JuiceFS is a high-performance distributed file system designed for the cloud. And JuiceFS Cloud Service is a SaaS (Software-as-a-Service) solution provided by Juicedata. Unlike the open-source edition of JuiceFS, it is shipped with an in-house developed, high-performance metadata engine, so your team can use and manage JuiceFS storage more easily.

JuiceFS is designed for high-performance computing needs, commonly utilized in AI & machine learning for training, inference, model distribution, as well as in big data processing and analysis. Within industries such as generative AI, autonomous driving, quantitative trading, and bioinformatics, JuiceFS serves as a unified storage solution in users' data pipelines. It efficiently manages column-based structured data, semi-structured data like CSV and JSON, and various forms of unstructured data used for computation, including videos, images, and audio files.

JuiceFS separates "data" and "metadata" storage. Files are split into chunks and stored in Amazon S3 to enhance storage performance, while metadata (directories,file names) is stored in our fully managed service. The built-in caching capabilities of JuiceFS can significantly improve I/O throughput by utilizing the local storage of computing nodes, while also substantially reducing the bandwidth overhead of S3. With full POSIX compatibility, it enables S3 to be utilized as massive local disks, allowing them to be mounted and accessed on different hosts across platforms and regions.