API Gateway 10.11 | Administering API Gateway | Deployment | Concepts | API Gateway Components
 
API Gateway Components
API Gateway consists of the following components:
The following table lists the purpose and characteristics of the API Gateway components:
Component
Purpose
Characteristics
Integration Server
Policy Enforcement Engine
*Stateless engine
*Two or more instances can operate as a cluster
*Supports vertical and horizontal scaling
*It runs an embedded Apache Tomcat server to host API Gateway console UI
Elasticsearch
Persistence layer for API configurations and API analytics
*Stateful persistence
*Externalized Elasticsearch is supported
*Needs quorum (odd number of instances) to operate in a cluster
Note:
Best practice is to separate the API analytics in a dedicated cluster.
*Supports horizontal scaling and data- sharing
Kibana
Dashboard for the API analytics
*Stateless engine
*Externalized Kibana is supported
*Two instances can operate in a cluster
*Scaling more than two is generally not required
Cluster Coordinator
In-Memory Distributed Caching
*Stateful engine for cluster coordination and service result caching
*Two options:
Ignite
Terracotta Server Array (TSA)
*Ignite is embedded and scales as you scale the Integration Server.
*TSA operates in an active-passive setup, typically, two instances are sufficient. Scaling more than two is not required and an anti-pattern .
Note:
All the components can run in containers.
For more information on the supported versions of external Elasticsearch, Kibana, and Terracotta Server Array, see API Gateway, Elasticsearch, Kibana, and TSA Compatibility Matrix.