Introducing the EonStor GS 5024U

Infortrend Technology has announced a significant advancement in its unified storage portfolio with the release of the EonStor GS 5024U. This new system is engineered to meet the rigorous performance requirements of rapidly growing AI workloads, providing a robust foundation for applications ranging from complex model training to immediate data analytics.

Key Features and Performance

At its core, the EonStor GS 5024U leverages a powerful Intel Xeon 6 processor to deliver exceptional speed and efficiency. The system achieves an impressive 125 GB/s throughput – representing a substantial increase over previous models – alongside 2.4 million IOPS. This performance boost is significantly enhanced through support for the latest U.2 NVMe SSD technology, designed to maximize data transfer rates and minimize latency.

The EonStor GS 5024U’s architecture incorporates several key technologies to further optimize its capabilities

  • PCIe 5.0 Support: Enables the system to handle demanding workloads with unparalleled speed and responsiveness, facilitating faster data access and processing.
  • 200GbE Networking: Integrated high-speed networking provides seamless connectivity for large-scale deployments and efficient data transfer between storage systems and servers.
  • NVMe-oF Technology: This protocol optimizes the performance of NVMe SSDs over network connections, dramatically improving data access speeds for distributed applications.
  • GPUDirect Storage: By directly connecting GPUs to the storage system, this technology eliminates bottlenecks and maximizes GPU utilization, significantly accelerating AI training processes.

Designed for Demanding Workloads

The EonStor GS 5024U is specifically designed to tackle the challenges presented by modern AI workloads. Its high throughput and low latency make it ideally suited for

20121107 - building Goliath - SSD - box - IMG_4950
  • Model Training: The system’s speed allows for faster iteration cycles during model development, significantly reducing training times.
  • Inference: Real-time data processing is accelerated, enabling rapid responses from AI models in various applications.
  • Real-Time Analytics: The EonStor GS 5024U can efficiently handle large volumes of streaming data, facilitating immediate insights and decision-making.

Furthermore, the system’s compatibility with parallel file systems such as Lustre ensures it can deliver hundreds of gigabytes of throughput – a critical requirement for many advanced AI applications.

Scalability and Reliability

Beyond its performance capabilities, the EonStor GS 5024U is built for scalability and reliability. Its unified storage architecture provides flexible capacity options to accommodate growing data needs. The system’s robust design incorporates redundant components and advanced error correction mechanisms, ensuring high availability and data protection.

EonStor GS 5024U - Image 1 EonStor GS 5024U - Image 2

Conclusion

The Infortrend EonStor GS 5024U represents a significant step forward in unified storage solutions for the AI era. Its combination of high performance, advanced technologies, and scalability makes it a compelling choice for organizations seeking to accelerate their AI initiatives.