ERC-4337: Elephant in the Room — Part 3
The Key Challenges and Unknowns for Account Abstraction
ERC-4337 introduced a new transaction layer on Ethereum to achieve account abstraction. In principle, it represents one of the largest value capture opportunities for blockchains over the next 3–5 years.
Today, however, there are a number of challenges, unknowns and risks associated with it that need to be solved before we can unlock its true potential.
In part three, we look at the key challenges, dependencies and unknowns that currently exist for account abstraction.
Multichain Compatibility
Thanks to the EVM, users have become accustomed to their wallet address/EOA being the same on every chain. Your address on Ethereum is the same as on Avalanche, Fantom, BSC etc.
Account abstraction wallets however, are, by design, unique smart contract instances deployed on individual chains. While users can create account abstraction wallets on different chains, they ultimately have distinct account addresses, which introduces a few immediate concerns:
Breaking assumptions: Most bridges today assume identical addresses on both sides of cross-chain bridging. With account abstraction, this may result in loss of funds as the address on the receiver side may be different.
Risk of lost funds: If users share their account abstraction wallet address without specifying which blockchain, it would lead to inaccessible funds if the sender sends assets on an unsupported chain.
Added complexity: Managing multiple addresses across chains imposes additional overhead for users.
These are problems for many of today’s blockchain users who would take this seamless multichain compatibility for granted.
The good news is that Safe has begun to tackle these limitations with the implementation of the ERC-3770 standard, which adds prefixes to the account address. The solution is a work in progress; more information can be found here.
Ultimately, the solution to the overarching problem will lie with the wallet provider who will decide whether to use deterministic addresses across the networks.
Signing Issues, ERC-1271 Support
For a dapp to function with ERC-4337, it must adhere to the ERC-1271 standard for signature validation.
While EOAs can sign messages using their associated private keys, smart contracts currently cannot. ERC-1271 is a standard that enables a smart contract to determine whether a given signature is valid on its behalf.
You can find a list of dapps that have integrated ERC-1271 support here.
Gas Fee Responsibilities
Deploying a smart contract requires a gas fee, whereas EOAs are created for free. If account creation is not sponsored, the user will be responsible for the cost of creating an AA wallet.
Due to the nature of smart contract calls, transactions with an AA wallet will also be more expensive than a similar transaction with an EOA.
ERC-4337 offers a potentially more cost-effective alternative to EOAs on rollup networks, owing to its ability to incorporate adaptable authentication logic through smart accounts. This approach enables a substantial reduction in the volume of data stored on mainnet by consolidating signatures with the Aggregator. This efficient compression of data significantly reduces gas costs, potentially by up to 7x.
Risks and Unknowns for the Infrastructure Layer
Although ERC-4337 outlines the high level interfaces for components such as bundlers, paymasters, and signature aggregators, a lot of confusion and unresolved issues sit in the infrastructure layer:
Bundler Network and Mempool Design: Current bundler implementations primarily rely on private mempools, with UserOperation bundlers relayed directly to specific block builders. The feasibility, design and behavior of public or P2P mempools within the bundler network remains an open question.
MEV Extraction and Distribution: Today, it is unclear whether bundlers can extract MEV (and how much) by ordering UserOperations or if block builders can do so by ordering bundles. The distribution of MEV between bundlers and block builders also needs to be determined. This key economic incentive needs to be understood and proven to encourage a healthy, decentralized and competitive bundler market.
UserOperation and Regular Transaction Coordination: The potential for UserOperations to pass validation and fail during execution due to state conflicts with regular transactions is currently unknown. This is a key item blocking large wallets from offering smart contract wallets to their users.
The implementation of account abstraction will likely increase the number of users and transactions processed on the network, which will require more bundlers, infrastructure improvements, and technical know-how.
Eden is committed to contributing to the continued development of Ethereum’s infrastructure with its expertise and technical know-how. Click here to find Eden Network’s bundler and how to use it.
Security: The Devil is in the Detail
Account abstraction alters and improves Ethereum’s existing security model, which traditionally has depended on private keys for transaction signatures.
The updated model unlocks smart contracts to manage transaction signing as the implementation of social wallet recovery, UX and robust general smart contract security is paramount to the adoption of account abstraction.
In theory, account abstraction makes user wallets and funds more secure, but if these features are implemented poorly, they will leave users open to multiple attack vectors such as session key and social engineering hacks. When developing an account abstraction wallet, and web3 UX in general, these attack vectors should be key considerations.
The Future of ERC-4337
While account abstraction is still a novel concept in the Ethereum ecosystem, the potential it has to transform the network is evident and obvious. Like all nascent tech, it will take time to develop, tinker and test use cases before we stumble upon the “killer app”. As an example, Ethereum went live in 2015, but it wasn’t until 2020 that DeFi and NFTs really took off.
Many questions still remain, but we at Eden Network are excited at the possibilities of account abstraction. We will continue to invest in research and development, and support other projects and businesses building in the AA space.
Are you excited too? Tell us your thoughts on account abstraction by tweeting at us, or in Discord.
