A signal delivery model defines how contextual facts reach authorization decisions. It covers embedded tokens, live queries, pushed event streams, local agents, and hybrid approaches, alongside freshness requirements. section 12.0
Design the model across three related choices:
- Select relevant signals, such as employment status, device trust, multifactor authentication, geolocation, risk score, or account balance. section 12.0
- Choose delivery through tokens, API queries, event streams, local agents, or hybrids such as device-posture tokens checked at request time. section 12.0
- Set freshness through live queries, brief caching, or token expiry, balancing assurance, latency, caching, and revocation. section 12.0
Different risk, different delivery
A low-risk collaboration feature may accept token-based context lasting an hour, while a production deployment decision may require live operational data and short-lived device-posture signals. section 12.0
