Community Blog Container Service for Kubernetes (ACK): A Fully Managed Service on Alibaba Cloud

Container Service for Kubernetes (ACK): A Fully Managed Service on Alibaba Cloud

Alibaba Cloud Container Service for Kubernetes (ACK) is a fully-managed service compatible with Kubernetes to help users focus on their applications rather than managing container infrastructure.

What is the Container Service for Kubernetes?

Alibaba Cloud Container Service for Kubernetes (ACK) is one of the first services to participate in the Certified Kubernetes Conformance Program. ACK provides high-performance management services for containerized applications. You can manage enterprise-level containerized applications throughout the application lifecycle. This service allows you to run containerized applications in the cloud efficiently.

The Advantages of Container Service for Kubernetes

Cluster management

  • Provides three types of Kubernetes clusters: dedicated clusters, managed clusters, and serverless clusters.
  • Supports up to 5,000 Elastic Compute Service (ECS) instances in one cluster.
  • Supports multi-cluster management, cross-zone clusters, and the cluster federation feature.
  • Provides cross-zone high availability and disaster recovery.

Elastic resource scaling

  • Automatically scales the number of containers based on container resource usage.
  • Scales up to thousands of nodes in minutes.
  • If your application is deployed on elastic container instances (ECIs) in a serverless Kubernetes cluster, up to 500 pods can be started in 30 seconds.
  • Supports vertical scaling with a few clicks.
  • Supports affinity policies and scale-out for your services.
  • Provides standard Horizontal Pod Autoscaler (HPA), Vertical Pod Autoscaler (VPA), and Cluster Autoscaler capabilities provided by the Kubernetes community.
  • Provides the scheduled scaling capability similar to cron HPA and provides serverless scalability by using virtual-kubelet-autoscaler.
  • Uses ack-kubernetes-elastic-workload to provide fine-grained scheduling for online business.
  • Provides alibaba-metrics-adapter to meet different scaling needs. Application scaling is optimized by using ingress gateways, Sentinel-based microservice rate limiting, and other methods.

All-in-one container management

  • Application management:

1.Supports phased release, blue-green release, application monitoring, and application autoscaling.
2.Provides a built-in application store that allows you to deploy applications with a few clicks by using Helm.
3.Supports Service Catalog to simplify cloud service integration.

  • Repository (Alibaba Cloud Container Registry):

1.Provides high availability and high concurrency of image pull requests.
2.Supports accelerated image retrieval.
3.Supports large-scale P2P image distribution. By using the optimized image distribution process, ACK can automatically distribute images to a maximum of 10,000 nodes. This improves the distribution efficiency by four times.

  • Logs:

1.Allows you to collect logs and deliver the collected logs to Log Service.
2.Supports the integration with third-party open-source logging solutions.

  • Monitoring:

1.Supports both container-level and VM-level monitoring.
2.Supports the integration with open source monitoring solutions from third-party providers.

Multiple types of nodes

  • x86-based computing resources: ECS instances based on the x86 architecture
  • Heterogeneous computing resources: GPU ECS instances, NPU ECS instances, and FPGA ECS instances
  • Bare metal computing resources: ECS Bare Metal (EBM) instances
  • Serverless computing resources: ACK virtual nodes
  • Edge nodes: Alibaba Cloud Container Service for Kubernetes@Edge (ACK@Edge) supports centralized management of cloud and edge nodes and unified application release. It increases the release efficiency by three times.

IaaS capabilities

  • Network:

1.Provides a high-performance plug-in to assign elastic network interfaces (ENIs) to pods. The performance of this plug-in is 20% higher than that of regular network solutions.
2.Supports container access policies and throttling.

  • Storage:

1.Supports Alibaba Cloud disks, Apsara File Storage NAS volumes, and Object Storage Service (OSS) buckets, and provides standard CSI drivers.
2.Allows you to dynamically create and migrate volumes.

  • Load balancing:

Allows you to create public-facing and internal Server Load Balancer (SLB) instances.

The Scenarios of Container Service for Kubernetes

DevOps & continuous delivery

The optimized continuous delivery pipeline
Container Service integrates with Jenkins to automate the DevOps pipeline that ranges from code submission to application deployments. The pipeline ensures that code is submitted for deployment only after passing automated testing, and provides a better alternative to traditional delivery models that involve complex deployments and slow iterations.


  • DevOps pipeline automation
    Automates the DevOps pipeline, from code updates to code builds, image builds, and application deployments.
  • Environment consistency
    Allows you to deliver code and runtime environments based on the same architecture.
  • Continuous feedback
    Provides immediate feedback on each integration or delivery.

Related services


Microservice architecture

Agile development and deployment to speed up the evolution of business models
Your workload in the production environment is divided into multiple microservice applications, which are managed by Alibaba Cloud image repositories. Alibaba Cloud can schedule, orchestrate, deploy, and implement the canary releases of microservice applications and you only need to focus on feature updates.


  • Load balancing and service discovery
    Forwards layer 4 and layer 7 request and binds the requests to backend containers.
  • Multiple scheduling and disaster recovery policies
    Supports different levels of affinity scheduling policies, and cross-zone high availability and disaster recovery.
  • Microservice monitoring and auto-scaling
    Supports microservice and container monitoring, and microservice auto-scaling.

Related services
ECS, ApsaraDB for RDS, and OSS


Hybrid cloud architecture

Unified O&M of cloud resources
You can centrally manage cloud and on-premises resources in the Container Service console. Containers hide the differences between infrastructures. This enables you to use the same images and orchestration templates to deploy applications in the cloud and on-premises.


  • Application scaling in the cloud
    During peak hours, Container Service can scale up applications in the cloud and forward traffic to the scaled-up resources.
  • Disaster recovery in the cloud
    Business systems can be deployed on-premises for service provisioning and in the cloud for disaster recovery.
  • On-premises development and testing
    Applications that are developed and tested on-premises can be seamlessly released to the cloud.

Related services
ECS, VPC, and Express Connect


Automatic scaling architecture

Traffic-based scalability
Container Service enables workloads to auto-scale their resources based on traffic. This prevents traffic spikes from bringing down your system and eliminates idle resources during off-peak hours.


  • Quick response
    A scale-out event can be triggered within seconds when traffic reaches the scale-out threshold.
  • Autoscaling
    The scaling process is fully automated without human interference.
  • Cost efficiency
    Containers are automatically scaled in when traffic decreases to avoid resource waste.

Related services
ECS and CloudMonitor


0 0 0
Share on

Alibaba Clouder

2,605 posts | 746 followers

You may also like


Alibaba Clouder

2,605 posts | 746 followers

Related Products