aigw-fundamentals

Configuring an Envoy AI Gateway requires applying CRDs including AIGatewayRoute, AIServiceBackend, BackendSecurityPolicy, GatewayConfig, MCPRoute, and QuotaPolicy. These resources define the routing hierarchy, backend mappings, and security policies for your AI deployment.

Securing AI service backends in Kubernetes involves applying the BackendSecurityPolicy CRD within your Envoy AI Gateway architecture. This policy manages provider authentication by defining TLS validation rules and API key integration for secure backend connections.

You can route traffic to specific AI models by defining AIGatewayRoute resources within the Envoy AI Gateway. This CRD maps incoming requests to appropriate AIServiceBackend resources, enabling model-specific routing across standard OpenAI-compatible providers.

Mapping the Envoy AI Gateway resource hierarchy involves linking AIGatewayRoute definitions to AIServiceBackend resources. This structure dictates how requests traverse from the gateway entry point to specific AI models across various configured providers.

Designing an AI gateway architecture with Envoy involves creating core CRDs to outline route and backend configurations. This approach ensures clear documentation of resource types, integration points with Gateway API components, and security considerations.

Envoy AI Gateway supports standard OpenAI-compatible backends through the AIServiceBackend CRD. This allows teams to integrate various AI providers into their cluster while maintaining consistent routing and security policies across the gateway.