The Dubbo Kubernetes Integration
⚠️ This is still an experimental version. ⚠️
The universal Control Plane and Console for managing microservices on any environment - VM and Kubernetes.
Quick Start (under development)
NOTICE: As the project has not been officially released yet, the following commands may not run properly. The best way for now is to refer to the Developer's Guide to learn how to download the source code and build it locally!
-
Download dubbo-control-plane
binary package.
curl -L https://raw.githubusercontent.com/apache/dubbo-kubernetes/master/release/download/dubbo.sh | sh -
cd dubbo-$version
export PATH=$PWD/bin:$PATH
-
Install control-plane
on Kubernetes
dubboctl install --profile=demo
-
Check installation
kubectl get services -n dubbo-system
-
Next, deploy Dubbo applications to Kubernetes as shown below:
apiVersion: v1
kind: Service
metadata:
name: demo-service
namespace: dubbo-demo
spec:
selector:
app: dubbo-demo
type: ClusterIP
ports:
- name: port1
protocol: http
port: 80
targetPort: 8080
---
apiVersion: apps/v1
kind: Deployment
metadata:
name: example-app
namespace: dubbo-demo
spec:
...
template:
metadata:
...
labels:
app: dubbo-demo
dubbo.apache.org/service: dubbo-demo
spec:
containers:
...
If you want to create your own Dubbo application from scratch and deploy it, please use dubboctl we provided below.
-
Open the following page to check deployment status on control plane UI:
kubectl port-forward svc/dubbo-control-plane \
-n dubbo-system 5681:5681
visit, 127.0.0.1:5681/admin
Architecture
The microservcice architecture built with Dubbo Control Plane consists of two main components:
- The
Dubbo Control Plane
configures the data plane - applications developed with Dubbo SDK, for handling service traffic. Users create policies that the dubbo control plane processes to generate configurations for the data plane.
- The data plane - the
Dubbo SDK
, connets directly to control plane and receives configurations that can work as the sources for service discovery, traffic routing, load balancing, etc.
Dubbo Control Plane supports two deployment modes: kubernetes
and universal
.
kubernetes
mode is like the classic Service Mesh architecture, with all microservices concepts bonded to kubernetes resources. Unlike classic service mesh solutions like istio, Dubbo favors a proxyless data plane deployment - with no envoy sidecar.
universal
is the traditional microservice architecture that all Dubbo users are already familiar with. Unlike the kubernetes mode, it usually needs a dedicated registry like Nacos or Zookeeper for service discovery, etc.
Kubernetes
In kubernetes mode, the control plane will interact directly with the Kubernetes API-SERVER, watching the kubernetes resources and transform them as xDS resources for service discovery and traffic management configurations.
We all know the service
definitions of Kubernetes and Dubo are different, Kubernetes Service
is more like an application concept run on a selected group of pods while Dubbo Service
can mean a specific RPC service inside the application process. So how does dubbo control plane manages to bridge the interface-application
gap, check here for more details.
Universal
In Universal mode, Dubbo still uses Nacos or Zookeeper as registries for service discovery, control plane then interact with registry directly to work as the console UI, as the entry point for viewing and managing the cluster.
Multiple clusters
Dubbo Control Plane supports running your services in multiple zones. It is even possible to run with a mix of Kubernetes and Universal zones. Your microservice environment can include multiple isolated services, and workloads running in different regions, on different clouds, or in different datacenters. A zone can be a Kubernetes cluster, a VPC, or any other deployment you need to include in the same distributed microservice environment. The only condition is that all the data planes running within the zone must be able to connect to the other data planes in this same zone.
Dubbo Control Plane supports a global
deployment mode that can connect different zone
region clusters. The picture below shows how it works.
Roadmap
- Security
- Metrics
- Cross-cluster communication
- Console
Refereces
- Dubboctl
- Console UI Design
- Dubbo java xDS implementation
- Dubbo go xDS implementation
竞赛文档:面向云原生的下一代微服务集群监测机制,涵盖Kubernetes、Nacos等
- 设计文档
- 用户手册