Sharding is a method of splitting blockchains ( or other types of databases) into smaller, partitioned blockchains that manage specific data segments. This setup takes the stress off a single chain dealing with all transactions and interactions on a network. Each partitioned blockchain is known as a shard and has its specific ledger.
These shards can then process their own transactions, but a beacon blockchain or main chain manages interactions between shards. This makes sharding a Layer 1 network scalability upgrade, as it’s a change to the mainnet of a blockchain. Implementing shards offers much more scalability than continuing to use a single network. Ethereum’s developers are introducing sharding during Ethereum’s upgrade process to a scalable Proof of Stake network.
In Ethereum’s example, a Beacon Chain will coordinate 64 individual shards. The original Proof of Work Ethereum chain will merge into the Proof of Stake Beacon chain, and sharding will arrive later.
One of the benefits of sharding a network includes the ease and accessibility of running a node. As the network’s data is split across shards, there is no longer the need for validator nodes to store the whole history of the entire blockchain. Instead, the validator only needs to keep confirmations of the data’s integrity.
Networks without sharding often resort to the use of rollups to improve scalability. These take transactions off-chain and package them together for validation on the main chain. So, while sharding may seem like a competitor to rollups, it actually makes them more efficient. Rollups in a sharded network will be able to report their state more efficiently, improving their speed.
The primary worry with sharding is the possibility of bad actors taking over a shard. This shard could then negatively impact other parts of the network. Without proper care and rules in place, the task of taking over a shard is more achievable than taking over an entire non-sharded network.