Enquire: Asia & Africa - +65-98008081 USA - +1-919-995-4114

Home Networking Servers & Storage HPE Mellanox InfiniBand EDR Managed Switch (834979-B22) 36-Port 100Gb QSFP28, Managed, RAF

HPE Mellanox InfiniBand EDR Managed Switch (834979-B22) 36-Port 100Gb QSFP28, Managed, RAF

Brand:
HPE

The HPE 834979-B22 Mellanox InfiniBand EDR Managed Switch is a high-performance 100Gb InfiniBand fabric switch designed for modern HPC clusters, AI infrastructure, machine learning, enterprise analytics, and high-performance storage networks. Equipped with 36 EDR (100Gb/s) QSFP28 ports, integrated management capabilities, and rear-to-front (RAF) airflow, it enables scalable, ultra-low-latency interconnects for mission-critical computing environments. Powered by the Mellanox Switch-IB 2 architecture, it delivers predictable application performance while simplifying fabric administration for growing enterprise and research data centers.

OVERVIEW

  • The HPE 834979-B22 Mellanox InfiniBand EDR Managed Switch is an enterprise-grade 36-port 100Gb InfiniBand EDR switch built for organizations requiring high-bandwidth, low-latency networking across HPC, AI, deep learning, and large-scale data analytics environments. Featuring 36 QSFP28 EDR ports, integrated management, and rear-to-front airflow (RAF) for optimized rack deployment, it provides a reliable foundation for scalable compute fabrics. Compared with unmanaged EDR models, this version includes built-in management capabilities that simplify monitoring, provisioning, and operational control while maintaining the same high-performance Switch-IB 2 architecture and SHARP acceleration. It is an excellent choice for enterprises expanding HPE Apollo or Mellanox-based infrastructures while protecting long-term investments through scalable, standards-based InfiniBand networking.

USE CASES

  • High-performance computing (HPC) clusters
  • Artificial intelligence (AI) and deep learning platforms
  • GPU cluster networking
  • Scientific and research computing
  • Enterprise machine learning infrastructure
  • High-speed RDMA-enabled storage networks
  • Financial modeling and quantitative analytics
  • Oil and gas simulation workloads
  • Engineering and manufacturing design simulations
  • Expansion of managed InfiniBand enterprise fabrics 

KEY FEATURES

  • 36 × QSFP28 InfiniBand EDR ports
  • 100Gb/s bandwidth per port
  • Managed InfiniBand switch architecture
  • Mellanox Switch-IB 2 switching ASIC
  • SHARP (Scalable Hierarchical Aggregation and Reduction Protocol) support
  • Up to 7 Tbps switching capacity
  • Approximately 90 ns port-to-port latency
  • Rear-to-front (RAF) airflow for optimized rack cooling
  • Ultra-low-latency fabric for HPC and AI workloads
  • RDMA support for high-performance computing
  • Hot-swappable redundant power supplies
  • Hot-swappable fan modules
  • 1U rack-mount form factor
  • Enterprise scalability for large compute fabrics
  • Designed for continuous, high-availability operation

TECHNICAL SPECIFICATIONS

  • Brand: HPE
  • Product: HPE Mellanox InfiniBand EDR Managed Switch
  • SKU: 834979-B22
  • Product Family HPE Mellanox InfiniBand EDR Switch Series
  • Switch Type: Managed InfiniBand Switch
  • Form Factor: 1U Rack Mount
  • InfiniBand Standard: EDR (100Gb/s)
  • Switch ASIC: Mellanox Switch-IB 2
  • Ports: 36 × QSFP28 EDR
  • Port Speed: 100Gb/s per port
  • Switching Capacity: Up to 7 Tbps
  • Port-to-Port Latency: Approximately 90 ns
  • SHARP Support: Yes
  • Management: Integrated management
  • Airflow: Rear-to-Front (RAF)
  • Power Supplies: Dual hot-swappable redundant power supplies
  • Cooling: Hot-swappable redundant fan modules
  • Dimensions (H × W × D): 4.37 × 42.77 × 68.58 cm
  • Operating Temperature: 0°C to 45°C
  • Compatible Platforms: HPE Apollo Systems, Mellanox InfiniBand adapters, InfiniBand EDR fabrics

Why Choose This Product?

  • The HPE 834979-B22 is designed for organizations that need both high-performance InfiniBand networking and centralized switch management. Unlike unmanaged EDR switches, this model enables easier configuration, monitoring, and lifecycle management while delivering 36 non-blocking 100Gb/s ports, ultra-low latency, and SHARP acceleration for distributed computing workloads. It integrates seamlessly with HPE Apollo servers, Mellanox adapters, GPU clusters, and enterprise HPC environments, making it a practical investment for AI, scientific computing, and data-intensive applications. Its managed architecture helps reduce operational complexity, improve infrastructure visibility, and maximize long-term return on investment.

CUSTOMER REVIEW

No Reviews Found

FAQ

No FAQs Found