
Safe Checkout
Secure Payments
Fast Delivery
Order Today
Free Shipping
Across the US
Easy Returns
Hassle-Free
HP 590203-B21 QDR 324 x Ports Switch 18 x Ports FC Line Module
- 324 x QDR InfiniBand ports
- 18 x FC (Fibre Channel) line module ports
- Designed for High-Performance Computing (HPC)
- Provides extremely low latency
- High bandwidth interconnectivity
- Supports advanced fabric management
- Scalable architecture for large clusters
Free U.S. Ground Shipping
Typically 1-2 handling + 3-7 transit days
Purchase orders accepted
For government, enterprise, data center, and small business customers.
Bulk Purchase Inquiry
Volume pricing and availability
Product Overview
The HP 590203-B21 is a QDR InfiniBand switch with 324 ports, featuring 18 ports for an FC line module. It is designed for high-performance computing (HPC) environments requiring extremely low latency and high bandwidth interconnectivity.
Technical Information
| Product Type | InfiniBand Switch |
| Technology | QDR InfiniBand |
| Total Ports | 324 |
| Line Module Ports | 18 x FC |
Additional Specifications
| Application | High-Performance Computing (HPC) |
| Latency | Sub-microsecond |
| Form Factor | Rack-mountable |
Product Description
The HP 590203-B21 is a high-density, high-performance switch designed for demanding High-Performance Computing (HPC) environments. It utilizes QDR InfiniBand technology, which offers significantly higher bandwidth and lower latency compared to traditional Ethernet networks. With a total of 324 QDR InfiniBand ports, this switch can connect a very large number of compute nodes, enabling massive parallel processing capabilities essential for scientific simulations, financial modeling, and big data analytics. In addition to its extensive InfiniBand connectivity, the switch features 18 ports dedicated to an FC (Fibre Channel) line module. This integration allows for seamless connectivity to storage area networks (SANs) or other Fibre Channel infrastructure, providing a unified fabric for both compute and storage traffic. This hybrid approach is often beneficial in HPC clusters where high-speed access to large datasets is critical for performance. The switch is engineered to provide sub-microsecond latency, which is paramount for tightly coupled parallel applications where inter-process communication must be extremely fast. Its scalable architecture allows for the creation of large, non-blocking fabrics, ensuring that all connected nodes can communicate efficiently without performance degradation. Advanced fabric management features enable administrators to monitor, control, and optimize the network for maximum throughput and reliability.



