Technology
Two Core Technologies
Compute-storage disaggregation and EBOF all-flash — fully proprietary. The only team in China achieving compute acceleration via hardware-accelerated all-flash.
Compute-Storage Disaggregation
A four-layer architecture: protocol decoupling, high-speed interconnect, node addressing, and elastic scheduling. IPv6-native modular design is our unique innovation.
NVMe-oF Protocol
Complete physical decoupling of compute & storage
RoCEv2 Interconnect
Eliminates TCP/IP congestion bottlenecks
IPv6 Dynamic Addressing
Node identity-address unified binding
SDN Smart Scheduling
Cluster-wide elastic resource scheduling
EBOF All-Flash Storage
PCIe Gen5 switch-direct NVMe SSD, gateway-free distributed pool, eliminating bottlenecks from the ground up. 8+2 EC erasure coding achieves six-9s protection with only 20% overhead.
PCIe Gen5 Direct
Gateway-free distributed storage pool
8+2 EC Erasure Code
20% overhead for 99.9999% reliability
Smart Data Migration
Wear leveling extends SSD life by 20%+
Two-tier Cache
Local NVMe + global RDMA shared cache
Architecture Gallery
Hardware Design
Fully self-designed PCB backplane supporting PCIe Gen5 switch and NVMe SSD direct connection. Modular hot-swap design enables scaling without downtime.
Combined Benefits
Native deep integration of two core technologies — not just functional stacking — delivering unified elastic compute + high-performance storage + local data security.
60%+
Response efficiency gain
40%+
Total cost reduction
90%
Scaling cycle reduction
0
Data leak risk
More Technical Details
Explore Our Products
Three products built on our core technology, covering small to large-scale compute needs.