Skip to main content

Total Order Broadcast

It's the idea where all nodes from a distributed system receive the same messages, either by logs or messaging, and agree upon the validity of the operation. On invalid operations, all nodes should abort with no side-effect. Total Order Broadcast can be significantly compared with Consensus Algorithm.

One example of implementation is: Server A emits message with order 3 and receives another with 5. It must wait for message N4 before it proceeds with committing 5.