2023 has undoubtedly been the year of account abstraction. Even though the concept is nowhere new, with [early discussions](https://blog.ethereum.org/2015/07/05/on-abstraction) dating back to the beginning days of Ethereum, the industry made its [biggest leap this year](https://sixdegree.xyz/research/Half-Year-Data-Report-of-ERC4337-by-Sixdegree.pdf) in bringing working smart contract wallets to hundreds of thousands of users.

Our team at CyberConnect has been at the forefront of building and growing consumer crypto applications. Here, we’re taking a new perspective on social media, building towards a decentralized future where users exercise complete ownership and control over their social data, while developers can innovate permissionlessly. In almost three years, over 50 teams have built web3 applications utilizing the CyberConnect social network protocol, and have cumulatively registered [1.2M+ users](https://dune.com/cyberconnecthq/cyberconnect-link3-metrics), creating 1.9M+ units of social content. In the past 4 months alone, we’ve deployed over 500,000 ERC-4337-compatible smart accounts on Optimism, Arbitrum, Polygon, Linea, opBNB, and Base. At the time of writing, that accounts for almost [50% of all the smart accounts](https://dune.com/niftytable/account-abstraction) that exist today across all blockchains. CyberAccounts have generated over [1M UserOperations](https://dune.com/cyberconnecthq/cyberaccount-cyberconnects-4337-smart-account) and more than USD 1.2M has been deposited in CyberWallet for gas.

### So, why CyberWallet?

One of the most pivotal insights on our journey in building and growing consumer crypto applications emerged while navigating the challenges in user onboarding. Even though decentralized applications and protocols are fairly accessible to crypto-native users, the intrinsic complexity of blockchains and crypto wallets creates a high barrier for regular users to explore the web3 ecosystem. The most significant of these complexities include:

1. Transaction signing and the gas fee system

2. Cryptographic key management

3. Isolated blockchains

Like most other crypto teams, we want to make web3 more accessible to a wider audience. Instead of waiting for someone else to build a solution, we addressed these challenges head-on and developed [CyberWallet](https://bit.ly/3svmoIZ), an ERC-4337-compatible smart contract wallet.

We bet on [ERC-4337](https://eips.ethereum.org/EIPS/eip-4337) because it aims to make blockchains user-friendly, increase their security, and enhance the overall functionality of wallets and transactions. Of all the enhancements it offers to the existing infrastructure, we believe the most crucial ones are:

**Flexible Authentication Layer**

ERC-4337 introduces a versatile system for managing who can access and control blockchain accounts. This flexibility means users can choose how to manage their private keys, which is essential for account access. They can either opt for:

1. **Self-Custodial Management**: Manage keys independently, using a familiar wallet interface like [MetaMask](https://metamask.io/), where users are solely responsible for their keys' safety.

2. **Shared Key Management**: Share the responsibility of managing keys with trusted parties or across multiple devices using advanced methods like Multi-Party Computation (MPC) or cloud enclaves. This can add extra security layers, as keys are not held in one place but distributed and managed collectively.

This flexibility opens the door to a number of innovative applications, like 2-factor authentication and session keys for blockchain accounts.

**More Efficient Transactions by Decoupling Gas Payments**

ERC-4337 introduces an alternative structure for processing blockchain transactions that simplifies the user experience. It does this by using a new kind of transaction, a UserOperation, which separates the transaction from the complexities of paying gas fees. This enables a 'gas-sponsored' experience where users no longer have to pay transaction fees themselves. Instead, a marketplace of gas sponsors—like blockchain companies—could emerge, where these sponsors voluntarily cover the transaction costs for users. This could play a significant role in the early stages of web3 applications by reducing barriers for users.

**Programmable Modular Account**

Smart accounts pave the way for a new generation of programmable accounts, greatly expanding the design space for developers to build novel account solutions in web3. For example, by utilizing standards like [ERC-6900](https://eips.ethereum.org/EIPS/eip-6900), we can offer users enhanced account functionality with modular and extensible features. This includes but is not limited to, session keys, multi-signature authentication, and on-chain guardians. The result is a ‘smart’ web3 account that can adapt to the diverse and evolving requirements of users, providing a seamless and secure interface for engaging with decentralized applications.

Unfortunately, despite these advantages, ERC-4337’s full potential has yet to be realized. While many in the industry have been touting its potential in onboarding the next billion users onto crypto, there’s a reason it hasn’t taken off at the speed many of us expected.

As we set forth on this journey to bring a robust account abstracted smart contract wallet to market, it quickly became clear to us that pioneering work often meets unanticipated obstacles. In the spirit of camaraderie and collective progress, we want to share the challenges we faced with the hope that we can move together quickly to address them as an industry.

### The Challenges

**Single Domain Stateful Architecture**

One of the key differences between an EOA and a smart contract account is that the EOA is stateless, while the smart contract account is stateful. States, like the initial access key setup, are bound to a specific smart contract account on a blockchain. This means that they are stored on each blockchain and [do not magically propagate to others](https://vitalik.ca/general/2023/06/20/deeperdive.html). Therefore, managing a smart contract account in a multi-chain setup is not as easy as switching a network inside an EOA wallet.

To get around this, at CyberWallet, we utilize [CREATE2](https://docs.openzeppelin.com/cli/2.8/deploying-with-create2) to generate deterministic account addresses from initial access keys. Important to note that for now, we assume the initial access keys are the same across different EVM-compatible blockchains. This is to provide a familiar user experience for multi-chain account addresses, similar to EOA wallets. However, because access keys may need to change over time, this presents a new challenge. While technologies like [Keystore](https://vitalik.ca/general/2023/06/20/deeperdive.html) are very appealing, we believe that in the short term, developing cross-domain infrastructure and messaging channels is definitely needed to address multi-chain state management.

![img](https://link3-image.cyberconnect.dev/82f46803-4c8c-4bfb-835b-17ca1265f634/postImageFolder/3d70482f39591c49)

**Gas: Estimation, Overhead, and Management Across Blockchains**

One of the biggest challenges we faced was conducting gas estimation for different blockchains. While L1 blockchains have straightforward gas estimation, ERC-4337 Gas Estimation for L2s is exponentially more complex. `preVerificationGas` for a UserOperation is proportional to `L1_price/L2_price`, and in the case where this ratio increases from the time gas estimation happens until bundling, the `UserOperation` will be rejected. While adding a buffer (~25%) to `preVerificationGas` increases the likelihood of bundling, that also means users are stuck paying for this overhead, regardless of price fluctuations in L1 and L2 gas. We work hard to provide the right balance that offers ‘most likely’ bundling without rejection and the lowest gas overhead for users. Our friends over at Alchemy wrote an excellent [blog](https://www.alchemy.com/blog/l2-gas-and-signature-aggregators) about this problem.

On top of this, another rather well-known challenge is the gas overhead caused by ERC-4337 components like the EntryPoint contract and Paymaster. Granted, this seems like a major blocker for the mass adoption of smart accounts-powered web3 applications. However, we believe that proposed Ethereum upgrades like [EIP-4844: Proto-Danksharding](https://www.eip4844.com/) will significantly lower the gas cost on L2s and make this gas overhead negligible. Of course, as ERC-4337 adoption rates increase, smart contract accounts will soon be cheaper than EOAs on L2s as they start aggregating signatures to reduce the amount of data that needs to be stored on L1.

Finally, the true success of any technology is when it disappears from the end user experience. For web3 applications, this must manifest by the complexities of blockchains completely disappearing, even as users move across apps on different networks. However, gas management across blockchains comes with its own set of challenges. In CyberWallet, we introduced **SmartGas**, a universal gas credit represented in USD value. With SmartGas, users can deposit tokens as a gas credit on any supported EVM blockchain, which is then converted into their corresponding USD value at the time of deposit. Users may utilize this credit to cover gas fees when doing transactions on any of our supported blockchains.

**Fragmented Mempool**

Ideally, the alternative mempool introduced by ERC-4337 should be accessible and shared across different bundlers to maximize the chances of a UserOperation being included in a block. However, as the P2P mempool standard is still under development, and each bundler provider has their own mempool, we arrive, albeit hesitantly, at the crossroads of a fragmented mempool. Staying dedicated to our commitment to ensuring that users’ transactions are included in the block, we decided to connect several bundler networks behind our service to increase availability and prevent censorship. UserOperations are assigned to different bundlers and their traffic split changes dynamically based on a bundler’s error rate, RPC latency, and operation pending time.

![img](https://link3-image.cyberconnect.dev/82f46803-4c8c-4bfb-835b-17ca1265f634/postImageFolder/cffd0a9b124d79c0)

Although helpful, this multi-bundler strategy comes with its own set of challenges. It may be possible that a user unknowingly generates duplicate operations on multiple mempools with one operation’s gas going to waste. In any case, the UserOperation path must be properly tracked should we want to support UserOperation replacement in case of network congestion. To address this, we developed a ‘UserOperation + PmData lifecycle’ to have state transition properly defined so that user won’t waste their gas credit and can do advanced UserOperation management, such as speed up transactions by paying higher gas.

**Testing the Limits: ERC-4337 Boundaries and Scalability**

Arguably, one of the most aggressive yet ambitious steps we’ve taken as a team was to test the boundaries of our ERC-4337 implementation by letting over 150k users claim their [CYBER token](https://link3.to/cyberconnect/post/ae8c8e2bac694aca8c7f8fd9aee3c3da6dfaf5c5427de9db52e030e66eed7e80)[ community rewards](https://link3.to/cyberconnect/post/ae8c8e2bac694aca8c7f8fd9aee3c3da6dfaf5c5427de9db52e030e66eed7e80). It was an audacious test in retrospect, to say the least, as over 185k UserOperations were generated on the claiming day, and 70 ETH was consumed as gas fees. Even the L1 gas price pumped up when we started the claiming event.

![img](https://link3-image.cyberconnect.dev/82f46803-4c8c-4bfb-835b-17ca1265f634/postImageFolder/1cc1fd0726537fe8)

Here’s what happened:

Way too many UserOperations were submitted to multiple mempools across different bundlers in a short period. Subsequently, as the gas price increased, the gas estimation in the UserOperation was too low to be included on-chain. This resulted in an increasing number of UserOperations getting stuck in the mempool, unfortunately breaking the underlying bundler infrastructures. Reasons for failure range from depleted database resources to issues with node memory, and the message queue getting stuck. Our production environment experienced about one hour of downgraded service. Now is probably a good time for a special shout-out to our friends over at [Pimlico](https://www.pimlico.io/), [Alchemy](https://www.alchemy.com/), and [Biconomy](https://www.biconomy.io/) for their endless efforts in providing support during this emergency.

Our biggest lesson from this event is to ensure a proper UserOperation rate limit based on the blockchain capacity. Developing an advanced gas estimation strategy is crucial here, as submitting many underpriced UserOperations quickly doesn’t make any sense and only creates pressure on the mempools. Not to forget, bundling multiple UserOperations in one transaction can somewhat mitigate this issue. We were specifically impressed with Alchemy’s bundler that bundled around 10 UserOperations in one transaction to accelerate transactions to be included onchain during this event.

### CyberWallet: What’s Next?

Ultimately, CyberWallet’s core mandate is to bring web3 to the world and onboard 1B+ new users into crypto. We realize the audacity of that goal and are moving quickly to achieve that. The next series of feature updates to CyberWallet include a number of fundamental changes that are sure to make the user experience of blockchain-based applications safer and more intuitive.

Here’s what’s coming:

**Improved User Onboarding with Passkeys**

Cryptographic key management remains a major complexity for users just starting their web3 journey. Currently, CyberWallet adopts a traditional EOA (like MetaMask) strategy for key management. We will continue to support that, and in fact, in the coming weeks, we will expand that to develop wallet interfaces such as a CyberAccount MetaMask Snap.

We also believe that there is no single key management framework that meets the diverse range of use cases like mobile environments, gaming, and social. Therefore, we are adopting the PassKey standard to bring web2 product UX to web3. The device-bound design of PassKey is important in terms of security and accessibility. Important to note that Ethereum doesn’t natively support such a signature yet, so for the time being, we’re using a cloud enclave solution. Fortunately, key management does not require vendor lock-in because of the flexible authentication layer of the smart contract account. We will continue to follow developments in Ethereum, like [EIP-7212](https://eips.ethereum.org/EIPS/eip-7212), to adopt more progressive, and decentralized solutions.

**Cross Domain Framework**

With the emergence of more L2s and L1s, the isolation of blockchain systems is becoming a serious problem in navigating UX challenges. This is a complex problem and needs wider industry collaboration. We are actively researching and participating in topics like cross-rollup interoperability, AMPs (Arbitrary Message Passing Protocols), and independent mempool frameworks like Flashbot SUAVE, Anoma, etc. We believe a properly implemented cross-domain framework will elevate web3 UX significantly, and, therefore, become a core foundation for the future development of CyberWallet.

**Intent-based Design**

Paradigm recently helped revive [intent-based architectures](https://www.paradigm.xyz/2023/06/intents). ‘Declarative paradigm’ is not a new concept; in fact, many web2 systems are already built upon this programming design principle. It’s interesting to see such design thinking being adopted into blockchain systems. We are actively exploring use cases where users can express and execute their intent with multiple staged transactions. These use cases will be tested in the new version of the CyberWallet app in the coming months.

**CyberID and Multi-Chain Naming Solution**

The account states, such as underlying access keys, tend to change over time. We’ve found that it’s not practical to build upon the assumption that users will always have the same account address in a multi-chain world. In fact, it’s impossible to get the same account address in some blockchains, such as zkSync, even with CREATE2. Gnosis Safe wrote an excellent [blog](https://safe.mirror.xyz/4GcGAOFno-suTCjBewiYH4k4yXPDdIukC5woO5Bjc4w) about smart account addresses in a multi-chain landscape. We are developing our in-house naming solution called [CyberID](https://link3.to/cyberconnect/post/94e7b558022376aff30fabe6c8ed07be82b6ffdf6bd2c6665cdd054dc54b4da0) to allow users to track their account-related info such as counterfactual address, historical validators, and deployment status in a multi-chain setup. Therefore, users and dApps only need to use this human-readable identifier when interacting with CyberWallet.

### **Conclusion**

User onboarding remains one of the most crucial bottlenecks in bringing web3 to the world. The developments in account abstraction, especially most recently with ERC-4337, have been extremely significant and have brought the industry one step closer to mass adoption. The onus now lies on the web3 developer community and our ability to mobilize industry-wide collaboration to build user-friendly products that spread the promise of blockchains and decentralization far and wide. Our team at CyberConnect is excited to collaborate with you to bring web3 to the world.

For questions, queries, or concerns, feel free to reach out to:

- Zhimao Liu:  [Twitter](http://x.com/zhimao_liu) | [Telegram](http://t.me/zhimaol)
- Ryan Li: [Twitter](http://x.com/ryanli_me) | [Telegram](http://t.me/ryanli_me)

**About CyberWallet**

CyberWallet is a user-friendly smart contract wallet available on leading EVM blockchains, including Optimism, Arbritrum, Base, and others. To learn more, visit [wallet.cyber.co](https://bit.ly/3svmoIZ).

[Website](https://bit.ly/3svmoIZ) | [Twitter](https://twitter.com/thecyberwallet) | [Link3](https://www.link3.to/wallet)

2023 has undoubtedly been the year of account abstraction. Even though the concept is nowhere new, with early discussions dating back to the beginning days of Ethereum, the industry made its biggest leap this year in bringing working smart contract wallets to hundreds of thousands of users. Our tea

Building CyberWallet: Challenges and its Future

0x777

core contributor

PM at Crypto/Web3

Operation

degen

0xruhua

66lin

6262

Web3 Investor

🦙crypto Llama🦙

Ambassador and motion designer

bigwinner

crypto bear

Founding Partner

Crypto and gorilla lovers

Ambassador

brainer

cto at at

owner

Director

DisneyPlus

Electronic Arts

founder

Cryptocurrency community.

2266880

Investor , Researcher 

CEO, CTO,...

Co-Founder & CEO

investor

Crypto Enthusiast

13145208.eth

Link3 to the world in Web3

Manager

Ansari👑

Production Planning Manager

打工仔

mehdi65

Pharmacist

Proprietor

Myself

💎💎CEO of RETRODROP DAO💎💎 

LoveWeb3

crypto

worker

🍭Сrypto endocrinologist🍭

Free

shadows

twitter

漫谈Coin

Founder

Owner/Content Creator

Crypto Guy

UX/UI Designer, Investor

Dev Learner | web3 enthusiast

winner

Co-founder 

互关

银河会所