Ethereum’s sharding is a critical scaling solution designed to significantly enhance the network’s transaction processing capacity and reduce congestion. This evolution moves the blockchain beyond its historical limitation of processing around 15-20 transactions per second (TPS), addressing key performance bottlenecks.
The concept, while rooted in database management, has undergone a substantial transformation within the Ethereum ecosystem, shifting from initial plans for execution sharding to the current focus on Danksharding.
Understanding sharding for blockchain networks
Sharding is a database architecture technique that improves scalability and performance by partitioning data into smaller, independent pieces, known as shards. Each shard processes its own transactions and queries in parallel, effectively distributing the workload and reducing potential bottlenecks.
In blockchain technology, sharding means dividing the network into smaller partitions, with each capable of processing transactions independently. This parallel processing allows the network to handle a much larger volume of activity simultaneously, thereby enhancing throughput, reducing congestion, and potentially lowering transaction fees.
The core idea moves from a system where every node processes every transaction to one where the workload is efficiently split into parallel segments. This represents a fundamental shift for distributed ledgers, providing a clear path towards unprecedented network scale and efficiency for the Ethereum blockchain.
Ethereum’s evolving sharding roadmap
The original vision for Ethereum’s sharding included “execution sharding,” which would have split the network into 64 distinct chains. These chains were intended to process transactions and smart contracts in parallel, creating multiple independent execution environments.
However, the Ethereum developer community strategically shifted towards a “rollup-centric roadmap.” This updated strategy now prioritizes Danksharding, which uses sharding to expand data availability for Layer 2 (L2) rollups rather than partitioning the execution layer itself. This approach aims to dramatically reduce the cost of storing rollup data on the Ethereum mainnet.
Proto-Danksharding: The first step
Proto-Danksharding, also known as EIP-4844, represents the initial phase of this scaling evolution. It was implemented as part of Ethereum’s Dencun upgrade in March 2024. This stage introduces a new transaction type capable of carrying “blobs” of data.
These blobs are large, fixed-size chunks of data attached to standard Ethereum transactions. They provide a significantly cheaper method for L2 rollups to post their batched transaction data to Ethereum. Before this, rollups relied on calldata, which was more expensive due to competition with regular Ethereum transactions for execution gas.
Blobs now operate with their own separate fee market, structured similarly to EIP-1559, which directly contributes to lowering data availability costs for L2s. Unlike regular transaction data, the contents of a blob are inaccessible to the Ethereum Virtual Machine (EVM) and are automatically purged after a fixed period, typically around 18 days (4096 epochs).
This design minimizes long-term data storage demands on the mainnet while ensuring data remains available long enough for rollups to function correctly and for users to verify their transactions. Under Proto-Danksharding, a single block can accommodate a target of three blobs, with a maximum capacity of six. Each blob is approximately 128 KB.
The vision of full Danksharding
Full Danksharding represents the complete realization of Ethereum’s rollup scaling ambitions. It plans to significantly increase the number of blobs attached to blocks, expanding from the current maximum of six in Proto-Danksharding to 64.
A crucial component of full Danksharding is Data Availability Sampling (DAS). DAS enables nodes to verify data availability without needing to download an entire block or all blob data. Instead, each node downloads a small, randomly selected subset of the total data.
This mechanism leverages data erasure coding, which adds redundant information to the dataset, allowing for reconstruction even if parts are missing. Successfully downloading these samples provides high confidence that all data is available, scaling the data availability layer without burdening individual nodes with excessive hardware requirements.
The Beacon Chain, which launched in December 2020, coordinates the network and manages the consensus mechanism. It plays an essential role in overseeing the shards, managing validator registration, distribution, and coordination to ensure overall network security and consensus under the Proof of Stake consensus model.
Key benefits of Ethereum’s scalability solution
The primary advantage of Ethereum’s sharding implementation is a substantial increase in transaction throughput, with the long-term goal of achieving over 100,000 transactions per second (TPS). The current roadmap envisions scaling Ethereum’s processing capacity to roughly 1 gigabyte of gas per second, which could translate to around 10,000 TPS on Layer 1.
By distributing the network workload more effectively, sharding will significantly reduce network slowdowns and backlogs. This improvement leads to faster transaction speeds and enhances the overall reliability of the Ethereum network, creating a more responsive user experience.
As network congestion decreases, transaction fees are projected to fall, making the Ethereum network more accessible and affordable for a wider range of users and applications. This cost reduction is vital for fostering broader adoption and supporting decentralized application growth.
Furthermore, sharding is expected to lower the hardware requirements for running an Ethereum client, making participation more accessible. Individuals could potentially run clients on consumer devices like personal computers or even mobile phones, thereby increasing network participation and promoting greater decentralization.
Danksharding specifically offers a more efficient and cost-effective method for Layer 2 rollups to post their transaction data to the Ethereum mainnet. This is absolutely critical for the continued scalability and cost-efficiency of the entire L2 ecosystem.
Debunking common sharding misconceptions
A frequent misunderstanding about Ethereum sharding is that it splits the execution layer into many independent “shard chains” that process transactions separately. While this was part of the initial vision, it’s no longer the case.
The current roadmap, embodied by Danksharding, does not fragment the base layer into multiple execution shards. Instead, its focus is on expanding data availability for Layer 2 rollups, enabling them to scale more efficiently on top of the mainnet.
Another misconception suggests that “every node processes every transaction.” While true for traditional monolithic blockchains, sharding is designed to distribute this workload. In Danksharding, the network shifts away from requiring every computer to process every transaction.
Instead, the workload is split into parallel segments, and with Data Availability Sampling (DAS), nodes will eventually check data availability by sampling small, random pieces rather than downloading every blob. This significantly optimizes resource usage across the network.
The future of Ethereum’s network performance
Ethereum’s evolution through sharding, particularly with the implementation of Proto-Danksharding and the future full Danksharding, marks a pivotal moment for the network. It’s a strategic move to overcome long-standing scalability challenges while preserving decentralization and security.
This phased approach underscores the Ethereum Foundation’s commitment to a robust, high-throughput blockchain capable of supporting a vast and increasingly complex decentralized ecosystem. The ultimate goal remains to create a highly scalable and accessible platform for global innovation.