Skip to main content

API Management

Agent Based

An agent in API management is an autonomous or semi-autonomous program that:

  • Interprets policies or rules
  • Monitors API usage
  • Takes actions (e.g., throttling, caching, logging, alerting)
  • May use AI/ML or decision trees to optimize or modify API behavior

Agents typically reside in:

  • API gateways (like Kong, Apigee, or AWS API Gateway)
  • Middleware platforms
  • Sidecars in service meshes (e.g., Envoy in Istio)
  • Client SDKs (for offline handling, caching, retries)

Proxy Based

A proxy is an intermediary between clients and servers. In API management, a reverse proxy is used to:

  • Accept client requests
  • Apply policies (security, rate limits, etc.)
  • Forward valid requests to the backend services
  • Log, monitor, and transform traffic as needed

A proxy-based API manager often acts like a traffic cop, enforcing rules before requests hit the actual API services.

Common Proxy-based API Gateways

ToolTypeFeatures
KongOSS & CloudPlugins, rate limiting, auth, logging
ApigeeCloudPolicies, analytics, monetization
AWS API GatewayCloudLambda auth, usage plans, CORS
Nginx/OpenRestySelf-hostedScripting, reverse proxy, caching
TykOSS & CloudJWT, auth, quotas, caching

Proxy vs Agent-Based API Management

FeatureProxy-Based API ManagementAgent-Based API Management
Central control point✅ Yes⚠️ Not always centralized
Requires client support❌ No✅ Sometimes (e.g. SDK agent)
Latency impact✅ Slight (adds hop)✅ or ❌ (depends on location)
Extensibility✅ via plugins✅ via scripts or AI logic
Offline behavior❌ Cannot handle✅ Client-side agent can