What Is Sharding?
Also known as: Shard, Shards, Danksharding, Database Sharding
In short
Sharding is a scaling method that splits a blockchain's workload into parallel pieces called shards, so each node processes only part of the transactions rather than all of them, with the aim of increasing network capacity.
In a classic blockchain, every full node verifies all transactions and stores the entire state; this improves security but limits capacity. In sharding, the network is split into multiple pieces that each manage their own transactions and state, and validators are distributed across these pieces. A coordination layer provides communication between the pieces and shared security. Networks such as Near Protocol and MultiversX use different sharding designs.
Ethereum's roadmap also originally planned to split transactions across 64 separate shard chains. With the rise of rollups, this plan changed: Ethereum now focuses on scaling not transactions but the data published by rollups. This approach is called danksharding; its first step, proto-danksharding (EIP-4844), introduced blob transactions with the Dencun upgrade in March 2024.
The main challenge of sharding is security. When validators are divided among shards, taking over a single shard can be easier than taking over the whole network; for this reason, validators must be reassigned randomly and frequently between shards. The complexity of cross-shard transactions and guaranteeing data availability are also technical problems that must be solved.
Last updated: · This content is for information purposes only and is not investment advice.