Sharding

How sharding works

The conceptual framework:

1. Chain split into multiple shards.
2. Each shard processes transactions in parallel.
3. Shards communicate through cross-shard mechanisms.
4. State synchronization ensures consistency.
5. Result — higher total throughput than single chain.

Like database sharding, blockchain sharding splits work across multiple parallel processes.

Why sharding is hard

Several challenges:

• Cross-shard transactions — transactions affecting multiple shards are complex.
• State management — keeping shard states consistent.
• Validator distribution — assigning validators to shards.
• Security — smaller validator sets per shard could be vulnerable.
• Implementation complexity — sharded systems are much harder to build than single-chain.

These challenges have made sharding one of the most-discussed but slowest-implemented scaling approaches.

Where sharding is used

Various implementations:

• NEAR Protocol — uses sharding (Nightshade).
• Ethereum's previous roadmap — included data sharding for data availability.
• Various other chains — assorted approaches.

Ethereum's roadmap evolved away from execution sharding toward rollup-centric scaling as the primary approach.

Ethereum's evolution

A specific case:

• Original sharding plan — ~64 shards for execution.
• "Surge" phase — focused on data sharding for rollups.
• "Danksharding" — current direction; data availability sharding only.
• Execution kept on single chain + L2 rollups.

The shift reflects engineering pragmatism — rollups proved easier to build than full execution sharding.

Sharding vs. rollups

Different scaling approaches:

• Sharding — split chain into pieces; each handles transactions in parallel.
• Rollups — execute off-chain; settle to main chain.

Both increase capacity; the trade-offs differ:

• Sharding — more architectural complexity.
• Rollups — simpler architecture but cross-rollup interactions are challenging.

What individuals should know

For most users, sharding is largely invisible:

• NEAR users transact normally without considering shards.
• Cross-shard transactions sometimes have specific patterns.
• Sharding is one of multiple scaling approaches.

For broader awareness:

• Sharding has been promised for years but slow to ship.
• Actually-deployed sharding at NEAR works but hasn't dominated.
• Rollup approach has been faster to deliver real impact.

Sharding represents one of the longest-discussed scaling approaches in blockchain. Its actual implementation has been slow; rollups have moved faster. Whether sharding eventually plays a larger role depends on continued architectural evolution.