Build Data Processing Systems Based on Event Sources

Function Compute provides multiple event sources. The event triggering mechanism can process data in real time with just a few lines of code and simple configurations. For example, the mechanism can decompress OSS packages, cleanse logs generated by Log Service or Tablestore data, and customize consumption of MNS messages.

Benefits

Integrates Multiple Easy-to-configure Event Sources

- Function Compute supports multiple event sources to process event source data by using simple configuration.

Flexibly Customizes Processing Logic

- Function Compute allows you to flexibly customize processing logic based on business scenarios.

Provides the Serverless Architecture for AI Engineers to Focus on Algorithms

O&M-free and elastically scalable Function Compute allows algorithm engineers to convert trained models into elastic and highly available reasoning services.

Benefits

Enables AI Engineers to Focus More on Algorithms and Avoid Complex O&M Operations

- Artificial intelligence (AI) engineers can focus on algorithm model training and business logic development. Function Compute performs cluster O&M operations to improve work efficiency.

Mobilizes Tens of Thousands of Computing Resources to Eliminate the Computing Power Bottleneck

- Function Compute has an abundance of computing resources that can be elastically scaled within milliseconds. Function Compute can quickly mobilize tens of thousands of computing resources to smoothly handle traffic peaks.

Provides Multiple Versions for A/B Testing to Reduce Model-launching Risks

- Function Compute provides the multi-version feature and supports the canary release strategy for models. This allows you to implement A/B testing on algorithms to reduce model-launching risks.

One-click Installation of Third Party Libraries

- Function Compute upgrades the toolchain to improve the installation process of third-party libraries such as TensorFlow and PyTorch. You can deploy your applications to Alibaba Cloud by one click.

Build Elastic and Highly Available Video Processing System Based on the Serverless Architecture

Function Compute and Function Flow can be used together to create elastic and highly available Serverless video processing systems that have enhanced performance and efficiency as well as lower costs.

Benefits

Flexible Transcoding: Supports Custom Transcoding Processing Logic

- You can customize video transcoding functions and logic.

Low Cost: Provides Costs Reductions of Over 75%

- Function Compute provides flexible billing methods for a variety of scenarios to save costs.

Parallel Transcoding: Automatically Scales Based on the Number of Video Files

- When multiple video files are uploaded at the same time, Function Compute automatically performs scaling in milliseconds to mobilize computing resources and process multiple files in parallel. You are only charged for the resources you have used.

Fast Migration: Lower Migration Costs and Simplified Operations

- FFmpeg-related commands can be directly run in Function Compute. FFmpeg-based services deployed on virtual machines can be migrated to Function Compute at low costs.