What is sharding in MongoDB?

What You'll Learn

• What sharding is
• Shard key selection
• Architecture components

What is Sharding?

Sharding distributes data across multiple machines for horizontal scaling. Each shard contains a subset of the data.

When to Shard

• Data too large for single server
• High write throughput needed
• Working set exceeds RAM
• Geographic data distribution

Sharded Cluster Components

mongos (Router)
         |
    Config Servers
    /     |     \
Shard1  Shard2  Shard3

• mongos: Query router
• Config servers: Cluster metadata
• Shards: Data partitions (replica sets)

Shard Key

// Enable sharding on database
sh.enableSharding("mydb");

// Shard a collection
sh.shardCollection("mydb.users", { region: 1 });

// Hashed shard key (better distribution)
sh.shardCollection("mydb.events", { _id: "hashed" });

Shard Key Selection

Good Shard Keys

• High cardinality
• Even distribution
• Query isolation (queries target single shard)

Bad Shard Keys

• Low cardinality (few unique values)
• Monotonically increasing (timestamp, ObjectId)
• Frequently updated fields

Query Routing

// Targeted query (includes shard key)
db.users.find({ region: "US", name: "John" });
// Routes to specific shard

// Scatter-gather (no shard key)
db.users.find({ name: "John" });
// Queries all shards