Range-based sharding assigns contiguous key ranges to each shard:
Shard 1: user_id 1-1000000
Shard 2: user_id 1000001-2000000
Shard 3: user_id 2000001-3000000
Advantages:
- Efficient range queries (single shard)
- Easy to understand and debug
- Natural for time-series data
Disadvantages:
- Hot spots if keys aren't uniformly distributed
- New users all hit the latest shard
Use range sharding for time-series data or when range queries are common. Combine with load monitoring to detect hot spots.