Ethereum Dencun

An Ethereum Dencun upgrade was deployed on 13 March 2024, introducing blobs — a new temporary data type that stores information for 18 days before being deleted. This dramatically reduced costs for layer-2 solutions, which use blobs instead of calldata to post transactions, reducing fees by 10–100 times.

The problem Dencun solves

Layer-2 solutions like Arbitrum and Optimism bundle transactions and post them to Ethereum for security. However, they posted transaction data as calldata — a permanent part of the ledger.

Storing data on-chain is expensive. Arbitrum and Optimism transactions could cost a few cents, far cheaper than Ethereum base layer (~$10–$100+) but not as cheap as they could theoretically be.

The bottleneck was data availability: layer-2 solutions needed to post enough data to Ethereum so that anyone could reconstruct the layer-2 state independently (a security property). But permanent storage on Ethereum is expensive.

Blobs and proto-danksharding

Dencun introduced blobs — temporary data storage. A blob is a commitment (a cryptographic hash) that proves data existed, without storing the data permanently on-chain.

Blobs are inspired by a proposal called “Danksharding” (a longer-term Ethereum scaling road map). Dencun implements a precursor: proto-danksharding (EIP-4844).

Here is how it works:

1. A layer-2 creates a blob (up to ~128 KB) containing transaction data.
2. The blob commitment is posted to Ethereum (small, constant size).
3. Nodes must temporarily store the blob (for ~18 days) to verify layer-2 validity.
4. After 18 days, the blob data can be deleted; only the commitment remains.

This gives layer-2 solutions all the security properties (anyone can reconstruct state from commitments and archived blobs) without permanent storage overhead.

Cost reduction

The impact was dramatic:

• Before Dencun: Arbitrum and Optimism transactions cost $0.05–$0.50 (depending on congestion).
• After Dencun: Transactions dropped to $0.005–$0.05, a 10–100x reduction.

This made layer-2 solutions competitive with alternative chains like Solana and Polygon in terms of cost, while maintaining Ethereum’s security.

Blob pricing

Blobs have their own gas market, separate from calldata. This allows layer-2 solutions to post cheaper data without competing with base-layer Ethereum users for block space.

Blob gas prices can spike if layer-2 solutions flood the network but generally remain low compared to calldata.

Data availability and archived data

Critics raised a concern: if blob data is deleted after 18 days, how do new layer-2 nodes sync? The answer: layer-2 services (like Arbitrum) run archive nodes that maintain blob data longer than 18 days. Additionally, some projects archive blobs off-chain using services like Arweave or IPFS.

This reintroduces a small trust assumption (that archive nodes persist data), but it is far weaker than relying on a centralised service.

Impact on layer-2 adoption

Dencun’s cost reduction significantly improved layer-2 user experience. Previously, a swap on Arbitrum might cost $0.10–$1.00; after Dencun, it costs $0.01–$0.10. This dramatic reduction likely accelerates layer-2 adoption.

Future work: full Danksharding

Dencun is proto-danksharding. Full Danksharding (a future Ethereum upgrade) would further optimise data availability by using techniques like data availability sampling and 2D sharding, potentially enabling thousands of transactions per second.

Full Danksharding is conceptually designed for, but not yet deployed. Dencun is a pragmatic intermediate step that achieves much of the benefit (low layer-2 costs) with simpler implementation.

Comparison with competing scaling solutions

Layer-2 solutions now compete with alternative chains on cost:

Layer-2 solutions maintain Ethereum’s security but now match competing chains’ costs.

See also