Distributed Systems for System Designs
Distributed systems theory explains what goes wrong when data lives on many machines. Topics include the CAP theorem, consistency models from eventual to strong, consensus algorithms such as Raft and Paxos for agreeing on a leader or value, and ways to handle transactions across services, such as two-phase commit and sagas.
Stage: Reliability · Part of the System Design roadmap · Last reviewed 19 November 2026
What to learn in Distributed Systems
- State the CAP theorem correctly and apply it to a real database
- Compare eventual, causal and strong consistency
- Explain how Raft elects a leader and replicates a log
- Describe a saga with compensating actions
- Explain why two-phase commit blocks on coordinator failure
- Use quorum reads and writes to tune consistency
Distributed Systems interview questions
What does the CAP theorem actually say?
When a network partition happens, a distributed data store must choose between consistency and availability. It cannot return both the latest data and a response from every node. Without a partition, you can have both.
How does the saga pattern replace a distributed transaction?
A saga splits a business operation into local transactions across services. If a later step fails, previously completed steps are undone by compensating transactions, such as refunding a payment, giving eventual consistency without a global lock.
Where Distributed Systems fits in the System Design roadmap
Distributed Systems belongs to the Reliability stage (advanced level), which takes about 32 hours. Design for failure.
Other units in this stage
Back to the full System Design roadmap