VSS3120 – 3U GPU Server for AI/ML Edge Platform

Artificial Intelligence, Machine Learning and Hybrid Edge Solutions

The Viking Enterprise Solutions (VES) VSS3120, a storage and GPU server with massive storage capacity, is a high-performance computing solution designed to handle demanding workloads, such as AI training, deep learning, and large-scale data processing. This platform typically integrates multiple powerful GPUs, enabling parallel computing and accelerating tasks like machine learning, scientific simulations, and video rendering. Coupled with extensive storage capacity, often in the range of hundred of terabytes, it ensures fast data access and retrieval, essential for big data applications at the edge. The compact form factor strikes a balance between powerful hardware and efficient space utilization, making it ideal for data centers or environments where both computational power and data storage are critical. Its scalable design allows for additional storage or GPU resources to be added as needed, ensuring the system can evolve with growing data demands.

AMD EPYCTM Genoa-based, the VSS3120 offers 6 PCIe Gen 5 slots (supporting up to three dual-slot cards or five single-slot cards) and up to ten NVMe SSDs. CPU cores, GPU compute power, storage and connectivity can be adjusted to balance and optimize the system based on application needs.

The system is designed for maximum deployment flexibility: it’s designed to fit in a 2-post or 4-post rack so it can be deployed at edge points of the network for minimizing network load. With its exceptional thermal management, the platform is highly reliable in environments up to 35C and 80% humidity. VSS3120 utilizes hot-swappable AC and HVDC redundant power supplies and can operate indefinitely with a fan fail (also hot-swappable), minimizing service complexity and frequency. Optional intrusion detection provides extra hardware and content security, even when the system is powered off.

VSS3120 Features

• Multiple High-End GPUs Support: Enables parallel processing for accelerated AI, deep learning, and high-performance computing tasks.
• Massive Storage Capacity: Ranges from several hundreds of terabytes, ideal for big data applications.
• High-Speed NVMe SSDs or Enterprise-Class HDDs: Optimized for fast data throughput and low-latency access.
• Scalable Architecture: Allows easy expansion of both GPU and storage resources to meet growing demands.
• Redundant Power Supplies: Ensures system reliability and uptime by protecting against power failures.
• Advanced Cooling Mechanisms: Maintains optimal operating temperatures even under heavy workloads.
• Efficient Space Utilization: Compact form factor designed for data centers and space-constrained environments.
• High Availability: Designed for continuous, 24/7 operation with minimal downtime.

The VES VSS3120, GPUs, and storage-server combination provide a powerful edge infrastructure for the most demanding workloads while eliminating the need for a high-speed network between the server and storage, which in turn helps simplify edge AI/ML deployment. Some key use cases that benefit from this synergy are:

Other Data-Intensive Workloads

Large-scale data analytics, real-time fraud detection, and complex financial modeling can all benefit from the processing power of GPUs combined with fast storage to handle massive datasets efficiently.