Definition

Sharding

A scaling technique where a blockchain is split into multiple parallel "shards," each processing a fraction of total transactions.

Traditional blockchains require every node to process every transaction. Sharding partitions the network into shards (e.g., 64 on Ethereum's original sharding plan), with each shard handling a subset of transactions. Validators only need to process their assigned shard, dramatically increasing total throughput.

Cross-shard transactions add complexity: if I send tokens from shard A to shard B, both shards need to coordinate. Most sharding designs include cross-shard messaging primitives, but this creates latency and additional attack surface compared to single-shard execution.

Ethereum originally planned execution sharding (multiple parallel virtual machines) but pivoted to a "rollup-centric roadmap" — instead of sharding at the L1 level, scale via L2 rollups. The remaining Ethereum sharding work is "data sharding" (introduced via blobs in EIP-4844) which lowers L2 costs without splitting execution.

Why it matters

Sharding shaped Ethereum's early scaling discourse. Knowing why Ethereum moved away from full execution sharding (toward rollups + data sharding) helps interpret its current architecture.

How CryptoRadar24 tracks it

CryptoRadar24 references sharding indirectly when discussing Ethereum scaling, EIP-4844, and L2 ecosystem dynamics.

Related terms

FAQ

Why did Ethereum drop execution sharding?

Rollups proved more practical: same scaling benefits, less L1 complexity, faster shipping. Ethereum kept data sharding (blobs) which makes rollups cheaper without requiring full execution sharding.

Do other chains use sharding?

NEAR uses Nightshade sharding. Polkadot has parachains (a related but distinct concept). Most major chains have moved toward L2 scaling instead of native sharding.

Is sharding a security risk?

Smaller validator subsets per shard mean less work needed to corrupt one shard. Designs use validator shuffling and global random sampling to mitigate, but this is still a meaningful design challenge.

Are blobs the same as shards?

Not exactly. Blobs are a form of data availability sharding — they let rollups post compressed data cheaply without requiring full execution sharding. Different scope, similar throughput motivation.