Definitions and setup

Shared state, blackboard, message passing

• Shared state: data that several agents can read and write.
• Blackboard: shared state used as the main coordination surface, where agents post and read partial results.
• Message passing: agents communicate by sending messages; no agent mutates another's state directly.

Lost update, last-writer-wins, race condition

• Race condition: the outcome depends on the timing of concurrent operations.
• Lost update: two concurrent read-modify-write operations; one write overwrites the other.
• Last-writer-wins: the store keeps the most recent write and silently discards the rest.

Optimistic concurrency, compare-and-set, reducer

• Optimistic concurrency: proceed without locking, and at write time check that nothing changed (a version number); if it did, redo the work.
• Compare-and-set (CAS): write only if the current value or version matches what the writer read.
• Reducer (merge function): a rule that combines concurrent updates, such as append to a list, add to a counter, or union of sets; it is safe when the operation is commutative and associative.

Delivery semantics, idempotency key, ordering

• At-most-once: send once; a lost message is lost.
• At-least-once: retry until acknowledged; messages may arrive more than once.
• Idempotency key: a unique id the receiver remembers, so applying the same message twice has the effect of applying it once.
• Ordering: sequence numbers let the receiver apply messages in the order they were sent even if they arrive out of order.