Listen to Arbitrum's founder talk about the future development path: Scaling technology parallel to Arbitrum One is on the way

Interview and Article by: Pan Zhixiong
Interviewees: Steven Goldfeder, CEO of Offchain Labs (Arbitrum) and the founding team

After the official launch and public testing of the Ethereum scaling network Arbitrum One, it has become the network with the highest asset lock-up volume among Ethereum L2s, even surpassing the second place by an order of magnitude. However, this is just the first scaling network launched by Offchain Labs, and there will be more scaling technologies launched in parallel with Arbitrum One.

For example, in a previous collaboration with social media and online forum Reddit, Arbitrum stood out among dozens of scaling solutions, and Reddit will adopt Arbitrum to expand its community points system (Community Points). The Arbitrum team told Chain News that based on our current estimates, various industry-specific and company-specific Rollup solutions will emerge in the future. Multiple Rollups can run in parallel, thereby increasing total capacity.

At that time, after researching and thoroughly reviewing multiple Ethereum scalability solutions, Reddit found that Arbitrum's Optimistic Rollups were the most promising scaling technology for the community points system, thus deciding to adopt the Arbitrum solution, which will first be launched on the Rinkeby testnet and then migrated to the Ethereum mainnet.

Regarding the current scaling effect of Arbitrum One, the team stated that there is still some room for improvement in the short term, but in the long run, it will still rely on Ethereum 2.0's data sharding technology to significantly reduce transaction costs.

In addition, we also discussed topics such as the centralization risks of a single Sequencer and views on other scaling technologies with the Arbitrum team.

What components make up the complete Arbitrum solution?

We are actually curious about what components make up Arbitrum as a novel and complex scaling technology, and what roles the frequently mentioned AVM and ArbOS play in this protocol.

The Arbitrum team stated that Arbitrum consists of four parts: smart contracts of the protocol, AVM (Virtual Machine), ArbOS, and asset bridge.

Smart contracts of the protocol: Sometimes referred to as EthBridge, this refers to the smart contracts of the Arbitrum Rollup protocol on Ethereum, ensuring that the business above this layer can operate correctly and providing the logic for mediating disputes that may arise in the Arbitrum protocol. (Chain News Note: The term Arbitrum comes from Arbitrium, which means arbitration, or resolving disputes).

Arbitrum Virtual Machine (AVM): The AVM virtual machine is positioned similarly to the EVM (Ethereum Virtual Machine), executing computer programs, reading inputs, and producing outputs.

ArbOS: ArbOS is an operating system (OS) running on the second-layer network, providing a fully compatible layer supporting EVM, and will also serve as the recorder, supervisor, and enforcer of smart contract execution on the Arbitrum chain. (For more about AVM and ArbOS, see the official documentation)

Asset bridge: Allows users to send Ethereum and other tokens between Arbitrum One and the Ethereum mainnet.

Is there still room for a decrease in transaction costs on Arbitrum One?

Although according to our previous estimation, the transaction costs of Arbitrum can be reduced by about an order of magnitude compared to Ethereum Layer 1, data from the L2 Fees website shows that it is still several times higher than the costs of some ZK Rollup networks.

However, it can also be seen that the L2 Fees website has made some notes for Arbitrum, indicating that the currently beta-stage Arbitrum One is artificially rate-limited, and this restriction may be lifted in the future official mainnet version, potentially further reducing transaction costs.

We also inquired about this topic with the Arbitrum team, and they stated that in the short term, they are working to reduce the base fees in each transaction, which means lowering the costs allocated to each transaction in batch transactions, and we will soon see a decrease in transaction costs.

However, in the long run, the Arbitrum One solution requires storing each transaction on the Ethereum mainnet via calldata, and as long as Ethereum's capacity remains unchanged, there will still be an upper limit. Therefore, the long-term scalability potential of Arbitrum One still relies on Ethereum 2.0's data sharding technology.

The Arbitrum team stated that ETH 2.0's data sharding will significantly reduce the cost of publishing data to Ethereum, which is crucial for Rollup scaling technology. ETH 2.0 data sharding means cheaper calldata, and we are excited to be part of Ethereum's (Rollup-centric) roadmap.

Are there more scaling solutions from Arbitrum?

The upper limit of Arbitrum One is constrained by Ethereum, so could other types of scaling technologies potentially break the ceiling for Rollups?

Teams have already begun exploring more efficient scaling solutions than Rollups, although this may sacrifice some security, the results are quite significant. For example, StarkWare's StarkEx solution provides a gas-free perpetual contract trading experience in the second-layer network through collaboration with dYdX. Meanwhile, Matter Labs will introduce zkPorter technology in zkSync 2.0 to reduce data availability, but at a lower cost.

Matter Labs' zkPorter off-chain data solution

In this regard, the Optimistic Rollup camp has not publicly disclosed more different scaling technologies, as removing the advantage of data availability may resemble Plasma technology. However, Arbitrum has already acknowledged that there will be more scaling solutions that will coexist with Arbitrum One.

The Arbitrum team stated that Arbitrum One is our flagship Optimistic Rollup solution, but in fact, we have other solutions that will coexist with Arbitrum One. According to our earliest published academic paper, Arbitrum was one of the first to explore hybrid data model scaling solutions, and we will continue to explore how these ideas can coexist with Arbitrum One, so stay tuned.

Can the Sequencer censor transactions?

In the L2 network, there is a role similar to miners in Layer 1 networks, responsible for processing the order of transactions, known as the Sequencer. For early L2 networks, there is currently only one Sequencer, which is the node operated by the team itself.

Therefore, the Ethereum community is very concerned about whether this single Sequencer may pose centralization or transaction censorship issues, such as selectively confirming transactions. Regarding this issue, the team stated that although there is currently only one Sequencer responsible for transaction ordering in the Arbitrum One mainnet, the Sequencer cannot censor transactions, because users can choose to bypass the Sequencer and directly publish their transactions on the blockchain, thereby forcing the transactions to be written into the blockchain.

Additionally, in the Arbitrum official documentation, the team stated that they have not yet added a mechanism to penalize the Sequencer, as this node is maintained by the official team and will not act maliciously, but in the future, Arbitrum One will switch to a decentralized fair ordering service.

Beyond the Sequencer, the network also needs more validation nodes (Validators) to review each transaction in real-time, making the network's security more reliable. They stated that although there is currently only one Sequencer, we hope to have more validators to ensure the correctness of the protocol, so we will announce well-known institutions that will participate in this validation process in the short term.