A company is evaluating AI security architecture for a new generative AI application. The security architect needs to implement a secure AI gateway pattern. Which architectural component sits between client applications and AI model endpoints to provide centralized security enforcement including rate limiting, authentication, content filtering, and logging?
Select an answer to reveal the explanation.
Short Explanation and Infographic
Here's the deal — b is correct because Azure API Management configured as an AI gateway provides centralized security enforcement including authentication/authorization via validate-jwt policies, rate limiting and quota management, request/response logging, content transformation, and custom policy execution. Microsoft has also published AI-specific APIM policies including Azure OpenAI token rate limiting and semantic caching.
Full explanation below image
Full Explanation
B is correct because Azure API Management configured as an AI gateway provides centralized security enforcement including authentication/authorization via validate-jwt policies, rate limiting and quota management, request/response logging, content transformation, and custom policy execution. Microsoft has also published AI-specific APIM policies including Azure OpenAI token rate limiting and semantic caching. A is incorrect because Azure Application Gateway with WAF provides L7 web application firewall protection against web attacks but does not provide API management capabilities like authentication policy enforcement, rate limiting per API key/user, or AI-specific content inspection. C is incorrect because Azure Front Door provides CDN and global load balancing but lacks the API management policy framework needed for comprehensive AI security controls. D is incorrect because Azure Load Balancer operates at L4 and provides no security policy enforcement capabilities.