Design a distributed key-value store
Company: TikTok
Role: Software Engineer
Category: System Design
Difficulty: medium
Interview Round: Technical Screen
## Scenario
Design a **distributed key-value (KV) store**.
## Requirements
### Functional
- API:
- `Get(key) -> value | not_found`
- `Put(key, value)` (optionally with versioning)
- (Optional) `Delete(key)`
- Support many machines; data must be partitioned.
### Non-functional
- High availability and fault tolerance (node failures).
- Low latency reads/writes.
- Define consistency target (e.g., strong vs eventual) and justify.
- Handle rebalancing when nodes are added/removed.
- Discuss how you’d do replication, leader election, and failure recovery.
Assume values are small-to-medium (e.g., <= 1 MB) and keys are strings.
Quick Answer: This question evaluates understanding of distributed systems concepts such as data partitioning, replication, consistency models, fault tolerance, scalability, rebalancing, leader election, and performance trade-offs in a key-value storage context.