Two-site link · operations console
A simulated 24-hour day on the 10 km BB84 fiber link of the SEE 510 case study: a session every 15 minutes, a key store that an application draws from, and scripted drift, attacks, a fiber bend, and a fiber cut. The monitor sees only what the sites can measure.
time
—
session
—
error rate
—
detections
—
of the morning baseline
key store
—
application
—
Error rate (%) · dashed: abort threshold · dotted: alert level
Key per session (bits, after the authentication key is replaced)
Keys in the store at Site A (256-bit keys) · ✕: application refused
Authentication key pool (bits)
replay
current session
operator log
the day's events and when the monitor flagged them
| from | to | event | kind | what happens | flagged after |
|---|
A session is flagged when it is rejected, when its error rate passes the alert level, or when its detections fall below half of the morning baseline. The error rate alone cannot tell drift or noise from interception; the protocol treats every error as if an adversary caused it. Within this model, under the assumptions of 03_assumptions.md; not a proof of real-world security. Evidence: scenario 9.