What is sharding and how do you choose a shard key?

Each shard holds a subset of the data and handles its own reads and writes. A router or the client maps a key to a shard. The design question is which key distributes load evenly without forcing most queries to visit every shard.

Choosing the key

• High cardinality, so there are many possible values and no single dominant key
• Even write distribution, so one shard does not absorb a disproportionate share
• Query alignment, so common reads can be served from one shard
• Stable values, because changing a shard key usually means moving the row

function shardFor(key, shardCount) {
  let hash = 2166136261;
  for (let i = 0; i < key.length; i += 1) {
    hash ^= key.charCodeAt(i);
    hash = Math.imul(hash, 16777619);
  }
  return Math.abs(hash) % shardCount;
}

const shard = shardFor("customer:7", 16);

Resharding is the hard part. Consistent hashing or virtual buckets reduce the amount of data that moves when nodes are added, but some migration is unavoidable. Plan for a routing layer that can move a key without changing application code.