Intro
A strong form of Reliable Broadcast. Guarantees that all correct nodes deliver the same messages in the same exact order. If any node sees [m1, m2, m3], all nodes see [m1, m2, m3].
Implementation
STUB
Replicated Logs
A replicated log, used in Distributed Consensus Algorithms like Raft, is a sequence of commands agreed upon by all replicas.
Key properties
- Every replica agrees on both the command itself, and the
- position that the command is in the log. Basically, the goal here is to have a log that has the same exact messages in the same exact order for all nodes in a cluster.
This is definitively Total Order Broadcasting.
Conclusion
Achieving consensus on a replicated log, which is the backbone of distributed consensus, is the same as implementing a Total Order Broadcast, making them super important.