Offline Asynchronous Task

In an offline asynchronous application scenario, a workload has one or more of the following characteristics:

Long execution time
Service processing takes minutes to hours, and the response time is insensitive.
Immediate return after submission
The system provides an immediate response upon receiving the call, ensuring that the lengthy processing does not obstruct the execution of the primary service logic.
Real-time task status monitoring
None.
Parallel processing
To handle large amounts of data and meet high GPU resource demands, offline GPU tasks require parallel processing through APIs to accelerate processing.
Data source integration
An offline GPU task has various requirements on data sources, and needs to frequently interact with multiple storage products (for example, OBS) and multiple message products (for example, message queue) during processing.

FunctionGraph offers the following benefits for offline asynchronous application workloads:

Simplified service architecture
Asynchronous processing improves system response, resource utilization, and availability.
Sufficient GPU resources
FunctionGraph provides sufficient GPU resources for any offline services that are always idle or have unpredictable traffic.
Data source integration
FunctionGraph supports multiple data source triggering modes, such as OBS and message queue.

Parent topic: Scenarios