A fundamental dilemma of blockchain is right here: no matter how sophisticated smart contracts are written, they can only operate within their own limited on-chain domain. Real-world events—commodity price fluctuations, contract signings, sports results—all of these are black boxes to them.

This is no small issue. Early DeFi participants suffered many losses because data delays or errors led directly to liquidations, costing real money. Later, everyone realized that without a trustworthy external data source, so-called decentralization is at best a semi-finished product.

Some people saw an opportunity from this predicament. These individuals have special backgrounds—they are developers who have been tinkering in blockchain for years, witnessing firsthand how many system crashes are caused by unreliable data. Rather than being marketing geniuses, they are thorough engineers. Before officially launching products, they kept asking themselves tough questions: How to verify the authenticity of external data? How to ensure speed and cost-effectiveness without compromising security?

In the early stages, it was quiet and arduous. No explosive news, just low-key development. Turning ideas into prototypes, then tearing them down and rebuilding if they didn't work. Off-chain processing is fast but trust costs are high; on-chain verification is reliable but transaction fees are expensive. They gradually explored a balance through repeated trade-offs.

What truly set them apart from competitors was their adoption of a "both-and" strategy. Instead of choosing between "push" and "pull" data models, they supported both. High real-time requirements used push mechanisms, while projects aiming to control costs used on-demand pull. This isn't just bragging; during actual development, engineers found that both approaches serve real user needs in different scenarios.