The evolution of Ethereum’s sharding architecture reflects a profound shift in blockchain scalability thinking — from complex, execution-heavy Layer 1 designs to a lean, data-centric approach that empowers Layer 2 solutions. This journey, marked by technical pivots and strategic simplifications, has culminated in what many now call danksharding. Let’s explore how Ethereum’s sharding vision transformed over time and where it’s headed.
The Original Vision: Execution Shards and Crosslinks
Initially, Ethereum’s sharding roadmap (known as “Phase 1”) was designed to follow “Phase 0” — the launch of the Beacon Chain. The idea was to introduce 1,024 execution shards, each capable of running smart contracts and communicating via crosslinks. These crosslinks would anchor shard data into the main chain, ensuring consistency and security across shards.
However, this ambitious design faced growing complexity:
- Supporting full EVM execution on every shard increased state bloat.
- Cross-shard communication introduced latency and coordination overhead.
- Statelessness, dynamic state access, and ewasm execution environments added layers of uncertainty.
As development progressed, it became clear that pushing full execution onto Layer 1 would delay progress and risk system stability.
👉 Discover how next-gen blockchain scaling is redefining performance and accessibility.
A Strategic Pivot: From Execution to Data Availability
By late 2020, a new direction emerged — one championed by Vitalik Buterin in his seminal post A Rollup-Centric Ethereum Roadmap. The core insight? Let Layer 1 focus solely on data availability, not computation.
This shift simplified the problem space dramatically:
- Instead of running complex logic across shards, Layer 1 would only guarantee that data is published and available for verification.
- Rollups (Optimistic or ZK) would handle execution off-chain, submitting compressed transaction data (“calldata”) to Ethereum.
- Scalability would come from increasing data throughput, not parallelizing execution.
This philosophy laid the foundation for data availability sampling (DAS) — a cryptographic technique allowing light clients to verify data availability with high confidence by randomly sampling small portions of a large dataset.
Birth of the Blob: Proto-Danksharding Emerges
Enter the blob — short for binary large object. Unlike traditional blocks, blobs are temporary data carriers attached to beacon blocks, designed specifically for rollup transaction data. They don’t interact with the EVM and expire after ~18 days, reducing long-term storage pressure.
The term “blob” first appeared informally in early research discussions around 2020. Protolambda used it in the context of “shard data blobs” before it was formally adopted in specifications. By 2021, the concept evolved into a more structured proposal centered around Beacon Chain proposers bundling blob transactions proposed by builders — an early nod to proposer-builder separation (PBS).
In early 2022, Vitalik Buterin outlined two paths toward full danksharding:
- Simple path: Modify existing execution payloads to include blob fields.
- Cleaner path: Decouple consensus and execution layers further, enabling modular upgrades.
The latter won favor for its future-proof design, leading directly to EIP-4844, also known as proto-danksharding.
EIP-4844: The Gateway to Full Danksharding
Launched as part of the Cancun-Deneb upgrade, EIP-4844 introduces blobspace — a dedicated area in each block for rollup data. Key features include:
- Each block can carry up to 6 blobs (expandable to 24 in future phases).
- Blobs use KZG commitments for efficient verification.
- Blob data is not processed by the EVM but is made available for sampling.
This change reduces rollup transaction costs by up to 90%, making Layer 2 solutions far more accessible.
But EIP-4844 is just the beginning. It sets the stage for full danksharding, which will enable:
- Thousands of blobs per slot.
- Distributed data sampling via peer-to-peer networks.
- Enhanced fault tolerance and decentralization through light client participation.
👉 Explore how cutting-edge consensus mechanisms are shaping the future of decentralized networks.
Network Layer Challenges and Innovations
Supporting mass data availability requires robust peer-to-peer networking. Early designs leaned on GossipSub for broadcasting shard data, but its limitations in scalability and DoS resistance became apparent.
Alternative approaches gained traction:
- Discv5 (a UDP-based discovery protocol) offers better privacy and efficiency.
- DHTs (Distributed Hash Tables) show promise for scalable data lookup, though initial implementations lagged behind GossipSub in readiness.
Researchers now favor hybrid models combining gossip for propagation and DHTs for targeted retrieval — especially critical for light nodes performing data sampling.
Additionally, future iterations may integrate ZK-based Sybil resistance in p2p networks, though practical deployment remains a longer-term goal.
Core Keywords
- Ethereum sharding
- Data availability sampling (DAS)
- Blob
- EIP-4844
- Proto-danksharding
- Rollup-centric roadmap
- Beacon Chain
- Danksharding
Frequently Asked Questions
Q: What is the difference between sharding and danksharding?
A: Traditional sharding involves splitting the network into parallel chains handling both data and execution. Danksharding focuses only on scaling data availability, leaving execution to Layer 2 rollups. It's more modular and efficient.
Q: Why are blobs temporary?
A: Blobs expire after ~18 days because their primary purpose is short-term data availability for rollups. Permanent storage isn’t needed, reducing node burden and improving scalability.
Q: How does data availability sampling work?
A: Light clients randomly request small portions of a blob. If they can retrieve these samples successfully, they statistically verify that the full data is available without downloading everything.
Q: Is EIP-4844 full sharding?
A: No. EIP-4844 is proto-danksharding — a stepping stone that introduces blobs and lays the networking foundation. Full danksharding will come later with expanded sampling and higher throughput.
Q: Will danksharding make Ethereum faster for everyday users?
A: Indirectly. By drastically lowering rollup fees, danksharding makes Layer 2 networks cheaper and more scalable — which translates to faster, lower-cost transactions for end users.
Q: When will full danksharding launch?
A: There's no fixed date yet. After EIP-4844, Ethereum will gradually increase blob count per block, improve p2p protocols, and implement distributed sampling — likely over several years.
👉 Learn how Ethereum’s latest upgrades are unlocking unprecedented scalability and security.
Looking Ahead: The Road Beyond 2025
The narrative of Ethereum’s evolution has shifted from "when will sharding arrive?" to "how fast can we scale blobspace?" The focus is now on optimizing:
- Bandwidth: Increasing blob capacity per slot.
- Efficiency: Improving KZG evaluation and peer discovery.
- Decentralization: Ensuring light clients can participate in sampling.
Ultimately, danksharding isn’t about making Ethereum do more — it’s about enabling others (rollups) to do everything on top of Ethereum securely and affordably.
As Protolambda noted, “blobspace” may be the most meme-worthy yet technically sound innovation in Ethereum’s recent history — a testament to the power of simplicity over complexity.