🧩 Sharding Sharding
Splitting one big blockchain into smaller parallel pieces called shards. Each shard handles its own slice of transactions, and a node only checks the shard it's assigned to — so the work is shared instead of every node doing everything.
🛒 The simple version — open more checkout lanes
Picture a supermarket with one cashier and a line out the door. Everyone waits, because every shopper goes through the same register. A blockchain hits the same wall: in a plain design, every node re-processes every transaction, so the whole network only moves as fast as one machine can keep up. Sharding opens more lanes. The network is divided into shards, each handling its own set of shoppers (transactions), and the lanes serve people at the same time. The line moves faster because the work is split, not because any one cashier got quicker.
🧱 What a shard actually is
Sharding is a database trick borrowed from regular software, where a big table is partitioned into smaller chunks. Applied to a blockchain, each shard holds its own subset of transactions — and in some designs its own accounts, balances, and smart contracts. A node is assigned to one shard and only validates that shard, instead of carrying the entire chain. That means less to store and less to process per machine. When a transaction needs to reach across shards, they talk to each other through cross-shard communication — and that hand-off is where most of the difficulty lives.
📈 Why beginners run into this word
You meet "sharding" when reading about speed. Base-layer (Layer 1) chains are slow by everyday standards: rough beginner figures put Bitcoin near 7 transactions per second and Ethereum around 15 to 30, far below a card payment network. Speeding a chain up without giving away decentralization or security is the hard part, known as the blockchain trilemma. Sharding is one proposed answer to the scalability corner of that triangle.
🛰️ A real example, and the plan that changed
| Project | What it does |
|---|---|
| 🛰️ NEAR Protocol | Uses sharding called Nightshade. It shows users a single chain while each shard processes transactions in parallel; rolled out in phases, with shard count and throughput growing over time. |
| ⟠ Ethereum | Was the famous sharding example, then changed course. It dropped shard chains for a rollup-centric path and now shards the data rollups use, via the blob upgrade that went live in March 2024. |
📊 Shard counts and headline TPS numbers vary a lot by source and by roadmap phase. Treat any single figure as illustrative, not a fixed spec.
🚨 The trade-offs honestly
- 🛡️ Smaller shards, thinner defense — A shard has fewer validators than the whole chain, so attacking one shard can be cheaper than attacking the full network
- 🔀 Cross-shard is hard — Moving value or state between shards is complex to design and a frequent source of bugs
- 🧭 "Sharding" isn't one thing — Classic chain sharding and Ethereum's data sharding (blobs for rollups) solve the speed problem in different ways
❓ FAQ
- Does sharding mean a coin gets split into many separate coins?
- No. The shards are pieces of one network, not separate coins. They share the same token and the same security model, and they pass data to each other when a transaction touches more than one shard. To a user it still looks like a single chain.
- Is Ethereum being split into 64 shard chains?
- Not anymore. That older plan was dropped. Ethereum now scales through Layer 2 rollups and gives them cheap data space instead. Proto-Danksharding (the blob upgrade that went live in March 2024) shards the data for rollups, not the chain itself.
- Does sharding make a blockchain less safe?
- It can add risk. A single shard has fewer validators than the whole chain, so it can be cheaper to attack one shard than the entire network. Moving value between shards is also harder to build and a common source of bugs. Designs add safeguards, but it is a real trade-off.