Distributed computing systems face a fundamental challenge: when multiple components must coordinate actions, the order in which messages are processed can be safety-critical. This talk uses an aircraft door control system as a motivating example to show that prevailing architectural patterns — pub-sub, actors, and service-oriented architectures — force application designers to solve distributed consensus problems that should instead be handled by the underlying framework.

Lingua Franca, an open-source polyglot coordination language, addresses this by assigning logical timestamps to messages and processing them in order. A single parameter, maxwait, gives system designers explicit control over the consistency-versus-availability tradeoff: larger values favor consistency at the cost of latency, while smaller values favor responsiveness at the risk of out-of-order execution. This unified mechanism subsumes a surprisingly broad range of classical distributed computing patterns, including Chandy-Misra, fault-tolerant variants, ACID/CRDT/CALM, optimistic execution with rollback, Logical Execution Time, remote procedure calls with futures, pub-sub, and actors.

Critically, all of these capabilities depend on clock synchronization with bounded error. This talk argues that precise, reliable clock synchronization is not merely an infrastructure concern — it is a foundational enabler of correct, deterministic distributed software.