Podcast Notes | Interview with Sui Co-Founder: What Makes Move-Based Sui Stand Out?
TechFlow Selected TechFlow Selected
Podcast Notes | Interview with Sui Co-Founder: What Makes Move-Based Sui Stand Out?
On the Sui platform, issues such as scalability, performance, and gas stability have already been resolved.
Navigating Web3 tides with focused insightsCompiled & Translated by TechFlow
In this episode of Lightspeed aired on December 28, 2023, Garrett, Mert, Sam, and Evan discussed Sui's origin story, the Move language, Sui's features, and more. Please read the notes below for further details.
Program Background
-
Garrett Harper (Host) – Co-host of Lightspeed
-
Mert (Host) – CEO of Helius Labs, Co-host of Lightspeed
-
Sam Blackshear (Guest) – Co-founder of Sui
-
Evan Cheng (Guest) – Co-founder of Sui
The Origin of Sui
-
Sam said he began developing Sui in November 2021, at the peak of the previous bull market. While market hype made it seem like a favorable time to launch a new blockchain project, other factors were also at play.
-
He added that the decision to launch Sui was not driven solely by market conditions, but rather by identifying gaps in the market and recognizing opportunities for improvement.
-
Sam mentioned that he and Evan had both been part of the Facebook Libra team since 2018. One of their responsibilities within the team was building a global, blockchain-powered compliant payment network. Sam led the creation and design of the Move language and implemented it within Libra.
-
Evan said he was part of the research team focusing on various aspects of Libra’s design and implementation. During their work on Libra, they encountered certain limitations that motivated them to explore ways to improve existing solutions.
-
He added that they developed a playbook for Libra’s next-generation version, including advanced consensus mechanisms, enhanced smart contract programming, and additional cryptographic primitives. Despite technical progress, Libra still faced challenges unrelated to the technology itself.
-
Evan said they realized many people needed what they wanted to build, so they decided to try launching their own blockchain project. The technical blueprint they developed for the next-generation Libra became the foundation of Sui.
-
He added that his background lies in building developer platforms and products, not just conducting research. He aimed not only to build a blockchain but to create a genuine developer platform for product builders. With Libra nearing completion, the timing and circumstances aligned perfectly for him to start his own venture.
Move: A New Programming Paradigm
-
Sam said when he initially joined the Libra project, creating a new programming language wasn’t part of the mandate. However, he was given the freedom to either explore existing options or develop smart contracts from scratch on Libra.
-
He added that before making design decisions for the Move language, he spent months carefully studying Solidity and the EVM.
-
Sam said a smart contract language should resemble a domain-specific language (DSL) more than a general-purpose programming language. These languages focus on specific tasks such as creating objects, transferring and sharing objects, and performing access control checks. The core abstractions in smart contract programming should revolve around these tasks to ensure security and efficiency.
-
He added that they decided to design a new language called Move instead of trying to adapt existing technologies like the EVM or Solidity. Starting from B code, Move incorporated features such as memory safety, type safety, resource safety, and human-friendly handling of objects/assets. The development process included implementing example code for Libra and comparing it with Solidity code to evaluate their respective advantages.
-
Sam said the focus was on attracting developers to build with Move by offering compelling incentives. Based on cryptocurrency wallet installations (approximately 60 million), the total addressable market is estimated at around 60 million. Overcoming barriers such as complex key management, wallet installation, and native token acquisition is critical to attracting developers.
-
He added that their goal is to transform these challenges into technical problems—such as enabling ZK login so users can transact without installing a wallet. Sponsored transactions allow others to pay transaction fees, enabling developers to explore traditional revenue models.
-
Evan said when pitching to web3-native developers, the emphasis is on the strengths of the language and the entire platform. It's essential to highlight that the design of the blockchain system must align with the chosen language. For large product makers, understanding their specific pain points and providing solutions is crucial.
-
He added that using a popular language like Rust does not automatically attract the entire community associated with that language. Rust lacks built-in functionality for fundamental blockchain operations such as asset transfer or account management. Libraries must be added to provide these capabilities, so it's necessary to go beyond the language itself.
Sui: Technical Deep Dive
-
Evan said many blockchains place excessive emphasis on the complexity of underlying technologies, requiring developers to consider various scaling strategies and component choices.
-
Sui aims to simplify this by providing a platform where issues such as scalability, performance, and gas stability have already been resolved.
-
He added that Sui focuses on:
-
Ultra-low latency
-
Horizontal scalability
-
Emphasis on efficiently solving real-world problems
-
-
Sam said in blockchain systems, ownership is represented using hash tables. Keys in the hash table are addresses, and values represent balances or other data.
-
He added that transactions are sent by owners, who can only perform certain actions based on the code. Different types of objects such as NFTs or smart contracts can be stored in the hash table.
-
Sam said in Sui, global storage is represented by a set of objects each with a unique ID. Metadata is embedded within objects, including an owner field. An owner can be an address, another object’s ID, immutable (smart contract), or shared (accessible by multiple users). This model differs from the account-centric model, where ownership is represented via hash tables.
-
He added that low latency is crucial for use cases such as gaming or real-time transactions. In traditional EVM systems, transactions undergo consensus ordering before execution. Solana introduced a fast path using Byzantine Consistent Broadcast, allowing transactions to execute without full consensus. Compared to systems requiring full consensus, this significantly reduces latency.
-
He added that Sui is designed to handle a high volume of transactions. Similar to Solana, the data model in EVM systems involves transactions operating on specific objects. This enables parallel execution and easy scheduling of tasks within the Sui system. Gas pricing and incentive mechanisms are designed to manage sequential contention on shared states without affecting other transactions.
-
Sam said Sui’s data model enables parallel execution by specifying input objects for each transaction. The scheduler can assign tasks to appropriate cores or machines, enabling efficient parallel processing. Gas pricing and incentive mechanisms aim to prevent sequential contention on shared states while maintaining service quality for other transactions.
-
He added that Sui delivers practical finality within 25 milliseconds, providing rapid confirmation for transactions. Full finality is achieved within 400 milliseconds, ensuring all necessary processes are completed. End-to-end latency is approximately 480 milliseconds, clearly outperforming systems relying entirely on consensus.
-
Sam said parallel execution allows the system to handle a large number of transactions simultaneously. The goal is to avoid limiting throughput based on the capacity of a single machine or validator. What’s needed is the ability to distribute execution across different workstations and elastically scale. Scaling up by adding more machines during transaction peaks and scaling down during normal traffic periods is essential.
-
He added that the object data model plays a crucial role in achieving scalability. Native ownership information and types within objects enable efficient organization and retrieval of data. Objects can be sorted by owner address, type, or common combinations. Flexibly changing schemas based on evolving workloads helps optimize throughput and storage scalability.
-
Sam said validation reads are performed by committing transaction results and validator signatures. Individual transactions are verified using BLS signatures instead of relying on a global Merkle account tree.
-
He added that in the fast path, there is no strong correlation between network latency and the increase in validator set size. Regardless of the number of validators, users broadcast transactions to validators in parallel. Only BLS signature aggregation time increases as the number of validators grows.
-
Sam said in Narwhal consensus, the relationship between network latency and validator set size resembles that of the fast path. Users broadcast transactions to validators in parallel, regardless of the number of validators.
-
Sam said the long-term roadmap includes elastic scaling through the ability to spin up new machines. Currently, SW and Mana run on a single-machine architecture but can leverage multiple cores. Workstations are treated as cores rather than individual machines. Transactions are routed to specific workers based on the types of objects involved.
-
He added that isolation can be achieved by assigning specific workers to handle certain types of objects or transactions. All necessary reads and writes can occur on the same machine, reducing cross-machine communication. System scalability depends on successful isolation and efficient resource allocation.
Validator Requirements
-
Sam said there are currently about 100 validators in the network. Validators can apply for shares from the Sui Foundation, which helps achieve geographic distribution and diversity of data centers.
-
He added that over time, as more people secure their own funding, they can launch new validators and compete with existing ones. Validator count is not the ultimate measure of decentralization; instead, the focus is on cost-effectiveness for users.
-
He added that the primary metric for decentralization is the cost for users to replicate transactions affecting their state. Sui leverages the object data model to enable "sparse nodes," allowing users to efficiently verify only transactions touching their objects.
-
Sam said as transaction volume increases, running a full node or validator becomes impractical for average users. By collectively tracking object changes, users can ensure validators comply with the protocol.
Optimistic VS Pessimistic Parallel Execution
-
Sam said Sui supports both static and dynamic techniques for parallel execution. Static information aids initial parallelization, but dynamic conflict detection can be added when needed. Incentives are required to provide workloads that can be parallelized.
-
He added that detecting conflicts at the static layer can prioritize or deprioritize transactions based on conflict scenarios. The system is expected to evolve further as more static information and advanced techniques emerge.
State Growth and Storage
-
Sam said the main reason for EVM’s slow performance and low transactions per second (TPS) is its measures to prevent state growth. Sui’s goal is to allow state growth while avoiding reliance on giant Merkle trees or other global authenticated data structures.
-
He added that traditional databases can be used for storage, easily scaling by adding more disks or upgrading hardware. Since storage costs are much lower compared to traditional solutions (like Sal), state growth is not an issue. NFT compression is currently being discussed as a way to reduce storage costs, but overall, state growth is manageable.
-
Sam said in Sui, the account/object database is stored on nodes for transaction processing. However, historical transaction logs require separate archival storage for syncing full nodes or conducting historical analysis. Archive nodes maintained by organizations such as Sweet Foundation and Miss ensure long-term storage availability.
-
He added that Solana faced similar challenges and has long implemented various solutions, including using Google and RWE for archival storage.
-
Sam said checkpoints serve as an audit mechanism, committing prior checkpoints into a sequence. Epochs occur every 24 hours, summarizing the checkpoints within. Each checkpoint commits the transactions it contains, ensuring data integrity.
-
He added that validator set changes and stake-weight status updates are also signed and reflected in checkpoints. To verify old data, one can start from genesis, track validator set changes until the desired data is found, check signatures, and validate transactions.
-
Sam said about 19% of transactions on Sui use the fast path. Fast-path transactions involve simple transfers, such as sending USDC to another account. Interactions with smart contracts like Uniswap on Ethereum do not qualify for the fast path due to involvement of multiple parties.
Move’s Security Advantages
-
Sam said Move’s security isn't solely based on the fact that no hacks have occurred so far. No smart contract language can prevent programmers from writing insecure code.
-
He added that Move addresses issues present in existing languages, such as reentrancy, dynamic dispatch, ownership checks, and object serialization. Move is designed to eliminate these problems, providing developers with a safer environment.
-
Sam said Move’s key value proposition lies in its typed objects and type safety. Type safety applies not only within programs but also when interacting with external code or untrusted sources.
-
He added that integrating Move into other systems may require considering account-based systems, interactions between Move and non-Move programs, and running validators to ensure security guarantees.
-
Sam said integrating Move into other ecosystems might require addressing account-based systems and interactions between Move and non-Move programs. Adding compilers could introduce risks, but Move’s key value proposition remains its typed objects and type safety.
-
Evan said to fully realize Move’s security advantages, all smart contracts in the system must consistently use Move. When all contracts are written in Move, validators can run and enforce strong assumptions about code and objects. He drew an analogy to early technologies—using abstraction techniques reduced errors compared to manual coding.
-
He added that having a blockchain alone is insufficient—it needs compatibility with different languages to replicate the same stack. Replicating the entire stack is crucial to gaining all the benefits of blockchain technology.
Economic Security
-
Sam said economic security is assessed by how much stake someone needs to acquire to cause undesirable outcomes in the system. From the beginning, new systems implement staking subsidies to ensure sufficient stake from day one.
-
He added that currently, out of a total supply of about 10 billion tokens, around 8 billion are staked, indicating good economic security.
-
Evan said economic security attacks on blockchains are rare, especially in the early stages when ecosystem growth makes such attacks prohibitively expensive.
-
Sam said while economic security is an important aspect, social consensus serves as the ultimate safeguard against attacks that could damage the network’s reputation.
Atomic Composability, Scalability, and Validator Sustainability
-
Sam said the utility of a smart contract platform can be measured by multiplying the value of assets on the platform by their programmability. Atomic composability plays a vital role in programmability, allowing developers to access all assets and build applications without worrying about implementation details or varying latencies. The vision is to place everything of value in one place and provide a programming model that enables frictionless automated access to all assets. Splitting state across multiple layers or sub-chains might create user experience issues, but this is seen as an acceptable trade-off to achieve atomic composability.
-
He added that the future of crypto envisions different types of rollups, just as there are new websites for different purposes. Atomic composability is considered crucial for blockchains because it allows easy access to all valuable assets and facilitates seamless interaction between users and applications. It enables users to interact without needing to understand implementation details or differences in security models of various assets.
-
Sam said scaling Layer 1 to accommodate all use cases on the same layer is crucial for expanding application scope. If bottlenecks arise that require splitting state or degrading user experience, it may be necessary to explore alternative solutions.
-
He added that pushing technological limits determines the scale and impact of Layer 1. Before considering sharding or bridging, improving throughput and scalability at the base layer is beneficial.
-
Sam said building a network with a large user base and consumer-grade experience may require higher hardware demands. Each community must decide the value of accessibility versus lowering the barrier to node operation.
-
Evan said due to scalability limitations, Ethereum’s user experience is not seamless for consumer applications. In contrast, staking on newer chains like Solana is relatively straightforward.
-
He added that the role of service providers in node operations raises concerns about centralization. Questions about the true Nakamoto coefficient and whether services count as a single validator or multiple validators remain controversial.
-
Evan said balancing hardware requirements, accessibility, and sustainability is an ongoing challenge. The pace of hardware advancement relative to blockchain demand affects long-term sustainability.
-
He added that there are few blockchain solutions available for consumers. Limiting solutions to those unfriendly to consumers hinders the realization of potential value.
-
Sam said decentralization doesn’t mean anyone can run their own node. That’s unrealistic. The cost of operating a node may exceed income from staking rewards or transaction fees. When evaluating the value of running a node, other factors must be considered, such as additional income streams and faster read speeds via RPC providers.
-
He added that running a node offers benefits beyond direct economic returns, such as offsetting costs through other revenue sources. Nodes can also act as RPC providers, offering users faster read speeds. Some operate nodes for security reasons or as side businesses, further promoting ecosystem development.
Sui’s Beachhead: Gaming and Commerce
-
Evan said the gaming industry is known for experimenting with new business models and technologies, making it an interesting area to explore. Blockchain’s ability to handle assets aligns well with gaming industry needs. Almost all consumer-facing commerce can benefit from blockchain technology.
-
He added that blockchain’s core value lies in its ability to facilitate asset-related activities in commerce. Commerce encompasses various activities, including payments, asset exchanges, and staking. Gaming serves as the initial use case to introduce blockchain technology to the public.
Sui vs Aptos and Future Outlook
-
Sam said the focus should be on how developers can reach as broad an audience as possible. Building on Sui offers many benefits, such as ZK login for users without wallets and sponsored transactions enabling monetization without requiring tokens.
-
He added that features offered by Aptos include enforced royalties for digital objects and restricted closed-loop tokens.
-
Evan said including features in the protocol makes life easier for developers, eliminating reliance on centralized service providers or monthly fees. Ethereum allows a more free-market approach, where developers must opt into certain features that may not be universally supported. Writing key features directly into the protocol is important for product builders, reducing competition over basic functionalities.
-
Evan said Ethereum faces challenges in adopting new standards, as high gas fees push some features off-chain or make them inefficient. In contrast, Sui aims to incorporate key features directly into its native protocol, making them easily accessible without relying on external services. Social consensus plays a crucial role in adopting new standards, but achieving universal agreement is challenging.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
深潮TechFlow
