Velocity • Durability

USE CASES

High-throughput parallel storage for AI training workloads.

Low-latency data access for real-time inference pipelines.

Persist KV cache and session context beyond GPU memory.

FEATURES

Per-tenant QoS, namespaces, and access policies.

API

REST, Kubernetes CSI, and CLI interfaces.

Direct memory access for maximum AI/HPC throughput.

MLEC and client erasure coding for data durability.

S3

S3-compatible object storage interface.

 

AES-256 encryption at rest and in-flight across every tenant data path.

Automatically moves data across NVMe, SSD, and object tiers by access pattern.

Performance

Up to 2.7 TB/s per rack for demanding workloads.

Observability

Real-time fleet telemetry and monitoring.

Freedom to choose your preferred hardware vendors.

Enterprise support from the VDURA expert team.