megaease/easegress
Easegress
Easegress est une API gateway cloud-native, un contrôleur d'ingress Kubernetes et un plan de contrôle de service mesh. Il sert de point d'entrée centralisé pour gérer le trafic réseau, fournissant un plan de contrôle dédié et un plan de données sidecar pour gérer le cycle de vie et l'interception des requêtes entre les services distribués.
Le projet fonctionne comme une passerelle de modèles d'IA qui relaie les requêtes vers des fournisseurs de grands modèles de langage (LLM) et traduit les formats de requête pour assurer la compatibilité. C'est un proxy extensible qui permet une gestion personnalisée des requêtes et des réponses via un pipeline de filtres implémenté en WebAssembly.
Le système couvre de larges capacités de gestion de trafic, incluant le routage multi-protocole pour HTTP, MQTT et WebSockets, ainsi que l'équilibrage de charge et la découverte de services. Il implémente des protections de stabilité telles que des disjoncteurs (circuit breakers) et des limites de débit (rate limits), et prend en charge des stratégies de déploiement comme les releases canary et blue-green.
L'observabilité est intégrée via OpenTelemetry pour le traçage distribué des requêtes et la surveillance des performances du trafic. La sécurité est gérée par la validation de jetons et de signatures ainsi que par la gestion des certificats SSL/TLS.
Features
- Ingress Controllers - Serves as a Kubernetes ingress controller to manage external access and route traffic into the cluster.
- Centralized API Gateways - Provides a centralized entry point for routing incoming network requests to various backend microservices.
- AI Gateways - Provides a unified interface for routing requests to multiple AI model providers while maintaining format compatibility.
- Control and Data Plane Separation - Separates network configuration and management in the control plane from the packet forwarding in the data plane.
- Ingress Gateways - Implements specialized ingress gateways for managing external entry points into Kubernetes clusters.
- Proxy Filters - Supports custom request and response handling through a filter pipeline implemented via WebAssembly code.
- Service Mesh Control Planes - Implements a centralized management system for configuring and orchestrating service-to-service network policies.
- Service Meshes - Provides an infrastructure layer to manage service-to-service communication, observability, and security.
- Dynamic Service Discovery - Maps logical service names to physical network addresses in real-time to enable dynamic microservices communication.
- Model Request Proxies - Acts as a gateway that intercepts and routes API requests to various large language model providers.
- System Stability Protections - Prevents cascading failures by implementing rate limiting, circuit breakers, and execution timeouts.
- Content-Aware Routing Rules - Implements routing rules that inspect request headers, paths, and IP addresses to direct traffic.
- Request Token Validators - Validates tokens and digital signatures on every request to enforce access control and security permissions.
- API Access Security - Secures API access through token validation, digital signatures, and stability mechanisms like rate limiting.
- Multi-Protocol Proxies - Functions as a network intermediary supporting multiple protocols including HTTP, MQTT, and WebSockets.
- Request Processing Filters - Executes custom request and response processing logic using portable WebAssembly binary modules.
- Deployment Strategies - Supports various deployment patterns including blue-green and canary releases to manage application rollout.
- Canary Deployment Controllers - Implements controllers to manage incremental traffic shifts for safe canary deployments of API updates.
- Traffic Load Balancers - Distributes network traffic across multiple service instances using round-robin and sticky session policies.
- Multi-Protocol Traffic Orchestration - Routes and manages traffic across multiple protocols including HTTP, MQTT, and WebSockets.
- Zero-Downtime Deployments - Ensures continuous service availability during updates through health-check-gated traffic switching.
- Zero-Downtime Upgrades - Updates service binaries and configurations in place to maintain active network connections without downtime.
- API Orchestration - Provides capabilities to aggregate and manage concurrent requests across multiple external service APIs into single responses.
- Cloud Network Orchestration - Directs network data flow within cloud environments to ensure efficient connectivity between distributed components.
- Circuit Breakers - Prevents cascading failures by automatically stopping traffic to unhealthy backend services.
- Traffic Pipeline Filters - Implements WebAssembly-based custom filters within the traffic pipeline for request transformation and routing.
- SSL/TLS Certificate Management - Manages the lifecycle and deployment of SSL/TLS certificates to secure network traffic.
- Dynamic Service Discovery - Dynamically locates and tracks available backend service instances using external service registries.
- Configuration Hot-Reloading - Refreshes routing rules and system settings in place without requiring a full process restart.
- Distributed Tracing - Generates vendor-neutral distributed traces using unique identifiers to analyze request execution flow across services.
- Request Traffic Monitors - Tracks request volume, processing duration, and error rates to monitor the health of backend servers.
- OpenTelemetry - Integrates with OpenTelemetry to export trace spans to OTLP-compatible backends for distributed tracing.
- OpenTelemetry SDKs - Integrates OpenTelemetry SDKs to generate and export vendor-neutral observability data and distributed traces.
- Infrastructure et conteneurs - A cloud-native traffic orchestration and management system.
- API Gateways - Cloud-native traffic orchestration and gateway system.
- API Gateways and Proxies - Traffic orchestration system for cloud-native architectures.
