Blob Space

Blob space is a specialized data storage compartment within an Ethereum block, designed to hold large, temporary data packets called blobs. Unlike standard transaction calldata, which is stored permanently on-chain, blob data is only accessible for a short period (approximately 18 days) before being pruned by nodes. This ephemeral storage model allows Layer 2 rollups to post their transaction data to Ethereum at a significantly lower cost, as it avoids the long-term storage burden and associated gas fees of the Ethereum Virtual Machine (EVM).

The primary function of blob space is to provide data availability. Rollups execute transactions off-chain and then post cryptographic proofs along with the underlying transaction data to Ethereum. By posting this data to blob space, rollups guarantee that anyone can independently verify and reconstruct the rollup's state, ensuring security and trustlessness. This mechanism separates the cost of data availability from the cost of EVM computation, creating a more efficient scaling pathway known as proto-danksharding.

Blob space is priced via a distinct fee market, separate from the standard gas market for Ethereum execution. This uses a base fee and priority fee system specifically for blobs, which adjusts dynamically based on demand for blob space in each block. The separation prevents competition for block space between rollup data posting and regular user transactions, leading to more predictable and stable costs for Layer 2 operations.

From a node's perspective, blob data is handled differently. Full nodes and consensus clients must store and propagate blobs for the short retention window, but execution clients do not execute this data. After the retention period, nodes can delete the blob data, significantly reducing long-term storage requirements compared to permanent calldata. This design is a foundational step toward the full danksharding vision, where the data availability layer will be distributed across a committee of validators.

Blob Space is a dedicated data availability layer on the Ethereum blockchain, introduced via EIP-4844 (Proto-Danksharding), designed to provide low-cost, temporary data storage for Layer 2 rollups. It functions as a separate resource market from standard block gas, where validators accept and attest to blob-carrying transactions that include large data packets called blobs. These blobs, each ~128 KB, are stored on the consensus layer for approximately 18 days (4096 epochs) before being pruned, providing a sufficient window for fraud or validity proofs to be verified. The primary goal is to decouple rollup data costs from mainnet execution gas fees, dramatically reducing transaction costs for end-users.

The mechanism operates through a distinct fee market, governed by a blob gas limit per block and a blob base fee that adjusts dynamically based on demand, similar to EIP-1559. Rollups post their compressed transaction data in blobs, and validators are responsible for making this data available. Clients and Layer 2 nodes can then download the blob data from the peer-to-peer (p2p) network using the blob identifiers (versioned hashes) included in the block. This separation ensures that high-volume data posting by rollups does not congest the execution layer where Ethereum Virtual Machine (EVM) transactions are processed.

From an economic perspective, Blob Space creates a more efficient and predictable pricing model for rollup data. The blob base fee is burned, creating a deflationary pressure on ETH, while validators receive priority fees from blob transactions. The temporary storage model—lasting only for the data availability sampling window—fundamentally reduces the perpetual storage burden on all consensus nodes compared to storing data directly in calldata. This architecture is the foundational step toward full Danksharding, a future upgrade that will scale Blob Space to handle 64 blobs per block, further decentralizing data availability.

Proto-Danksharding, implemented as EIP-4844, introduced blob-carrying transactions and blob space as a dedicated data layer for Layer 2 rollups. This interim upgrade, also known as The Surge, significantly reduced rollup data costs by providing a new, cheaper transaction type for posting data commitments, without yet implementing data sharding across the network. It established the core cryptographic and consensus framework—using KZG commitments and a separate blob gas market—required for the full vision.

The evolution to full Danksharding will horizontally partition the blob space into multiple data shards, dramatically increasing Ethereum's data availability bandwidth from ~0.75 MB per slot to a target of 16-32 MB. This final stage eliminates the need for proposers to locally store all blob data by employing Data Availability Sampling (DAS), where light nodes and validators sample small, random pieces of data to probabilistically verify its availability, enabling secure scaling.

The key architectural shift in Danksharding, pioneered by Dankrad Feist, is proposer-builder separation (PBS) with crLists. This ensures block builders, not validators, are responsible for constructing blocks with abundant, cheap blob space, preventing centralized validators from censoring transactions. The crList mechanism allows validators to force the inclusion of certain transactions, preserving censorship resistance in this new model.

This phased approach allows core components—the execution layer's blob handling, the consensus layer's validation rules, and the peer-to-peer networking for blob propagation—to be developed, tested, and deployed incrementally. Proto-Danksharding delivered immediate cost relief for rollups and real-world validation of the blob gas market, de-risking the path toward the more complex, fully sharded system that will ultimately enable mass scalability.