Many people think that Layer2 security verification is done in real-time, but that's not the case. Take Plasma as an example—fraud proofs are not just an optional feature but the core of the entire security design.

The workflow is actually quite straightforward. The side chain first generates blocks, and operators submit block headers to the mainnet. But here’s a key point—the mainnet does not immediately verify the authenticity of transactions; it simply records Plasma’s commitment to the state at that moment. The real contest happens later.

When a user initiates an asset exit, the fraud proof mechanism is truly activated. The exiting party needs to provide proof: that the asset indeed exists within a specific Plasma block. Then it enters the so-called challenge period—that’s Plasma’s most clever design.

During the challenge period, any observer can submit a fraud proof. For example, pointing out that the exit request involves double-spending, that the asset has been spent in a subsequent transaction, or that the state itself is invalid. Plasma’s logic is interesting: it doesn’t require you to prove everything upfront, but rather to prove "where the problem is" afterward.

The beauty of this time window lies in game theory. Plasma’s assumption is that as long as the challenge period is long enough, someone will step forward to point out issues either for self-protection or to gain incentives. Once the window closes, if no one challenges, the mainnet recognizes Plasma’s commitment as valid, and the exit is finalized.

So you see, Layer2 security is not built on real-time calculations but is "stretched over the dimension of time." This design is both radical and pragmatic, and it’s precisely what makes Plasma worth pondering.