Provide storage cache acceleration, support Posix and CS interfaces, cache hot-spot files from storages, and deliver high-performance and high-throughput data access service.
It is training-oriented, with bandwidth increasing linearly with the capacity to achieve million-level IO and throughput and improve the speed of model training.
It supports high availability up to 99.9% to ensure uninterrupted high performance during training acceleration.
It supports SDK, Posix, and CSI interfaces to meet different training scenarios.
It supports cache acceleration for object and file storage, and provides posix and CSI interfaces to meet the performance requirements of different training scenarios.
Allow users to scale up the capacity according to business demands to meet the performance requirements and scale down the capacity to reduce costs.
Before training, allow users to warm up the dataset to be trained to the cache for improving the training efficiency.
It provides cache services for S3 object storages.
Connect with various computing clusters to provide ultra-high throughput and ultra-high IOPS capabilities.
Optimize the cache service deeply to support large model training scenarios.
Help you achieve new breakthroughs in business with professional AI solutions and advanced AI products