<aside> 📘 TL;DR;
智能合约需要满足严格确定性（deterministic），这导致其无法和 off-chain 交互，因为 off-chain 数据是无法满足确定性的。
所以引入新的组件 Oracle，该组件扮演 on-chain/off-chain 数据交互的桥梁，可以将 off-chain 数据固化到链上，也可以将链上数据推送到外部。Oracle 所面临的挑战为真实性、可用性和可追溯性。
</aside>
Oracles are data feeds that make off-chain data sources available to the blockchain for smart contracts. This is necessary because Ethereum-based smart contracts cannot, by default, access information stored outside the blockchain network.
Giving smart contracts the ability to execute using off-chain data extends the utility and value of decentralized applications. For instance, on-chain prediction markets rely on oracles to provide information about outcomes that they use to validate user predictions. Suppose Alice bets 20 ETH on who will become the next U.S. President. In that case, the prediction-market dapp needs an oracle to confirm election results and determine if Alice is eligible for a payout.
This page assumes the reader is familiar with Ethereum fundamentals, including nodes, consensus mechanisms, and the EVM. You should also have a good grasp of smart contracts and smart contract anatomy, especially events.
Oracles are applications that source, verify, and transmit external information (i.e. information stored off-chain) to smart contracts running on the blockchain. Besides “pulling” off-chain data and broadcasting it on Ethereum, oracles can also “push” information from the blockchain to external systems, e.g., unlocking a smart lock once the user sends a fee via an Ethereum transaction.
Without an oracle, a smart contract would be limited entirely to on-chain data.
Oracles differ based on the source of data (one or multiple sources), trust models (centralized or decentralized), and system architecture (immediate-read, publish-subscribe, and request-response). We can also distinguish between oracles based on whether they retrieve external data for use by on-chain contracts (input oracles), send information from the blockchain to the off-chain applications (output oracles), or perform computational tasks off-chain (computational oracles).
Many developers see smart contracts as code running at specific addresses on the blockchain. However, a more general view of smart contracts is that they are self-executing software programs capable of enforcing agreements between parties once specific conditions are met - hence the term “smart contracts.”
But using smart contracts to enforce agreements between people isn't straightforward, given that Ethereum is deterministic. A deterministic system(opens in a new tab) is one that always produces the same results given an initial state and a particular input, meaning there is no randomness or variation in the process of computing outputs from inputs.
To achieve deterministic execution, blockchains limit nodes to reaching consensus on simple binary (true/false) questions using only data stored on the blockchain itself. Examples of such questions include:
If blockchains received information from external sources (i.e. from the real world), determinism would be impossible to achieve, preventing nodes from agreeing on the validity of changes to the blockchain’s state. Take for example a smart contract that executes a transaction based on the current ETH-USD exchange rate obtained from a traditional price API. This figure is likely to change frequently (not to mention that the API could get deprecated or hacked), meaning nodes executing the same contract code would arrive at different results.
For a public blockchain like Ethereum, with thousands of nodes around the world processing transactions, determinism is critical. With no central authority serving as a source of truth, nodes need mechanisms for arriving at the same state after applying the same transactions. A case whereby node A executes a smart contract’s code and gets "3" as a result, while node B gets "7" after running the same transaction would cause consensus to break down and eliminate Ethereum’s value as a decentralized computing platform.