Zero-Stale Protocol: Never Trust Old Data

When refreshing prices, ZISO prioritizes one rule: "Never present stale data as new."

If the latest price point hasn't been retrieved, the system would rather wait briefly than risk displaying a previous cache cycle as the current state.

What Problem Does It Solve?

• Ghost Data: Prevents edge or browser caches from misrepresenting old prices as "live."
• Timestamp Lag: Ensures that even if the page is open for a long period, you aren't looking at data from an hour ago without a warning.
• Network Jitter: Prevents "flickering" where old data makes the signal look valid when it has actually expired.

How It Works

1. Cache Bypassing: Price requests are appended with unique timestamp parameters to bypass aggressive edge caches.
2. Explicit Headers: API responses explicitly return no-store / no-cache headers to minimize middle-layer pollution.
3. Tiered Processing:
   • Price Pipeline: Returns the raw, fastest database state without overhead.
   • Decision Pipeline: Maintains a short server-side cache for heavy AI reasoning to prevent page weight issues.
4. Validation: If the returned data timestamp doesn't match the browser's current reality window, the UI will treat it as "Loading/Missing" rather than forcing the display of a stale result.

What You Will Experience

• Higher Fidelity: Prices are far less likely to get "stuck" during market hours.
• Honest Loading: If data is delayed, you'll see a loading state or a refresh prompt rather than a fake "Live" status.

Our Design Principles

• Wait for truth, don't mask with stale data.
• Fidelity over vanity "instant" loads.
• Segmented rendering for real-time accuracy.

ZISO AI: AI does the research. You keep the decision.