The oracle problem

A blockchain can execute logic in a trust-minimized way, but contracts cannot fetch external data on their own, so they need an intermediary. That intermediary is the risk: if a single source or node supplies the data, it creates a centralized point that can be manipulated or fail, and a false value, such as a tampered price, can trigger improper liquidations. Good-practice mitigations are decentralized oracles (DON, Decentralized Oracle Networks) that aggregate multiple sources and independent nodes, median or weighted aggregation to resist outliers, economic incentives and penalties for misreporting, and signatures and proofs that let the data's origin be audited.

It is a foundational concept that defines how far an on-chain application can minimize trust. All the usefulness of DeFi, tokenized real-world assets, parametric insurance and derivatives depends on solving it. A contract is only as secure as its weakest oracle. A known vector is using an oracle based on the spot price of a single market, mitigated with time-weighted average prices (TWAP) and feeds aggregated across several markets.

Chainlink is the canonical decentralized oracle through its DON network. Pyth Network provides high-frequency price feeds with institutional publishers, and UMA works as a dispute-based optimistic oracle.