Thank you for your support for Alibaba Cloud Batch Compute. To improve our technology and service experience, the original features of Alibaba Cloud Batch Compute will be merged with Alibaba Cloud Elastic High Performance Computing (E-HPC). We will use E-HPC Instant Computing Service to provide you with a simpler, more economical, and more efficient intelligent batch processing service. Alibaba Cloud plans to stop new user registration for Batch Compute on May 15, 2025. Users who have activated the service will not be affected and can continue to submit jobs. The service will be completely discontinued on May 15, 2026, at which time the Batch Compute console, technical support, and ticket services will also be discontinued.
To ensure your business continuity, we recommend that you migrate your workloads to Alibaba Cloud E-HPC Instant Computing Service at the earliest opportunity. During the transition period, we will provide technical support to existing users to help them complete a smooth transition. If you have any questions, contact our customer service team at any time.
Submitted jobs: In most cases, the jobs stop after running for a period of time, from a few minutes to several hours. Data is stored in Object Storage Service (OSS) or File Storage NAS and does not need to be migrated.
New jobs: We recommend that you use E-HPC Instant Computing Service to submit jobs. Compared with Batch Compute, E-HPC Instant Computing Service provides a simpler, more economical, and more efficient intelligent batch processing service. However, E-HPC Instance Computing Service involves API adaptation costs. You can go to E-HPC Instant Computing Service OpenAPI Explorer and view SDK examples in different programming languages, such as Python and Java.
The following table describes a feature comparison between E-HPC Instant Computing Service and Batch Compute.
✅ indicates supported and ❌ indicates unsupported.
Feature | Batch Compute (being discontinued) | E-HPC Instant Computing Service |
Manage clusters | ✅ | Not required. Fully serverless computing without the need to create or manage clusters. |
Run batch jobs | ✅ | ✅ |
Run background services | ❌ | ✅ |
Use virtual machines or Docker containers and custom images | ✅ | ✅ |
Use ECS instances | ✅ | Not required. You can specify computing resources by CPU and memory, which is more economical. |
Support Directed Acyclic Graph (DAG) workflows | ✅ | ❌ Integrates with mainstream third-party workflow engines. |
Support Apps | ✅ | ✅ |
Access OSS or File Storage NAS | ✅ | ✅ |
Access the Internet | ❌ | ✅ |
O&M and monitoring | ✅ | ✅ |
Notifications | ✅ | ✅ |