Content

Target customers: Users who have requirements on AI application development. Users who intend to perform training or inference on petabytes of data. Users who intend to perform algorithm-driven business that depends on complex model iterations, such as precise advertising and real-time risk control. Features released: AnalyticDB for MySQL supports fully managed Ray clusters. Ray clusters provide the parallel data computing and cluster resource management capabilities, suitable for various scenarios such as machine learning, real-time analysis, and high-performance computing. AnalyticDB for MySQL allows Ray customers to quickly develop AI applications and perform data processing, model inference orchestration, and distributed execution in the native Python language. The following mainstream AI frameworks are supported: PyTorch, TensorFlow, and Dask. The fully managed Ray clusters of AnalyticDB for MySQL support hybrid resource deployment to improve resource utilization and reduce costs. The Ray clusters also support high availability of head nodes to improve stability and utilize the serverless capability to implement flexible resource scaling. √ Reduces O&M complexity of distributed training frameworks. √ Implements seamless connections between data lakehouse storage and model training. √ Scales GPU and CPU resources as needed to prevent idle computing power.