Search⌘ K
AI Features
Home
Courses
System Design Deep Dive: Real-World Distributed Systems

SMR in Practice Via a Log

Explore the implementation of state machine replication through log data structures to achieve fault tolerance in distributed systems. Understand how client commands are ordered and committed using consensus algorithms such as Paxos and Raft. This lesson covers log states, client interaction handling, command sequencing, and methods like log compaction to maintain consistency and efficiency across replicas.

We'll cover the following...

By now, we have learned what state machine replication is and how it is implemented to toleratett faults. This lesson foreshadows the upcoming chapters on Paxos State Machine and Raft State Machine replications, both practical and widely used consensus algorithms. We will see how Paxos and Raft take the SMR ideas and implement them.

Log data structure

Every state machine of the replica group maintains a data structure called a log. We can consider a log as a file that consists of slotA slot can be considered a placeholder for client command or SMR's own internal commands. Each slot can store one command. numbers, starting from the left and extending indefinitely on the right. Often, these slot numbers are sequentially numbered starting from ...