In the world of blockchain, oracle machines are essential for smart contracts to access real-world data. Actually, I think this is something many people underestimate.

Smart contracts are very strong when it comes to data inside the blockchain, but they cannot access external world information at all. For example, if you want to reflect today’s crude oil trading volume in a contract, that information is off-chain, so some kind of intermediary is needed. That’s the role of oracle machines.

The mechanism of these oracle machines is actually a simple request-response system. The smart contract sends out a request saying “Please provide this data,” and the oracle picks it up, retrieves real-world data, and returns it to the chain. This entire process is the basic operation of an oracle machine.

In the blockchain industry, the project with the overwhelming market share is Chainlink. Chainlink is a decentralized oracle machine, specialized in providing real-world data to blockchains in the safest way possible. It builds an ecosystem centered around the LINK token, with incentive design.

To operate Chainlink’s oracle machine, transferring LINK tokens is required. These LINK tokens are implemented using the ERC677 standard on the Ethereum network. This is an important point. Standard ERC20 tokens cannot perform payments and service requests simultaneously. That’s why Chainlink proposed its own ERC677.

The core method of ERC677 is “transferAndCall.” This allows token transfer and service request to be completed in a single transaction. When a user executes transferAndCall, in addition to the ERC20 transfer, the onTokenTransfer method of the recipient address is automatically called.

When the oracle machine receives a request, it first performs multiple security checks. Is the sender really the LINK token contract? Is the data size appropriate? Is the function selector correct? If these validations pass, a unique ID called requestId is generated and recorded in the event log.

This event is monitored by off-chain nodes. The nodes read the necessary information from the event and retrieve real-world data via network APIs. Then, they call the fulfillOracleRequest method to return the data to the chain. This flow constitutes the complete cycle of an oracle machine.

In terms of implementation, Chainlink already provides price feeds for existing trading pairs, called Aggregators. Using these, developers don’t need to specify their own APIs. Major pairs like ETH/USD are already available, and simply calling the latestRoundData() method retrieves the latest price.

For most applications, just reading the latest price data is sufficient. Especially when dealing with USD-denominated prices, the precision is uniformly set to 8 decimal places, so no complex processing is needed. In other words, implementing price feeds with oracle machines is easier than you might think.

Because of this system, DeFi protocols and complex financial products can be built. Without oracle machines, bridging the blockchain and the real world would be impossible. You can also see various projects utilizing such oracles on Gate.io, so if you’re interested, it’s worth checking them out.