Ethereum’s Centralization Dilemma: Balancing Performance and Decentralization

On July 27th, Ethereum’s core developer Péter Szilágyi tweeted about the increasing level of centralization of Ethereum. In response, Vitalik Buterin emphasized that the current upgrades, particularly the PeerDAS proposal and Blob expansion, aim to enhance user experience rather than exclude individual stakers. He assured that the Ethereum community would continue to prioritize decentralization measures.

However, Vitalik’s response appears rather general, lacking a more technical perspective on the centralization tendencies that come with a series of upgrades. Looking at the development of Bitcoin and Solana, hardware upgrades inevitably lead to specialized, centralized operations, such as Solana’s specialized node configurations and the concentration of Bitcoin mining pools.

Yet, considering Vitalik’s personal preferences and Ethereum’s strategies against external competitors like Solana, Ethereum’s trend toward centralization seems inevitable.

Combining Software and Hardware: PeerDAS Upgrades and Blob Expansion

Since Ethereum transitioned from PoW to PoS, centralized staking systems represented by Lido and Binance have captured over 50% of the market share, with Lido alone maintaining around 30%. From a centralization perspective, Ethereum has gone further than Bitcoin.

With the upcoming PeerDAS upgrade and Blob expansion, professional stakers are expected to edge out individual players, further pushing Ethereum towards de facto centralization.

The PeerDAS (Peer Data Access System) proposal is a key component of EIP-7594. It aims to extend data availability beyond what EIP-4844 covers by leveraging well-known, battle-tested P2P components. The proposal’s primary change is increasing the size of Ethereum Blobs to 32 MB, significantly enhancing the mainnet’s ability to support Rollups and their data availability needs. Typically, Blobs are used to improve the data availability of the Ethereum network, a central element of EIP-4844. Blob transactions, introduced to Ethereum, provide an efficient, low-cost data storage method for Layer 2 solutions like Rollups.

Specifically, Blobs are binary large objects and the smallest storage units for Ethereum’s data availability. Unlike traditional call data, Blob transactions are invisible to Ethereum contracts and can only be stored for 18 days. By shifting data from L1 to L2, Blob transactions reduce L1’s fee, offering a more economical format for large data transactions on-chain, meeting the cheap data availability needs of Rollups.

Blob transactions also feature specific parameters and cryptographic techniques, with an improved gas accounting method tailored for their transactions. Each Blob is set at 128KB, and each block can contain up to six Blobs, with potential expansion to 256 Blobs per block in the future.

The EIP-7594 proposal includes the following components besides PeerDAS:

• Node Data Handling and Transmission: PeerDAS changes how nodes process and transmit data through Ethereum’s P2P network, significantly improving data transmission efficiency.
• Blob Gas Limit Increase: To benefit users, especially Layer-2 Rollups, developers must raise the current limit of six Blobs per block to a higher level, allowing for greater Blob (arbitrary data) throughput and improved overall performance.
• Decentralized Storage of Historical Records: In the context of EIP 4444, PeerDAS involves decentralized methods for storing historical records, a critical part of Ethereum’s development.
• Improved Permissionlessness and Decentralization: Vitalik Buterin noted that PeerDAS is a crucial improvement in Ethereum’s short and mid-term plans for enhancing permissionlessness and decentralization, indicating its importance in increasing network transparency and accessibility.

However, this upgrade has sparked controversy and concerns. As mentioned at the beginning, Ethereum core developer Péter Szilágyi criticized the PeerDAS proposal. He stated that it might disrupt decentralization, particularly for individual stakers with simpler computational setups. This could significantly increase their difficulty in participating in the Ethereum network. As a result, it could lead to more centralized validation nodes. Szilágyi also pointed out that the PeerDAS upgrade could undermine the interests of individual stakers.

From Vitalik’s perspective, the PeerDAS upgrade will focus on the scalability of Rollups, further enhancing the efficiency of Layer 2 networks like Base and Arbitrum.

Signs of Inevitable Ethereum Centralization

In 2023, Vitalik acknowledged that Ethereum’s node centralization problem might take 20 years to resolve. A more stark reality is that many so-called decentralized Ethereum nodes are hosted on Web2 cloud servers. AWS alone holds over 33% of the share, making Lido and AWS the de facto backers of Ethereum. The concept of decentralization is built on centralized entities, especially since both Lido and AWS must comply with U.S. laws, leading to Ethereum’s Americanization.

Vitalik has proposed numerous solutions, but generally, there is more talk than action. Considering his personal investment in MegaETH, it is hard to believe he opposes the trend towards centralization.

Notably, unlike other Rollup solutions, MegaETH is distinctive in fully adopting Solana’s operational model, stacking unlimited hardware and network resources to achieve theoretically optimal results. However, this approach will likely reduce network nodes to single digits or even just one. As the MegaETH team stated, “MegaETH aims to push ETH L2’s performance metrics to the limits of its running hardware, bridging the gap between blockchain and traditional cloud computing servers.”

Theoretically, MegaETH’s TPS could reach 100,000 transactions per second, accepting various centralized processing methods, which Vitalik has already endorsed. More similar solutions may emerge.

We might have caught a glimpse of Vitalik’s intentions as early as the 2016 DAO incident. After the DAO was hacked, Ethereum chose rollback and hard fork, leading to the major split between ETH and ETC. Later development shows that ETH’s fork did not hinder its dominance in the public chain field, while ETC, adhering to the crypto spirit, remains relatively obscure.

Conclusion

Decentralization is dubbed as blockchain’s spirit. However, beyond Bitcoin, we see many compromises suggesting that trading and compliance might be the inevitable paths for the industry. No one genuinely cares about the ideology, only its monetization.

If Ethereum continues to head towards centralization, will we see new game-changers emerge?