Overview

Why We Need Ingresses

A Service is generally used to forward access requests based on TCP and UDP and provide layer-4 load balancing for clusters. However, in actual scenarios, if there is a large number of HTTP/HTTPS access requests on the application layer, the Service cannot meet the forwarding requirements. Therefore, the Kubernetes cluster provides an HTTP-based access mode, ingress.

An ingress is an independent resource in the Kubernetes cluster and defines rules for forwarding external access traffic. As shown in Figure 1, you can customize forwarding rules based on domain names and URLs to implement fine-grained distribution of access traffic.

Figure 1 Ingress diagram

The following describes the ingress-related definitions:

Ingress object: a set of access rules that forward requests to specified Services based on domain names or URLs. It can be added, deleted, modified, and queried by calling APIs.
Ingress Controller: an executor for request forwarding. It monitors the changes of resource objects such as ingresses, Services, endpoints, secrets (mainly TLS certificates and keys), nodes, and ConfigMaps in real time, parses rules defined by ingresses, and forwards requests to the target backend Services.

Ingress Controllers provided by different vendors are implemented in different ways. Based on the types of load balancers, Ingress Controllers are classified into LoadBalancer Ingress Controller and Nginx Ingress Controller. Both of them are supported in CCE. LoadBalancer Ingress Controller forwards traffic through ELB. Nginx Ingress Controller uses the templates and images maintained by the Kubernetes community to forward traffic through the Nginx component.

Ingress Feature Comparison

**Table 1** Comparison between ingress features
Feature	ELB Ingress Controller	Nginx Ingress Controller
O&M	O&M-free	Self-installation, upgrade, and maintenance
Performance	One ingress supports only one load balancer.	Multiple ingresses support one load balancer.
	Enterprise-grade load balancers are used to provide high performance and high availability. Service forwarding is not affected in upgrade and failure scenarios.	Performance varies depending on the resource configuration of pods.
	Dynamic loading is supported.	Processes must be reloaded for non-backend endpoint changes, which causes loss to persistent connections. Lua supports hot updates of endpoint changes. Processes must be reloaded for a Lua modification.
Component deployment	Deployed on the master node	Deployed on worker nodes, and operations costs required for the Nginx component
Route redirection	Supported	Supported
SSL configuration	Supported	Supported
Using ingress as a proxy for backend services	Supported	Supported, which can be implemented through backend-protocol: HTTPS annotations.

The LoadBalancer ingress is essentially different from the open source Nginx ingress. Therefore, their supported Service types are different. For details, see Services Supported by Ingresses.

LoadBalancer Ingress Controller is deployed on a master node. All policies and forwarding behaviors are configured on the ELB side. Load balancers outside the cluster can connect to nodes in the cluster only through the IP address of the VPC. Therefore, LoadBalancer ingresses support only NodePort Services.

Nginx Ingress Controller runs in a cluster and is exposed as a Service through NodePort. Traffic is forwarded to other Services in the cluster through Nginx-ingress. The traffic forwarding behavior and forwarding object are in the cluster. Therefore, both ClusterIP and NodePort Services are supported.

In conclusion, LoadBalancer ingresses use enterprise-grade load balancers to forward traffic and delivers high performance and stability. Nginx Ingress Controller is deployed on cluster nodes, which consumes cluster resources but has better configurability.

Working Rules of LoadBalancer Ingress Controller

LoadBalancer Ingress Controller developed by CCE implements layer-7 network access for the internet and intranet (in the same VPC) based on ELB and distributes access traffic to the corresponding Services using different URLs.

LoadBalancer Ingress Controller is deployed on the master node and bound to the load balancer in the VPC where the cluster resides. Different domain names, ports, and forwarding policies can be configured for the same load balancer (with the same IP address). Figure 2 shows the working rules of LoadBalancer Ingress Controller.

A user creates an ingress object and configures a traffic access rule in the ingress, including the load balancer, URL, SSL, and backend service port.
When Ingress Controller detects that the ingress object changes, it reconfigures the listener and backend server route on the ELB side according to the traffic access rule.
When a user accesses a workload, the traffic is forwarded to the corresponding backend service port based on the forwarding policy configured on ELB, and then forwarded to each associated workload through the Service.

Figure 2 Working rules of shared LoadBalancer ingresses in CCE standard clusters

Services Supported by Ingresses

Table 2 lists the services supported by LoadBalancer ingresses.

**Table 2** Services supported by LoadBalancer ingresses
Cluster Type	ELB Type	ClusterIP	NodePort
CCE standard cluster	Shared load balancer	Not supported	Supported
CCE standard cluster	Dedicated load balancer	Not supported (Failed to access the dedicated load balancers because no ENI is bound to the associated pod of the ClusterIP Service.)	Supported
CCE Turbo cluster	Shared load balancer	Not supported	Supported
CCE Turbo cluster	Dedicated load balancer	Supported	Not supported (Failed to access the dedicated load balancers because an ENI has been bound to the associated pod of the NodePort Service.)