Chainlink 백서

Chainlink

원문

bitpro가 원문에서 글자만 모은 내용이에요. 표·그림·수식은 빠지거나 깨질 수 있어요. 전체 19쪽 중 1쪽

Chainlink 2.0: Next Steps in the Evolution of Decentralized Oracle Networks Lorenz Breidenbach1 Christian Cachin2 Benedict Chan1 Alex Coventry1 Steve Ellis1 Ari Juels3 Farinaz Koushanfar4 Andrew Miller5 Brendan Magauran1 Daniel Moroz6 Sergey Nazarov1 Alexandru Topliceanu1 Florian Tram` er7 Fan Zhang8 15 April 2021 v1.0 1Chainlink Labs 2The author is a faculty member at University of Bern. He co-authored this work in his separate capacity as an advisor to Chainlink Labs. 3The author is a faculty member at Cornell Tech. He co-authored this work in his separate capacity as Chief Scientist at Chainlink Labs. 4The author is a faculty member at University of California, San Diego. She co-authored this work in her separate capacity as an advisor to Chainlink Labs. 5The author is a faculty member at University of Illinois Urbana-Champaign. He co-authored this work in his separate capacity as an advisor on Chainlink. 6The author is a PhD candidate at Harvard University. He co-authored this work while on leave from Harvard, in his capacity as a researcher at Chainlink Labs. 7The author is a PhD candidate at Stanford University. He co-authored this work in his separate capacity as an advisor to Chainlink Labs. 8The author will be joining the faculty of Duke University in fall 2021. He co-authored this work in his current capacity as a researcher at Chainlink Labs. 1 Abstract In this whitepaper, we articulate a vision for the evolution of Chainlink be- yond its initial conception in the original Chainlink whitepaper. We foresee an increasingly expansive role for oracle networks, one in which they comple- ment and enhance existing and new blockchains by providing fast, reliable, and confidentiality-preserving universal connectivity and off-chain computation for smart contracts. The foundation of our plan is what we call Decentralized Oracle Networks, or DONs for short. A DON is a network maintained by a committee of Chainlink nodes. It supports any of an unlimited range of oracle functions chosen for deployment by the committee. A DON thus acts as a powerful abstraction layer, offering interfaces for smart contracts to extensive off-chain resources and highly efficient yet decentralized off-chain computing resources within the DON itself. With DONs as a springboard, Chainlink plans to focus on advances in seven key areas: • Hybrid smart contracts: Offering a powerful, general framework for aug- menting existing smart contract capabilities by securely composing on-chain and off-chain computing resources into what we call hybrid smart contracts. • Abstracting away complexity: Presenting developers and users with simple functionality eliminates the need for familiarity with complex underlying protocols and system boundaries. • Scaling: Ensuring that oracle services achieve the latencies and throughputs demanded by high-performance decentralized systems. • Confidentiality: Enabling next-generation systems that combine blockchains’ innate transparency with strong new confidentiality protections for sensitive data. • Order-fairness for transactions: Supporting transaction sequencing in ways that are fair for end users and prevent front-running and other attacks by bots and exploitative miners. • Trust-minimization: Creating a highly trustworthy layer of support for smart contracts and other oracle-dependent systems by means of decen- tralization, strong anchoring in high-security blockchains, cryptographic techniques, and cryptoeconomic guarantees. • Incentive-based (cryptoeconomic) security: Rigorously designing and ro- bustly deploying mechanisms that ensure nodes in DONs have strong eco- nomic incentives to behave reliably and correctly, even in the face of well- resourced adversaries. We present preliminary and ongoing innovations by the Chainlink community in each of these areas, providing a picture of the broadening and increasingly powerful capabilities planned for the Chainlink network. 2 Contents 1 Introduction 6 1.1 Decentralized Oracle Networks . . . . . . . . . . . . . . . . . . . . . . . 7 1.2 Seven Key Design Goals . . . . . . . . . . . . . . . . . . . . . . . . . . 8 1.3 Organization of this Paper . . . . . . . . . . . . . . . . . . . . . . . . . 19 2 Security Model and Goals 19 2.1 Current Architectural Model . . . . . . . . . . . . . . . . . . . . . . . . 20 2.2 Consensus Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . 20 2.3 Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 2.4 Note on Trust Models . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 3 Decentralized Oracle Network Interface and Capabilities 23 3.1 Networking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 3.2 Computation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 3.3 Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 3.4 Transaction-Execution Framework (TEF) . . . . . . . . . . . . . . . . . 27 3.5 Mempool Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 3.6 Stepping Stones: Existing Chainlink Capabilities . . . . . . . . . . . . . 29 3.6.1 Off-Chain Reporting (OCR) . . . . . . . . . . . . . . . . . . . . 29 3.6.2 DECO and Town Crier . . . . . . . . . . . . . . . . . . . . . . . 30 3.6.3 Existing On-Chain Chainlink Services . . . . . . . . . . . . . . . 31 3.6.4 Node Reputation / Performance History . . . . . . . . . . . . . 33 4 Decentralized Services Enabled by Decentralized Oracle Networks 34 4.1 Proof of Reserves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 4.2 Interfacing with Enterprise / Legacy Systems . . . . . . . . . . . . . . 35 4.3 Decentralized Identity . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 4.4 Priority Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 4.5 Confidentiality-Preserving DeFi / Mixicles . . . . . . . . . . . . . . . . 41 5 Fair Sequencing Services 43 5.1 The Front-Running Problem . . . . . . . . . . . . . . . . . . . . . . . . 45 5.1.1 Oracle Front-Running . . . . . . . . . . . . . . . . . . . . . . . 45 5.1.2 Front-Running User Transactions . . . . . . . . . . . . . . . . . 47 5.2 FSS Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 5.2.1 Transaction Processing . . . . . . . . . . . . . . . . . . . . . . . 48 5.2.2 Transaction Atomicity . . . . . . . . . . . . . . . . . . . . . . . 51 5.3 Fair Transaction Sequencing . . . . . . . . . . . . . . . . . . . . . . . . 53 5.4 Network-Layer Considerations . . . . . . . . . . . . . . . . . . . . . . . 56 5.5 Entity-Level Fairness Policies . . . . . . . . . . . . . . . . . . . . . . . 57 3 6 The DON Transaction-Execution Framework (DON-TEF) 57 6.1 TEF Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 6.2 Transaction Routing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 6.3 Syncing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 6.4 Reorgs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 7 Trust Minimization 65 7.1 Data-Source Authentication . . . . . . . . . . . . . . . . . . . . . . . . 65 7.1.1 The Limitations of Authenticated Data Origination . . . . . . . 67 7.1.2 Confidentiality . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 7.1.3 Combining Source Data . . . . . . . . . . . . . . . . . . . . . . 68 7.1.4 Processing Source Data . . . . . . . . . . . . . . . . . . . . . . . 69 7.2 DON Trust Minimization . . . . . . . . . . . . . . . . . . . . . . . . . . 70 7.2.1 Failover Clients . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 7.2.2 Minority Reports . . . . . . . . . . . . . . . . . . . . . . . . . . 70 7.3 Guard Rails . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 7.4 Trust-Minimized Governance . . . . . . . . . . . . . . . . . . . . . . . . 72 7.5 Public-Key Infrastructure . . . . . . . . . . . . . . . . . . . . . . . . . 73 8 DON Deployment Considerations 75 8.1 Rollout Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 8.2 Dynamic DON Membership . . . . . . . . . . . . . . . . . . . . . . . . 76 8.3 DON Accountability . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 9 Economics and Cryptoeconomics 78 9.1 Staking Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 9.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 9.3 Modeling Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 9.3.1 First-Tier Incentive Model: Rational Actors . . . . . . . . . . . 85 9.3.2 Second-Tier Adjudication Model: Correctness by Assumption . 85 9.3.3 Adversarial Model . . . . . . . . . . . . . . . . . . . . . . . . . 86 9.3.4 How Much Cryptoeconomic Security Is Enough? . . . . . . . . . 87 9.4 Staking Mechanism: Sketch . . . . . . . . . . . . . . . . . . . . . . . . 87 9.4.1 Further Mechanism Details . . . . . . . . . . . . . . . . . . . . . 89 9.4.2 Quadratic Staking Impact . . . . . . . . . . . . . . . . . . . . . 90 9.4.3 Realization of Second Tier . . . . . . . . . . . . . . . . . . . . . 91 9.4.4 Misreporting Insurance . . . . . . . . . . . . . . . . . . . . . . . 93 9.5 Single-Round Variant . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 9.6 Implicit-Incentive Framework (IIF) . . . . . . . . . . . . . . . . . . . . 95 9.6.1 Future Fee Opportunity . . . . . . . . . . . . . . . . . . . . . . 96 9.6.2 Speculative FFO . . . . . . . . . . . . . . . . . . . . . . . . . . 97 9.6.3 External Reputation . . . . . . . . . . . . . . . . . . . . . . . . 98 4 9.6.4 Open IIF Analytics . . . . . . . . . . . . . . . . . . . . . . . . . 98 9.7 Putting It All Together: Node Operator Incentives . . . . . . . . . . . 99 9.8 The Virtuous Cycle of Economic Security . . . . . . . . . . . . . . . . . 100 9.9 Additional Factors Driving Network Growth . . . . . . . . . . . . . . . 101 10 Conclusion 102 A Glossary 116 B DON Interface: Further Details 119 B.1 Networking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 B.1.1 Integrity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 B.1.2 Confidentiality . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 B.1.3 Availability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 B.2 Computation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 B.2.1 Trusted Execution Environments (TEEs) . . . . . . . . . . . . . 122 B.2.2 TEE Security . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 B.2.3 Secure Multi-Party Computation (MPC) . . . . . . . . . . . . . 124 B.3 Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 B.4 Resource Pricing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 C Adapter Examples 126 C.1 Oracle-Mediated Data-Source Access ( MediatedReport) . . . . . . . . . 126 C.2 Cross-Ledger Reports ( XL-Report-Read) . . . . . . . . . . . . . . . . . . 127 C.3 Confidential Switch ( ConfSwitch) . . . . . . . . . . . . . . . . . . . . . 128 D Functional Signatures 129 D.1 Functional Signatures for Combining Data . . . . . . . . . . . . . . . . 130 D.2 Discretized Functional Signatures . . . . . . . . . . . . . . . . . . . . . 130 E Prospective Bribery 134 F Random vs. Committee-Based Oracle Selection 135 5 1 Introduction Blockchain oracles are often viewed today as decentralized services with one objective: to forward data from off-chain resources onto blockchains. It’s a short step, though, from forwarding data to computing on it, storing it, or transmitting it bidirection- ally. This observation justifies a much broader notion of oracles’ functionality. So too do the growing service requirements of smart contracts and increasingly multifaceted technologies that rely on oracle networks. In short, an oracle can and will need to be a general-purpose, bidirectional, compute-enabled interface between and among on- chain and off-chain systems. Oracles’ role in the blockchain ecosystem is to enhance the performance, functionality, and interoperability of smart contracts so that they can bring new trust models and transparency to a multiplicity of industries. This transfor- mation will come about through broadening use of hybrid smart contracts, which fuse blockchains’ special properties with the unique capabilities of off-chain systems such as oracle networks and thereby achieve far greater reach and power than on-chain systems in isolation. In this whitepaper, we articulate a vision for what we call Chainlink 2.0, an evo- lution of Chainlink beyond its initial conception in the original Chainlink whitepa- per [98]. We foresee an increasingly expansive role for oracle networks, one in which they complement and enhance existing and new blockchains by providing fast, reli- able, and confidentiality-preserving universal connectivity and computation for hybrid smart contracts. We believe that oracle networks will even evolve to become utilities for exporting high-integrity blockchain-grade data to systems beyond the blockchain ecosystem. Today, Chainlink nodes run by a diverse set of entities come together in oracle net- works to relay data to smart contracts in what are known as reports. We can view such oracle nodes as a committee similar to that in a classical-consensus blockchain [72], but with the goal of supporting existing blockchains, rather than providing freestand- ing functionality. With verifiable random functions (VRF) and Off-Chain Reporting (OCR), Chainlink is already evolving toward a general-purpose framework and infras- tructure for providing the computational resources that smart contracts require for advanced functionality. The foundation of our plan for Chainlink 2.0 is what we call Decentralized Oracle Networks, or DONs for short. Since we introduced the term “oracle network” in the original Chainlink whitepaper [98], oracles have developed ever richer functionality and breadth of application. In this paper, we offer a fresh definition of the term according to our future vision for the Chainlink ecosystem. In this view, a DON is a network maintained by a committee of Chainlink nodes. Rooted in a consensus protocol, it supports any of an unlimited range of oracle functions chosen for deployment by the committee. A DON thus acts as a blockchain abstraction layer, providing interfaces to off-chain resources for both smart contracts and other systems. It also provides access to highly efficient yet decentralized off-chain computing resources. In general, a DON supports operations on a main chain . Its goal is to enable secure and flexi- 6 ble hybrid smart contracts, which combine on-chain and off-chain computation with connection to external resources. We emphasize that even with the use of committees in DONs, Chainlink itself remains inherently permissionless. DONs act as the foundation of a permissionless framework in which nodes can come together to implement custom oracle networks with their own regimes for node inclusion, which may be permissioned or permissionless. With DONs as a foundation, we plan to focus in Chainlink 2.0 on advances in seven key areas: hybrid smart contracts , abstracting away complexity , scaling, confidential- ity, order-fairness for transactions, trust minimization, and incentive-based (cryptoeco- nomic) security. In this paper introduction, we present an overview of Decentralized Oracle Networks in Section 1.1 and then our seven key areas of innovation in Sec- tion 1.2. We describe the organization of the rest of this paper in Section 1.3. 1.1 Decentralized Oracle Networks Decentralized Oracle Networks are designed to enhance and extend the capabilities of smart contracts on a target blockchain or main chain through functions that are not available natively. They do so by providing the three basic resources found in computing systems: networking, storage, and computation. A DON aims to offer these resources with strong confidentiality, integrity, and availability properties, 1 as well as accountability. DONs are formed by committees of oracle nodes that cooperate to fulfill a specific job or choose to establish a long-lived relationship in order to provide persistent services to clients. DONs are designed in a blockchain-agnostic way. They promise to serve as a powerful and flexible tool for application developers to create off-chain support for their smart contracts on any supported main chain. Two types of functionalities realize the capabilities of a DON: executables and adapters. Executables are programs that run continuously and in a decentralized man- ner on the DON. While they do not directly store main-chain assets, they have im- portant benefits, including high performance and the ability to perform confidential computation. Executables run autonomously on a DON and perform deterministic operations. They work in hand with adapters that link the DON to external resources and may be called by executables. Adapters, as we envision them for DONs, are a generalization of the external adapters in Chainlink today. While existing adapters typically only fetch data from data sources, adapters may operate bidirectionally; in DONs, they may additionally leverage joint computation by DON nodes to achieve additional features, such as encrypting reports for privacy-preserving consumption by an executable. To provide a sense of a DON’s basic operation, Fig. 1 shows conceptually how a DON might be used to send reports to a blockchain and thus achieve traditional, exist- ing oracle functionality. DONs can provide many additional features, however, beyond 1The “CIA triad” of information security [123, p. 26, §2.3.5]. 7 Chainlink’s existing networks. For example, within the general structure of Fig. 1, the executable could record fetched asset-price data on the DON, using such data to compute, e.g., a trailing average for its reports. Figure 1: Conceptual figure showing as an example how a Decentralized Oracle Net- work can realize basic oracle functionality, i.e., relay off-chain data to a contract. An executable uses adapters to fetch off-chain data, which it computes on, sending output over another adapter to a target blockchain. (Adapters are initiated by code in the DON, represented by small blue boxes; arrows show the direction of data flow for this particular example.) The executable can additionally read and write to local DON storage to keep state and/or communicate with other executables. Flexible network- ing, computation, and storage in DONs, all represented here, enable a host of novel applications. A major benefit of DONs is their ability tobootstrapnew blockchain services. DONs are a vehicle by which existing oracle networks can quickly stand up service applications that would today require the creation of purpose-built networks. We give a number of examples of such applications in Section 4. In Section 3, we provide more details on DONs, describing their capabilities in terms of the interface they present to developers and users. 1.2 Seven Key Design Goals Here we briefly review the seven key focuses enumerated above for the evolution of Chainlink, namely: 8 Hybrid smart contracts: Central to our vision for Chainlink is the idea of securely combining on-chain and off-chain components in smart contracts. We refer to contracts realizing this idea as hybrid smart contracts or hybrid contracts.2 Blockchains are and will continue to play two critical roles in decentralized-service ecosystems: They are both the loci where cryptocurrency ownership is represented and robust anchors for decentralized services. Smart contracts must therefore be repre- sented or executed on chain, but their on-chain capabilities are severely limited. Purely on-chain contract code is slow, expensive, and insular, unable to benefit from real-world data and a variety of functionalities that are inherently unachievable on chain, includ- ing various forms of confidential computation, generation of (pseudo)randomne
프로젝트로 돌아가기