enormous concurrency without falling over — telecom switches on Erlang, streaming and messaging backends on Akka, and the stateful cores behind many high-throughput APIs. Its message-passing discipline maps cleanly onto event-driven and asynchronous API styles, and as the agentic turn pushes toward many small, stateful, independently-failing units of compute talking over well-defined channels, the actor’s mailbox-and-address abstraction is a useful mental model for reasoning about agent-to-agent communication and back-pressure at scale.