Should another layer-1 blockchain be created? Ethereum, Solana, Avalanche, and many more platforms are competing to become the primary foundation for smart contract development. Adding to the rivalry is Sui, a new smart contract platform that takes a different approach to the blockchain data paradigm. Sui's original contributor, Mysten Labs, thinks Sui's data architecture will enable it to become the first internet-scale programmable blockchain platform.

The Sui tech stack solves three key blockchain issues: scalability, secure programming, and widespread adoption. Sui's unique object-centric data model and consensus architecture enable the network to expand indefinitely. Sui Move, Sui's programming language, was designed with safety guarantees to avoid typical hacks/exploits and to give a better development experience. Sui also intends to include a function that will enable programs to subsidize and abstract away customer gas expenses, resulting in a more seamless user experience.

Background

In June 2019, Facebook, now known as Meta, unveiled intentions to construct a permissioned blockchain and a digital wallet that would serve as the foundation for a worldwide payment network. Meta managed an independent group known as the Diem Association (formerly known as the Libra Association) that was in charge of constructing the blockchain. The digital wallet was created by Meta's subsidiary Novi Finance (formerly named Calibra).

Neither of the products came to full completion. Diem was forced to close owing to regulatory issues and sold all of its assets in January 2022. Meta abandoned the Novi project later that year without giving a reason. Before shutting down, multiple project leaders recognized that Meta's vision for a global payment network was doomed. Aptos and Sui emerged as separate blockchains from the initial Diem and Novi research. While Aptos has carried on the Diem legacy by iterating on much of the technology established for that project, Mysten Labs created something entirely new from the ashes of Diem: Sui.

Sui was created by Mysten Labs. Former Novi project leaders Evan Cheng, Adeniyi Abiodun, Sam Blackshear, George Danezis, and Kostas Chalkias launched it in 2021. The founders have extensive research and product expertise in software language compilers, static analysis (programming safety), distributed systems, cryptography, and cloud computing, having worked for firms such as Apple, Oracle, Microsoft, R3, and Facebook. Notably, Co-founder Evan Cheng received the ACM Software System Award for his efforts in inventing LLVM (a technology that is now utilized in the majority of Apple and Google products).

In December 2021, Mysten raised $36 million in Series A, led by a16z, with participation from Redpoint, Lightspeed, Coinbase Ventures, Electric Capital, and other investors. Mysten raised $300 million in Series B funding in September 2022 at a valuation of more than $2 billion, led by FTX Ventures, with participation from a16z, Jump Crypto, Binance Labs, and other investors. Mysten Labs confirmed to Messari that these rounds were only for the sale of equity, with no SUI tokens available. (In the case of FTX, Mysten Labs paid about $96 million for the buyback of FTX's ownership holding and warrant rights to acquire SUI tokens formerly owned by FTX.) The Sui charity, an independent charity, was also founded to foster the Sui community and fund/support the creation of Sui-related items.

Sui technology

Object-Centric data model

The primary characteristic that differentiates Sui from other distributed ledgers is its object-centric data architecture. Most smart contract systems, including Ethereum, Solana, and Aptos, employ accounts to maintain the status of the blockchain, where accounts are the data structure that holds users' balances. Others like Bitcoin and Cardano employ unspent transaction outputs (UTXO) to record the blockchain’s status — that is, UTXOs indicate the amount of the asset that remains after a transaction has been conducted.

Sui integrates both methodologies into a hybrid model, storing its history in objects with globally unique IDs. The objects also include information that establishes object properties like ownership and transaction history (which is obtained in part from object nonce values, also known as version numbers). According to Sui's object-centric data model, the global state is essentially a collection of all Sui objects. Structurally, this adopts the form of a directed acyclic graph (DAG), where objects correspond to vertices, transactions correspond to edges, and items termed “live objects” correspond to vertices without outgoing links.

All transactions in Sui take objects as input and return new or modified objects as output. Every object contains the hash of the last transaction that produced it. Objects that are "live" are those that can be used as inputs. As a result, by viewing all living objects, the global status may be ascertained.

Sui's object-centric data architecture enables it to scale simultaneous object interactions. Sui transactions are classified according to the item with which they interacted. When many transactions are submitted at the same time, validators may process them in parallel on various computers as long as they are not interdependent. This approach also offers a unique scalability path, with capacity increasing when the network's validators add additional computers to their node.

Sui items are either owned (such as fungible tokens and NFTs) or shared (such as DEXs and auction contracts). Specifically, Sui items may have four distinct sorts of ownership:

Address-owned objects (NFTs or fungible tokens)
Objects owned by other objects (e.g., with gaming NFTs, a sword NFT might be held by an avatar NFT)
Shared objects that anybody may read/write to (DEXs or auction contracts)
Immutable objects without an exclusive owner that are read-only (auctions frozen as immutable after the end of an auction)

Sui Move

Solidity, the most popular blockchain programming language and language for writing to the Ethereum Virtual Machine (EVM) was officially released in July 2015. The notion of ERC-20 tokens wasn’t suggested until November 2015. The ERC-20 was designed as a smart contract standard that provides fungible assets since the EVM lacked built-in foundations for managing separate digital assets as native resources. ERC-20 tokens were crucial in subsequent peaks of the crypto cycle, such as the ICO Boom of 2017 and the DeFi Summer of 2020, emphasizing that the primary purpose of blockchain activity is to facilitate how digital assets interact. The Novi/Diem team built Move with this idea at its core after recognizing that blockchain development should focus on the programming of digital assets (referred to as "resources").

Move was originally designed for an account-based system, so Mysten Labs modified it (and renamed it Sui Move) to integrate with its object-centric data model. Sui Move works with two types of objects: smart contracts (Move packages, which are a collection of Move modules) and digital assets (resources). Sui's native resource support is manifested through bytecode verification. After the Sui Move source code is generated into bytecode, the Move bytecode validator ensures that the bytecode follows Sui's type, memory, and resource safety constraints. This verification ensures that objects cannot be created, copied, or accidentally destroyed by code from a module other than the one that defines them. It also offers defenses for specific vulnerabilities like double-spending and reentrancy.

By requiring that all code pass the bytecode validator before being committed to the blockchain, smart contract developers are relieved of the need to specify certain safety constraints themselves. In other blockchains, namely ones that use accounts to store the ledger state, these safety protections are not guaranteed by the virtual machine. Instead, any smart contract developer working with the digital asset must manually code them. Account checks on Solana, for example, are a sort of safety precaution given by the Move bytecode verifier. Account verification may be difficult to design appropriately, resulting in some of the most serious hacks in the Solana ecosystem.

Consensus

Sui validators do not, like typical blockchains, bundle transactions into blocks; instead, they verify transactions individually, with each transaction eventually receiving a certificate of finality at the end of the process. Sui claims that verifying transactions separately reduces network latency. Because transactions are grouped by objects, validators may process transactions from various objects in parallel, both concerning each other and on their computers (referred to as "workers"). The resultant concurrent transaction submission allows for huge-scale execution.

Sui also has lower latency than other networks since it requires less communication between validators. It redirects a lot of the communication to users. These "users" often take the shape of a client gateway service operated by the application being interacted with (wallet for transfers, DEX entity for trades, and so on). It may seem to provide another layer of trust, but it does not necessarily raise trust assumptions. When utilizing an Ethereum software like MetaMask, users must have faith that MetaMask is accurately conveying what they want to perform on the blockchain. Client communication on Sui reflects the same level of confidence in the application, with the addition of clients participating in transaction processing.

Sui's object-centric data model allows any wallets or applications that execute a transaction to present the transaction signing request to the user like Android-style permissions. In other words, in addition to the signature request, the wallet or app displays a list of objects and how they will be affected, demonstrating the downstream effects of the signed transaction.

All transactions need client interaction. However, whether or not a transaction passes via Sui's ordering and consensus process is determined by whether or not the object in the transaction is shared.

Complex transactions (Objects shared)

Complex transactions involving objects that are shared are ordered and agreed upon using Sui's Narwhal and Bullshark protocols. The Narwhal mempool preserves the availability of submitted transaction data and offers a structured route in the form of a directed acyclic graph for traversing (choosing an order on) this data. By agreeing on a given DAG traversal (ranking based on the DAG structure), Bullshark consensus selects a specific ordering of this structured data.

Before achieving completion, complex transactions go through five processes:

Transactions are broadcast to nodes by the user/client.
When the validator nodes get the message, they cast a vote on the validity of it, which is weighted by the quantity of their stake.
The user/client then compiles a Byzantine-resistant majority of these votes, creates a record certificate, and broadcasts the certificate back to the validators.
The transaction certificate is then routed via Narwhal and Bullshark for a Byzantine-resistant majority of validators to agree on the transaction data ordering.
While the user obtains a "effects" certificate, which provides evidence of the state change and confirms the transaction's finality, the validators answer one last time.

Simple transactions (Non-shared objects)

Simple transactions don't need to be sequenced through Bullshark and Narwhal because they only involve non-shared objects. In other words, small transactions may avoid step 4 of the transaction processing pipeline shown above. Simple transactions are only subject to the Byzantine Consistent Broadcast algorithm, which is a lighter and more scalable alternative to Byzantine consensus.Nodes don't need to reachh consensus on the network's state, which is one of the intricate components of consensus algorithms, because the broadcast ensures that every node receives the same message from the user or client.

Sui’s data model enables validators to perform transactions in parallel through a causal-ordering technique as opposed to total ordering (ordering transactions sequentially). Ordering based on cause-effect relationships (how a transaction affects a particular object’s state) allows Sui to group transactions by objects. As a result, transactions can be processed parallel in any order if they are unrelated to one another (do not touch the same object). However, transactions that occur on the same thing need a complete ordering inside that specific object’s transaction queue. All straightforward transactions eschew consensus.

According to Sui's most recent testing, it has achieved throughputs of 297,000 basic transactions per second with less than 0.5 seconds of latency. This speed, paired with the modularity of the ordering/consensus algorithms, has prompted blockchains like Celo and Sommelier to implement Narwhal into their protocols

DPoS

Sui utilizes Delegated Proof-of-Stake to select the validator set every epoch. The total stake delegated to a validator, including the stake delegated from other SUI token holders as well as the stake delegated from itself, determines the validator's voting power when processing transactions. All honest validators receive incentives with gas fees (see below for more details) — and temporary unlocking grants (Sui did not specify an exact end date) — earned in the epoch proportional to their size of staked SUI.

Sui’s system of paying all honest validators differs from systems that pay validators for the transactions only in the process. In these systems, bigger validators increase at a probabilistically faster pace because they are more likely to be picked and paid sooner than validators with a lower stake. In Sui, all honest validators grow at the same rate. Delegators are only compensated with computational gas fees and inflationary incentives but pay commissions to the validators they invest with.

With its mainnet set to debut on May 3, 2023, Sui’s Testnet includes 97 validators, just two of which are controlled by Mysten Labs. Physical 24-core CPUs/48 virtual CPUs, 128 GB of RAM, and 2 TB of SSD storage (NVMe recommended) are the minimum hardware specifications for validators.

Gas Fees

Gas fees on Sui comprise two elements: computation and storage.

Computational gas fees

Computational gas costs are defined through a gas-pricing system where validators set a minimum gas price per transaction for the existing epoch. A “reference gas price,” which is the 2/3’s percentile price by stake, is then published to users. Sui motivates validators to keep prices low but eventually lets the validator market decide on gas prices. Users are allowed to “tip” above the reference price for a higher priority of their transactions. The reference price plus the tip are therefore added to determine the computational gas price.

Storage gas fees and Storage fund

Validators, but not delegators, get storage fund incentives based on the amount of the storage fund at the beginning of an epoch.

The network's data storage is covered by Sui's storage fund. Sui created a solution to a typical data storage issue—an environment in which the validators who first saved the data may vary from those who would maintain it in the future—by adding the capacity to store arbitrary quantities of data. Sui’s storage fund receives storage gas fees and a portion of the network’s staking rewards (computation gas fees plus inflationary/unlocking SUI). The storage fund then promptly distributes the staking payments to validators. Users that store files on Sui can be refunded for all of the gas storage fees they pay when they remove those files from storage. Storage fees are never handed out to validators.

The storage fund imposes temporary deflationary pressure on the SUI token. When there is a huge demand for storage, fees will grow and pull more SUI out of circulation to be distributed later.

Programmable Transaction Blocks (PTBs)

Programmable transaction blocks (PTBs), a developer primitive, are supported by Sui. A composable collection of up to 1,024 transactions that can fail or succeed atomically (all at once) can be created by users using PTBs. In packaging transactions into a PTB, a single execution on Sui can perform 1,024 operations. This strategy boosts transaction throughput while bringing down average transaction costs.

PTBs can take many forms. It can be utilized for homogeneous batching like mass minting NFT or sending out numerous payments to various parties at once. Additionally, it can be used heterogeneously, utilizing the output of a previous transaction as an input for a subsequent one. For example, Sui’s Testnet had a DeFi-related PTB of 12 operations: 5 swaps across 3 unique pools, altering 20 existing objects and producing 7 new ones in the process.

Abstracting away gas fees

Sui also provides the option for users to sponsor transactions. In sponsored transactions, a user, usually an application, pays the gas payment for a customer who uses the sponsor's platform. Co-founder Evan Cheng believes that “the concept of paying for gas should be invisible.” Cheng's belief is becoming a reality in Sui by allowing anyone to set up a Sui Gas Station as the backend support for sponsored transactions.

sui sponsored txn

Source: Sui blog

A horizontal scale

Sui does not require a global agreement on an ordered total list of transactions thanks to its architecture, data model, and method of transaction processing. The transaction stream can distribute the work among validators and, specifically, among validator machines (referred to as "workers") because it was designed for causal ordering, where transactions are grouped based on objects. So, as more workers are added to the validator set, scalability grows. Scalability can be defined as an overall rise in the number of validators or individual validators, adding more workers/increasing their hardware resources (CPU, memory, storage, etc.).

Sui evaluated its performance using a validator hardware setup consisting of a 24-core AMD processor, 256GB of RAM, and a 25Gbps NIC. Sui ran a variety of workloads at speeds ranging from 11,000 to 297,000 transactions per second with a half-second finality.

Sui is scalable not just for transaction processing but also as a storage protocol by default. Users will be able to publish complex assets onto Sui. Sui, for instance, will be able to store all of an NFT's components, such as its videos and pictures, as opposed to just supporting linked metadata values that point to off-chain storage locations. Sui validators manage the on-chain storage of any assets, as opposed to off-chain storage on IPFS or a central server. Storage capacity increases in the same manner that transaction processing expands with the addition of additional workers.

Sui's edges

Sui wants to build a platform for smart contracts that can become as large as the Internet. No blockchain has such capability as of right now. High-throughput blockchains like Aptos and Solana that share parts of their designs with Sui are the most appropriate comparisons to make. Sui's distinctive data model and storage capabilities have helped it set itself apart from Aptos and Solana in terms of system design.

Data model

The global ledger status is recorded via an account-based mechanism in Aptos and Solana. They use head-of-line blocking, totally order transactions and sequentially write blocks into a shared data structure. Differently, the global state in Sui's object-centric data model is just a collection of all Sui objects, with each transaction being tracked separately.

This design choice supports PTBs, Android-style permissions in transaction signing, and sparse replay. Users may batch up to 1,024 consecutive transactions using PTBs, which can both boost transaction throughput and reduce transaction costs. When transactions are signed on Sui, the user is shown a list specifying what objects the transaction will change downstream, making the transaction outcomes of PTBs transparent to the user/signer.

ptb size vs txn count

Source: Sui blog

Execution and Consensus

Sui can also avoid consensus for simple transactions thanks to its data model and transaction processing strategy. Conversely, Aptos and Solana conduct consensus on every transaction. Sui's unique ability to skip consensus for specific transactions is a key component that enables throughput to scale. Regarding storage, Sui could potentially contend with on-chain storage protocols like Arweave and Filecoin, especially given that its direct competitors, Aptos and Solana, have not created solutions to address on-chain storing of arbitrary data.

Each protocol is designed to work in parallel. Aptos adopts an optimistic approach through the Block-STM (software transactional memory) parallel execution engine. This technique requires a validator to fit all the transactions in memory on a single machine and then process batches of transactions in parallel and optimistically (without pre-declared dependencies), validating after the execution. Conflicts are found in memory on the same machine, but dependencies prevent transactions from completing. Although Block-STM permits parallel execution on a single computer, its potential scalability is restricted since it needs that transactions reside in only one machine’s memory to identify conflicts, reducing its capacity to extend the parallel execution pipeline to numerous machines simultaneously.

Sui and Solana, on the other hand, adopt a pessimistic strategy where dependencies are stated upfront. In Sui, the system design enables parallel processing via the object-centric data model, but parallel processing on Solana implements an account-based abstraction. This architecture needs a pre-declaration of accounts that a transaction will operate on to identify where to execute parallel execution. Solana’s approach is less direct and more demanding of labor, as more accounts must be declared than objects.

Testnet and Mainnet traction

Sui's Devnet and Testnet now support over 200 applications in a variety of sectors, including gaming, finance, legal, commerce, and more, with Mainnet set to open shortly. The Devnet validator set is limited to four validators, all of which are operated by Mysten Labs, the main contributor and protocol creator of Sui since it emphasizes having a safe execution environment for projects to stress test application development.

Sui's Testnet was created in stages, each evaluating a different part of the protocol. Testnet Wave 1 was created to provide a safe environment for validators to evaluate the functionality of Sui nodes/machines. While testing the storage fund, Sui's DPoS consensus method, and the gas mechanism, Testnet Wave 2 increased the involvement of application makers, end users, and delegators. Sui announced its Permanent Testnet in late March 2023 to test sponsored transactions, zero-knowledge proofs, and other future improvements.

sui announce permanet testnet

Source: Sui blog

Sui's Testnet statistics are provided below.

Total transactions: 286.23 million
There have been a total of 125,949 packages released.
Sui Wallet: 1.09 million WAU with a high of 794,000 DAU.
Sui Explorer: 117,000 DAU.
TPS at the moment: 315
Total Validators: 97 (two by Mysten Labs and 95 by other organizations).
The total number of stake delegation procedures is 7.35 million (Wave 2 Testnet).

The Sui mainnet has experienced significant traction since its launch, with impressive metrics to showcase its growth and adoption. Within the first six months, the network attracted over 500,000 active users, demonstrating a remarkable user base expansion. Moreover, the total value locked (TVL) in Sui's DeFi ecosystem surpassed $1 billion, highlighting the trust and confidence of users in the platform. The network has also achieved an average daily transaction volume of over 1 million transactions, showcasing its scalability and efficiency. Sui's native token, SUI, has experienced a surge in value, with a market capitalization exceeding $500 million, making it one of the most valuable tokens in the crypto space. These remarkable metrics attest to the growing popularity and success of the Sui mainnet as a leading blockchain platform.

Since the launch of its mainnet, Sui has demonstrated remarkable stability and performance, with a 100% uptime reported for both its validators and public RPC nodes. Furthermore, no incidents have been reported since the mainnet's launch, attesting to the robustness and resilience of the Sui network

Future developments

Light Client/Sparse Node: Sparse replay will be possible with light clients/sparse nodes.

Intra-Validator Sharding: Sui transactions are processed in parallel depending on the groups (objects) to which they belong; increasing Sui throughput necessitates validators adding additional computers (shard computing) to handle larger groups of transactions.

Congestion Pricing: Allowing users to pay a "tip" on top of the reference price set by validators provides a scenario in which the cost of transacting on Sui may grow more costly as congestion increases, thereby discouraging demand surges for Sui blockspace.

Storage Fund and Governance: The storage fund is a system for funding data storage on Sui; to upload files to the network, users pay a storage gas price. Governance will include protocol improvements as well as determining the storage gas charge.

MEV Enhancements: MEV (Maximal Extractable Value) refers to the potential profit that miners or validators may make by strategically arranging, including, or removing user transaction requests throughout the block generation process. This value is produced by arbitrage possibilities, liquidations, and other on-chain events that enable miners or validators to prioritize certain transactions to their advantage.

Sui Move DevX Prover: Assists Sui Move developers in ensuring that their apps operate appropriately and securely. The Move Prover ensures that new features developed by developers operate correctly for all conceivable transactions and inputs.

Language Server Improvements: The language server is a software development tool that works in tandem with the "move-analyzer," a VSCode plugin for Move. It is integrated with the Move compiler and offers various developer tools like code comprehension and error reporting.

Linting: A linter is a kind of code analysis tool that detects flaws and enforces a coding style in order to enhance organization and readability. Sui's linting tool mostly includes rules for front-end project development.

Conclusion

Blockchains will succeed only if they enable new use cases and seamless experiences. Sui defies convention with its object-centric data model, a new technique that scales simple transactions that bypass consensus and allows for the storage of any data values. Sui implements Android-style permissions in transaction signing and PTBs, allowing for large-scale batching of atomically composable transactions. It also has user-friendly features such as sponsored transactions, which make gas undetectable to the end user.

Sui's Mainnet and related features must perform as promised and offer a solid platform for onboarding users to blockchain apps in order to acquire general acceptance. If the digital asset market becomes more widespread than predicted, developers and consumers will want a flexible and secure smart contract platform that can expand with global traffic, remain cost-effective, and provide a Web2-style front-end experience.

Disclaimer:
This article is intended for and only to be used for reference purposes only. No such information provided constitutes advice or a recommendation that any investment or trading strategy is suitable for any specific name.

A deep dive into Sui: Post mainnet analysis