How do you implement data distribution strategy when sharding?

Data sharding strategy determines how data is partitioned and distributed across different shards (parts of a database). Main approaches:

1. Horizontal sharding by key (hash-based sharding) — a hash of the key value (e.g., userId) is computed, and the shard is selected based on the result.

2. Range-based sharding — data is distributed across key ranges, for example, users with IDs from 1 to 1000 in one shard, from 1001 to 2000 in another.

3. Geographical or business logic-based sharding — data is distributed based on regions or other criteria.

Implementation usually includes:

• Defining the sharding key.
• Routing logic for requests to the appropriate shard server.
• Ensuring consistency and load balancing.

Example in Java (simplified):

int shardCount = 4;
int shardId = Math.abs(userId.hashCode()) % shardCount;
// Then the request is directed to the database with index shardId

The choice of strategy depends on the nature of the data and performance requirements.