An Automatic Discovery and Association Method for API Sensitive Fields Oriented to Data Security

With the increasing adoption of API-based architectures in business systems, sensitive data are frequently exchanged between systems through APIs. However, existing data classification and grading results at the database layer are difficult to directly map to API fields at the network layer. To address the challenges caused by inconsistent naming conventions and significant contextual differences between API fields and database fields, this paper proposes an automatic discovery and association method for API sensitive fields oriented to data security. The method uses Suricata to passively capture API runtime data and integrates database field metadata with a small number of samples. Multi-feature association is then performed based on field semantics, data characteristics, and instance-value matching, enabling database data classification and grading labels to be mapped to corresponding API fields. A simulated business system related to electric power dispatching is used as the experimental scenario. The results show that the proposed method can improve field association accuracy under complex naming conditions. It can be used to generate an inventory of sensitive API fields and provide support for subsequent monitoring of sensitive data flows.