Synchronous vs Asynchronous Models
What you'll learn: How bounded vs unbounded communication delays affect the feasibility of reaching consensus in distributed systems.
The Core Distinction
In distributed systems, the synchronous model assumes bounded delays—there's a known maximum time for message delivery and processing. Think of it like a well-organized postal service where letters always arrive within 3 days. You can safely wait 3 days, then assume a letter was lost if it hasn't arrived.
The asynchronous model assumes unbounded delays—messages can take arbitrarily long to arrive (but eventually do). It's like sending a message in a bottle across the ocean: it might arrive tomorrow or in ten years. You can never distinguish between a slow message and a lost one.
Impact on Consensus
This distinction profoundly affects consensus feasibility:
Synchronous systems make consensus solvable even with failures. Because you can detect timeouts reliably, you know when a node has crashed. If a leader doesn't respond within the time bound, elect a new one! Algorithms can make progress with guarantees.
Asynchronous systems make deterministic consensus impossible with even one crash failure (remember the FLP impossibility result from lesson 2650). You cannot reliably distinguish a slow node from a dead one, so algorithms must either risk waiting forever or risk making incorrect decisions.
Real-World Reality
Most real systems are partially synchronous—they behave synchronously most of the time but occasionally experience asynchronous periods during network issues. Practical consensus algorithms like Raft and Paxos rely on this: they use timeouts for failure detection (synchronous assumption) but remain safe even when timing assumptions are violated (asynchronous safety).
Key Takeaway: Synchronous models with bounded delays enable reliable failure detection and consensus, while asynchronous models with unbounded delays make deterministic consensus impossible—understanding this trade-off is crucial for designing robust distributed systems.