Design a Local Leader Election System
Company: Discord
Role: Software Engineer
Category: System Design
Difficulty: medium
Interview Round: Technical Screen
Design and implement a local leader-follower distributed system.
You have multiple nodes running on one machine, each as a separate process or service instance. At any time, every live node should know which node is the current leader. The system must support:
- Starting an initial cluster of several nodes.
- Adding a new node to the existing network.
- Detecting when the current leader fails.
- Automatically electing a new leader after failure.
- Propagating the new leader information to all live nodes.
The interview is practical rather than purely theoretical: be prepared to explain the design and then implement the core logic locally. You may use local networking, in-memory state, files, or a local dependency such as Redis if you can justify and implement it.
Quick Answer: This question evaluates a candidate's understanding of distributed systems concepts such as leader election, failure detection, membership management, and inter-process communication in a local multi-process environment.