JuiceFS

A cloud-native distributed file system that stores data on cost-efficient object storage while providing POSIX, HDFS, and S3 access with distributed caching and a separated metadata architecture for high throughput and low latency. It is for developers, operations and IT teams, and data teams in businesses that run AI/ML, big data, and Kubernetes workloads across cloud and multi-cloud environments. It is delivered as an Apache 2.0 open-source edition and as a fully managed cloud service and enterprise edition.

Key features

• POSIX, HDFS, and S3 protocol support
• Data-metadata separated architecture
• Supports Redis, PostgreSQL, TiKV metadata engines
• Distributed metadata service
• Elastic scalable throughput
• Distributed caching for object storage
• Mount as local disk without SDK
• Kubernetes PersistentVolume via CSI Driver
• Multi-cloud and hybrid cloud support
• Automatic cross-cloud data replication