AIC Unveils the F2032-G6 JBOF Storage System to Enhance AI Inference Performance

AIC, a leading name in server and storage solutions, has recently announced the launch of its innovative F2032-G6 2U JBOF storage system. This new addition to the NVIDIA BlueField accelerated storage portfolio is specifically designed to optimize AI inference operations by efficiently managing KV cache context storage.

With the growing demands of data processing, the F2032-G6 stands out due to its High Availability architecture featuring dual-active nodes, ensuring that mission-critical workloads receive the resiliency and performance they require. This advanced system offers next-gen capabilities in performance, capacity, and availability, particularly suitable for flash-based storage deployments.

The F2032-G6 is equipped with two to four NVIDIA BlueField-4 Data Processing Units (DPUs), each capable of delivering an impressive 800Gb/s throughput. Additionally, it supports the integration of NVIDIA ConnectX-9 SuperNICs, further enhancing its performance. The architecture utilizes NVIDIA DOCA microservices, which facilitate breakthrough acceleration for the NVIDIA Inference Context Memory Storage Platform. Its dual-port design emphasizes link redundancy and continuous operation, crucial for maintaining high-throughput storage services in environments that prioritize scalability and distribution.

One of the significant advantages of the F2032-G6 is its ability to support up to 32 E3.S/L NVMe drives, providing a massive storage capacity of up to 8 petabytes within a compact 2U chassis. This is achievable through utilizing various configurations of SSDs (15, 30, 60, 122, or 256 TB). This not only enhances storage efficiency but also presents enterprises and cloud operators with a scalable solution for high-performance, power-efficient flash tiers.

Michael Liang, President and CEO of AIC, emphasized the company's commitment to utilizing the NVIDIA BlueField platform as a storage controller. 'As data demands continue to grow at unprecedented rates, the F2032-G6 delivers resilient, high-performance, and scalable flash storage that empowers critical workloads with unmatched speed, reliability, and efficiency,' said Liang.

NVIDIA's VP of Networking, Yael Shenhav, also highlighted the need for a fundamental architectural shift to accommodate the increasing complexity in AI inference workloads. The collaboration between AIC and NVIDIA aims to deliver accelerated, AI-native data storage platforms powered by the BlueField-4 DPU, designed to scale AI inference capabilities efficiently during this evolving era.

The F2032-G6 is particularly advantageous for various modern workloads, which include:
  • - Efficient KV cache context storage for AI inference
  • - Parallel and scale-out file systems that require robust performance
  • - Large-scale object storage, including mission-critical enterprise flash tiers

Through ongoing collaboration with NVIDIA and sustained alignment in product development, AIC ensures strong support for the BlueField-4 ecosystem in future platforms. This strategic partnership positions AIC at the forefront of high-density storage solutions, ready to meet the demands of an AI-driven marketplace.

AIC Inc. has nearly three decades of experience in delivering high-density storage servers and solutions, with significant advancements in AI storage and edge appliances, securing its place as a market leader. Focused on innovation and flexibility, AIC's in-house design, manufacturing, and validation processes enable a wide range of product configurations to meet diverse customer needs. For more information about their products, visit AIC's website or contact them directly.

Topics Business Technology)

【About Using Articles】

You can freely use the title and article content by linking to the page where the article is posted.
※ Images cannot be used.

【About Links】

Links are free to use.