KDSS is a production-ready distributed storage system built with Reed-Solomon erasure coding, delivering exceptional space efficiency and data reliability.
Enterprise-grade features designed from the ground up for reliability, performance, and operational simplicity.
Reed-Solomon EC with configurable data/parity ratios. Up to 90.6% space efficiency, far surpassing 3-way replication.
Bypass filesystem overhead with direct disk I/O. Sequential append-only writes with intelligent merge and tunable fsync dramatically reduce random I/O and head seek wear, significantly extending HDD lifespan, lowering failure rates, and reducing rebuild frequency and hardware replacement costs.
Full S3 API compatibility including multipart upload, presigned URLs, and streaming for large files.
Mount storage as a local filesystem with POSIX interface. Transparent EC encoding/decoding with read/write buffering.
Automatic fault detection and EC reconstruction. Failed disks are isolated and data migrated to healthy nodes.
Full-featured dashboard for cluster management, monitoring, alerting, and RBAC access control.
Choose the right erasure coding profile for your workload.
| Configuration | Recommended Servers | Data Shards | Parity Shards | Fault Tolerance | Space Efficiency |
|---|---|---|---|---|---|
EC 5+2 |
9 | 5 | 2 | 2 | 71.4% |
EC 9+2 |
13 | 9 | 2 | 2 | 81.8% |
EC 18+3 |
24 | 18 | 3 | 3 | 85.7% |
EC 29+3 |
36 | 29 | 3 | 3 | 90.6% |
EC 31+2 |
36 | 31 | 2 | 2 | 93.9%* |
* EC 31+2 Ultra-Density Mode is designed exclusively for brand-new enterprise-grade servers and drives, deployed in Tier 3+ data centers with 24/7 on-site staff for immediate hardware fault response. With only 2 parity shards, the fault tolerance margin is minimal — any delay in disk replacement increases the risk of data loss.
A distributed architecture with clear separation of concerns between metadata, storage, and access layers.
Interested in KDSS? We'd love to hear from you.