Understanding the new public chain Aptos: the “Exodus” of the original Diem team rebuilding the public chain

Although this is the first round of financing, Aptos cannot simply be regarded as a start-up project in terms of technical reserves or ecological cooperation. Starting from Diem, can Aptos become a disruptor in the public chain?

The aborted Diem stablecoin project is being reborn in another form.

On March 15th, Aptos, a new public chain project established by the original Diem team members, completed a strategic financing of US$200 million, led by a16z, and participating investors include Multicoin Capital, Katie Haun, 3 Arrows Capital, ParaFi Capital, IRONGREY, Hashed, Variant, Tiger Global, BlockTower, FTX Ventures, Paxos, and Coinbase Ventures, among others.

Aptos kept “distance” from Diem. As Kyle Samani, a partner at Multicoin Capital, points out, Meta (formerly Facebook) and the Diem Association are not investors in Aptos, nor anyone at Meta — including David Marcus, Mark Zuckerberg or anyone currently Meta’s leadership — none of them invested in the project. In the article “The Origin of Aptos”, founder and CEO Mo Shaikh also said, “The team is the original creator, researcher, designer and builder of Diem, the first blockchain built for this purpose.” , ” Since leaving Meta, we’ve been able to put our ideas into action, get rid of the bureaucratic red tape, and build a whole new network from the ground up to bring them to life”.

Understanding the new public chain Aptos: the "Exodus" of the original Diem team rebuilding the public chain

Judging from the project concept that has been announced so far, Aptos, as a Layer 1 blockchain, is still based on Diem’s ​​open source code base, uses the Move language as the development language, and adopts the iterative Byzantine consensus protocol to improve the security of the blockchain. scalability and scalability to create an infrastructure network that can reach billions of people. At present, Aptos’ development network devnet has been launched, and it is expected to launch the main network in the third quarter of this year.

Aptos, which means “people” in Oroni, is the name of a small town in Santa Cruz County, California, traditionally inhabited by the Oroni, the Native American people of the San Francisco Bay Area. The team takes this name and gives a glimpse of its vision. As Mo Shaikh puts it, “As a team, our drive is to believe that we are the right people to build the most secure, scalable, and broadest network. It may sound cliché or overly idyllic, but we are wholehearted: A world utility cannot succeed in the hands of a few. Aptos will only succeed when everyone collectively uses, owns and maintains it .”

Modified Byzantine Consensus?

While blockchains have come a long way, they are still fraught with danger and confusion. It’s no secret that the decentralization movement is working hard to achieve escape velocity, such as a batch of new public chains that have emerged, and the Ethereum Layer 2 network under construction, but the limitations of reliability, security, and availability remain Blocks the widespread adoption of blockchain by groups ranging from ordinary individuals to large corporations.

“Billions of internet users are already familiar with purchasing goods and services and exchanging or storing value. To accelerate the adoption of web3 by internet users, we need to make some progress at the blockchain level to further strengthen other parts of the web3 stack.” Another co-founder of Aptos, Avery Ching, pointed out.

When Libra first released the project white paper, it clearly used a variant of the Byzantine Fault Tolerance (BFT) consensus protocol, the HotStuff protocol. The Aptos team said that in the past three years, the team has iterated the protocol to the fourth generation.

The synchronization of the BFT consensus protocol has always been a difficulty. We assume that the messages in the network can arrive within a known time Δ, but it is difficult to guarantee this assumption in actual engineering practice. Therefore, in the semi-synchronized state, each replica node must maintain a timer, and once the timeout expires, the protocol will be triggered to elect a new leader node. When the BFT consensus protocol was first deployed, the Aptos team added an active pacemaker, which directly uses a timeout to synchronize validating nodes, much faster than waiting for the timeout to keep increasing. Through the improvement of the protocol, submitting a block only needs to go back and forth between the two networks, and the finality routine can reach within one second.

Aptos has a novel reputation system to analyze on-chain state and automatically update leader node rotation to adjust unresponsive validator nodes without any human intervention, making it well suited for decentralized environments. Furthermore, the protocol explicitly separates effectiveness and safety. Regardless of whether the network is unreachable, or the non-secure core is compromised in some way, as long as the integrity guarantees of BFT are adhered to, the chain will not fork. Aptos said that this security has been audited and verified.

“Our research shows that there can be simple extensions that increase the robustness of our current BFT protocol by taking advantage of the period in which more than 2f+1 nodes vote on a block. These stronger commitments provide more confidence , i.e. the network will not fork in the future, and further increase users’ confidence in finality beyond standard BFT properties.”

Modularized code, Move Move Move

The Move language is a programming language originally designed specifically for the Libra blockchain, and Aptos continues to use this language. When Kyle Saman, a partner of Multicoin Capital, wrote an article explaining why he invested in Aptos, he directly used “Move Move Move” as the title. Avery Ching said the team “loves the security of Move and is excited to help build a broad ecosystem around it”, and Kyle Saman also noted that Move is comparable in efficiency to Rust and C++ while offering many security features, It can reduce the vulnerability of smart contract writing, and I believe that Move will become one of the most important development environments in the coming year.

Avery Ching said the Move language was designed for secure resource management and verifiable execution on the blockchain. Execution of transactions is certain, sealed and measured. Deterministic and hermetic means that the output of transaction execution is completely predictable and based only on the information contained in the transaction and the current ledger state. Metering is an important defense against denial-of-service attacks at the transaction execution level. Move validators can formally verify the properties of Move modules via an expressive specification language fast enough to run as part of continuous integration testing. Move resources, inspired by linear types, statically ensure that resources are preserved and not copied or accidentally destroyed – completely avoiding a class of potential attacks, Aptos accounts, transaction fees, standard library, validator node management and configuration All implemented with Move.

Avery Ching further pointed out that a large number of validator nodes operating in different environments is important for both decentralization and security. Additionally, the software for validating nodes must be securely designed to prevent attacks — which is one of the main reasons why the project chose Rust and Move as the languages ​​to implement the protocol and smart contract logic. The core security properties of the Aptos blockchain depend on the correct implementation of validator nodes, Move modules, and Move VMs. The team has modularized the code and identified a minimal Trusted Computing Base (TCB) to isolate key security properties (e.g., the security of the consensus protocol and correctness of execution). The use of modularity and TCB enables teams to take advantage of high-assurance development techniques. Additionally, for security-critical components, teams can enforce stricter requirements on the security of code reviews, dependencies, testing, and execution environments.

Performance: 130,000 TPS achieved in 32 core execution benchmarks

Over the years, there has been a huge debate around blockchain performance metrics, with both throughput (tps) and finality being influenced by numerous factors. The first is the complexity of the transaction. A simple peer-to-peer transaction to transfer tokens from Bob’s account to Alice’s account costs much less than the transaction cost of pre-generating 1000 NFTs. The size and distribution of validator nodes and account universes can also have a big impact on performance. How many validator nodes are there and what are their hardware specifications? How many accounts are there (eg, 10,000 vs. 1 billion accounts)? What is the distribution of account sizes (eg 1k vs 10 MB)? What is the distribution of access patterns (eg, from zero transaction conflicts to fully sequential dependencies)? What are the ordering guarantees between transactions (partial ordering or total ordering) and are they a good fit for smart contracts? These factors make it difficult to objectively compare the performance of different networks. The comparison is even more confusing when one considers the different ways of measuring performance.

Avery Ching said that based on research on the DAG consensus protocol family, the throughput of such protocols can reach 125,000 to 160,000 tps or even higher. However, these throughput numbers do not represent end-to-end blockchain throughput, as they only factor in consensus, network, and partial storage. They do not take into account other important factors such as transaction execution time, account access patterns, or authentication data structures (such as Merkle trees) in the production blockchain. When it comes to finality, block time is often mistaken for finality, when in fact, block time is only an input to finality.

The Aptos team will share benchmarking frameworks in the future and compare the performance characteristics of various use cases across different blockchains. An important step in this is the complete decoupling of the consensus protocol from transaction execution. The consensus protocol agrees on the proposed transaction ordering, outside of a separate protocol and critical path, the validators execute the transactions and agree on the final transaction ordering and execution results. Protocols that integrate consensus and execution are simpler, but pay a higher price in throughput and latency due to their interdependence.

After the two are decoupled, the next bottleneck is transaction execution time. The Aptos team says it has achieved over 130,000 transactions per second in an execution benchmark with only 32 cores, a benefit that developers get for free as the execution framework automatically takes advantage of the parallelism inherent in any workload. No additional information is required. Perhaps most importantly, a high-throughput blockchain means low transaction fees for users.

The final bottleneck in performance — and the most overlooked — is the authentication data structure and associated state store. One challenge is that in-memory Merkle trees are efficient at small scales when authenticating ledger state (such as account balances, smart contracts, etc.), however, writing large Merkle trees to persistent storage is a bottleneck. The Aptos team says it is designing its own authentication data structure to make it database-friendly by exploring higher branching factors, access-pattern-optimized caching, and more careful version management. The team is also currently developing support for large accounts – being able to access separate Move resources for each account, rather than inefficiently as a single blob, and looking into future paths by splitting them into chunks optimized for access patterns , enabling finer-grained access in Move resources, tiered storage and state leasing are important priorities to support all types of new use cases.

Mainnet and ecological construction: The mainnet is planned to be launched in the third quarter

Although this is the first round of financing, Aptos cannot simply be regarded as a start-up project in terms of technical reserves or ecological cooperation. Mo Shaikh said, “Over the past few months, we have discussed their projects with hundreds of developers, brands and companies; there is a lot to be excited about. They are building the next generation of social media platforms, rich NFT experiences , Web3 gaming, creator-first media and entertainment businesses, cheap and secure payments, DeFi-integrated fintech offerings, and more”, like Anchorage, Binance, Blockorus, Coinbase, Livepeer, Moonclave, Paxos, Paymagic, Rarible, and Good companies like Streaming Fast are already providing feedback in the Aptos community and contributing code on Devnet.

The construction schedule of Aptos is as follows:

  • Q1 2022 (March 15) – Developer testnet launch

Collaborate with strategic partners and the web3 developer community to gather feedback and make improvements to the Move developer experience and the Move language

  • Q2 2022 – Incentivized testnet launch

Build a larger, mainnet-like testing ground for strategic partners and web3 developers

Collaborate with the node operator community to join and build a co-operating decentralized network

Launch a bug bounty program to improve developer experience, node operations, and resolve infrastructure issues

Incentives for all participants who help keep the network safe

  • Q3 2022 – Mainnet launch
  • Q4 2022 – Q1 2023 – Deploy the next major release on the Aptos mainnet and deliver the next set of important features

Avery Ching concluded by writing: We are committed to launching our mainnet as soon as possible so that the wider community can start building – many have been waiting for our technology to roll out for years. On the way to mainnet, we will be announcing a series of global hacking events and look forward to your participation. We will leverage our ability to upgrade the network with new features, functionality, security, and scaling improvements once an enterprise-level response is achieved during our testnet and deployment.

Posted by:CoinYuppie,Reprinted with attribution to:https://coinyuppie.com/understanding-the-new-public-chain-aptos-the-exodus-of-the-original-diem-team-rebuilding-the-public-chain/
Coinyuppie is an open information publishing platform, all information provided is not related to the views and positions of coinyuppie, and does not constitute any investment and financial advice. Users are expected to carefully screen and prevent risks.

Like (0)
Donate Buy me a coffee Buy me a coffee
Previous 2022-03-16 09:58
Next 2022-03-16 09:59

Related articles