Community Edition
Self Hosted
Cloud Service
Cloud Hosted
Everything in Community Edition, plus:
Enterprise Edition
On-premises (starting from 100TB)
Everything in Cloud Service, plus:
| Community Edition | Cloud Service | Enterprise Edition | ||
|---|---|---|---|---|
| Metadata Engine | Redis, TiKV, MySQL, PostgreSQL, etc. | High performance distributed metadata engine developed by Juicedata | High performance distributed metadata engine developed by Juicedata with horizontal scalability | |
| Data Storage | 30+ Object Storage | 30+ Object Storage | 30+ Object Storage | |
| File Amount | 100 million - 10 billion | 100 million | 100 billion | |
| Data Volume | 1PB - 100PB | 10PB | 100PB | |
| Metadata Throughput | 10,000 - 1,000,000 IOPS | 100,000 IOPS | Scales horizontally to 10,000,000 IOPS | |
| Number of Clients | 100 - 10,000 | 1,000 | 10,000 | |
| POSIX | ||||
| Kubernetes CSI Driver | ||||
| Hadoop SDK | ||||
| S3 Gateway | ||||
| Local Metadata Cache | ||||
| Local Data Cache | ||||
| Metadata Cache Expiration by TTL | ||||
| Active Metadata Cache Invalidation | ||||
| Distributed Data Cache | ||||
| Independent Cache Cluster | ||||
| Global Quota | ||||
| Data Compression | ||||
| Subdirectory Quota | ||||
| Cross-cloud and Cross-region Replication | ||||
| Web Console | ||||
| Data Encryption | ||||
| Trash | ||||
| Snapshot | ||||
| POSIX ACL | ||||
| Access Control | ||||
| Kerberos | ||||
| Apache Ranger | ||||
| Support | Community support: GitHub, Slack channel | Professional team online support | Customizable support | |
| Deployment | Self-hosted | Fully managed cloud service | On-premises |
We are proud to announce that our Cloud Service has achieved SOC 2 compliance, demonstrating our commitment to maintaining the highest standards of security, availability, and confidentiality for our customers.
This compliance validates that our security controls are designed and operating effectively to protect your data.
Thanks to JuiceFS, which supports concurrent access and is fully compatible with POSIX API, we are able to implement AI distributed learning and large-scale service applications on our AI platform. With JuiceFS, our system processes massive amounts of data with stability and efficiency.
NAVER Engineering Center
We built a general-purpose storage solution using JuiceFS, tailored to meet the specific needs of AI: high performance, reliability, low cost, and multi-cloud compatibility. At the same time, it can be mounted on GPU instances just like a regular filesystem. This allows us to focus on what we do best—building a cloud-native AI platform and optimizing high-performance GPU compute—while leaving storage optimization to the experts.
Lepton AI CEO
JuiceFS not only provides a storage system that meets our hybrid cloud needs, but also optimizes the data processing workflow with high-performance metadata services and multi-interface compatibility, significantly reducing operational costs. It is the ideal choice for running our LLM platform.
MiniMax Technical Director
