Analysis of Anomaly Detection Techniques Applied to Web API Network Scenario

Web API networks comprise a multitude of API services, and compared to traditional network environments, they possess unique security requirements and face distinct threats. This article delves into the security characteristics of Web API networks and the particular threats they encounter, establishing evaluation criteria for anomaly detection models tailored to Web API networks and their associated scenarios. Through an integrated assessment of numerous studies, we meticulously compare and analyze various models in terms of false positive rates, robustness, detection latency, and their adaptability and detection capabilities. The findings reveal that while many models excel in detecting singular attacks within the Web API network context, their scope of attack detection is often constrained, making it challenging to comprehensively address the myriad threats Web API networks confront. Moreover, several unresolved challenges persist in the research, such as the scarcity of real-world datasets, inadequate model generalization capabilities, and subpar robustness.