Improve Cost Management Efficiency
Container Service for Kubernetes Serverless (ACK Serverless) is a secure and reliable container service that runs on top of the elastic computing infrastructure of Alibaba Cloud. ACK Serverless is fully compatible with the Kubernetes ecosystem. ASK allows you to deploy containerized applications in Kubernetes clusters without managing and maintaining the clusters. You are charged based on the CPU and memory resources used by applications. ACK Serverless helps you focus on application development rather than managing the underlying infrastructure.
Ease of Use
Enables easy and fast cluster creation and launches containers within seconds. Frees you from node management and allows you to focus on application development
Auto Scaling
Frees you from node capacity planning and enables on-demand application scaling
Kubernetes Compatibility
Allows you to deploy containerized applications in all Kubernetes use scenarios using the Kubernetes CLI or API without the modifications
Observability
ASK is integrated with CloudMonitor and Prometheus Service to monitor workloads and pods in ASK clusters. ASK is also integrated with Log Service to enable log analytics and event-based alerting for ASK clusters
Security Isolation
Pods in ACK Serverless clusters run in the Alibaba Cloud elastic computing architecture. Each pod is deployed in a secure and isolated runtime environment. The underlying computing resources of each pod are isolated by lightweight virtual sandboxes. Pods do not affect each other
Features
Simplified Cluster Management
Eliminates the need to maintain nodes and resource pools, reduces O&M costs, and allows you to focus on application development
Rich Elastic Container Instance-Based Pod features
Provides configurable pod specifications and supports EIP mounting, vGPU-accelerated instances, preemptible instances, and snapshots
Highly Elastic Serverless Architecture
Adopts a highly elastic Serverless architecture to deal with user traffic spikes and supports up to 10,000 pods per cluster
Rich Kubernetes Features
Supports Kubernetes-native features (such as Services, Ingresses, Deployments, Jobs, CRDs, and Helm) and provides a streamlined platform for application lifecycle management
Per-Usage and Per-Second Billing
Bills resources based on actual usage (only when the instances are in use), helping reduce expenses
Scenarios
Maintain Kubernetes Clusters at Low Costs
All nodes are cloud-managed, which saves you the effort of handling node anomalies. Pods are isolated from each other to ensure data security. This significantly reduces cluster maintenance costs.
ASK resolves the following issues:
-
Complicated Infrastructure Management and High Maintenance Costs
Kubernetes clusters that contain Elastic Compute Service (ECS) instances or self-managed compute nodes are high-maintenance in infrastructure management. ACK Serverless clusters can significantly reduce your maintenance costs. This allows you to focus on application development.
Automate Pod Scaling to Withstand Traffic Spikes
ACK Serverless automates pod scaling to withstand traffic fluctuations and avoid resource waste.
ASK resolves the following issues:
-
Dynamic Node Scaling Cannot Withstand User Traffic Fluctuations
ACK Serverless can easily scale out the number of pods when it detects a sharp increase in user traffic.
Cost-Efficient Data Computing Based on the Elasticity of ACK Serverless
ASK is commonly used with data processing engines (such as Spark and Presto). ASK allows you to create a large number of jobs within a short time to meet business requirements in the big data computing and AI-assisted computing sectors.
ASK resolves the following issues:
-
A High Ratio of Idle Resources Drives Up Computing Costs
ACK Serverless offers significant savings because ACK Serverless clusters are automatically sized and pods are dynamically created based on business requirements.
Cost-Effective Services for CI/CD Pipelines
ACK Serverless supports Jenkins and GitLab Runner and allows you to create Jobs on demand. This significantly reduces computing costs.
ASK resolves the following issues:
-
Challenges in Separating CI/CD Pipelines and High Maintenance Costs of Resource Pools
Pods in ACK Serverless clusters are forcefully isolated from each other when they are running. Computing resources are deployed and used on demand to avoid time-consuming maintenance of resource pools and to reduce computing costs.
Customers Success Stories
TuSimple
Rapid business development increases the complexity of models used by TuSimple. As a solution, TuSimple benefits from the combination of ACK and ACK Serverless services, which the enterprise uses for short-term and long-term computing jobs. This helps TuSimple improve resource utilization and reduce costs.
Prudence Medical
ACK Serverless clusters are low-maintenance, and the underlying ECS instances are cloud-managed. This helps Prudence Medical shift its focus from operating systems and OS images to container images.
Resources
Quickly and Automatically Build Container Images Using Serverless Kubernetes and Kaniko
Introduces how to automate image building by using Serverless Kubernetes and Kaniko.
Kubeless: A Deep-Dive into Serverless Kubernetes Frameworks
Introduces the Serverless architecture of Kubeless and describes how the architecture helps developers focus on innovation
From Serverless Containers to Serverless Kubernetes
Analyzes the architecture and infrastructure of Serverless Kubernetes.
Building Serverless Kubernetes Clusters for Serverless Jobs
Introduces the concept of Serverless clusters and describes the Serverless jobs that can run in managed clusters
