There’s one fallacy that’s not a part of the above set, but still often causes confusion amongst people new to distributed systems. . If we follow the same style as above, we can phrase this fallacy as:

“Distributed systems have a global clock, which we can use to identify when events happen.”

This assumption is quite deceiving since it’s somewhat intuitive and holds true even in non-distributed systems. For instance, an application that runs on a single computer can use the computer’s local clock to decide when events happen, and in what order. However, this is not true in a distributed system, where every node in the system has its own local clock that runs at a unique rate.

While there are ways to keep the clocks in sync, some are very expensive and don’t completely eliminate these differences. Physical laws also bind this limitation. An example of this is the TrueTime API built by Google, which exposes the clock uncertainty explicitly as a first-class citizen.

However, as we’ll see in the upcoming lessons that discuss cause and effects, there are other ways to reason about time using logical clocks.