20" deep XR6 Compute, Storage and Expansion Modules

Capability-dense "plug-and-pull" blades that can be combined, configured and repurposed in multiple chassis to meet precise application needs

  • slide

XR6 RES HD Modules

The growth of big data, 5G networks and artificial intelligence (AI) continues to increase compute and data storage demands on edge systems. IT administrators and systems architects must consider trade-offs between performance, bandwidth, latency and cost. Our capability-dense "plug-and-pull" RES HD blades feature the latest Intel® Xeon® scalable processors and are purpose-built to deliver cost savings, keep pace with growing edge workloads and drastically reduce your overall server footprint.

DOWNLOAD COMPUTE, STORAGE AND EXPANSION BLADES DATASHEET

Rapidly Launch the Latest Technology

Rack server refreshes are difficult and costly endeavors that require “rip and replace” when an entire server must be taken offline, removed from the rack, then upgraded or replaced. Mercury’s hot swappable “plug-and-pull” blades can be individually replaced while the server remains in the rack and stays operational to maximize uptime and cost-effectively refresh technology as your application evolves.

Ensure Uptime with Highly Available Solutions

Mix and match blades in edge-ready 2U, 3U or 4U slim chassis options to construct granulate architectures and achieve precise functionality for any platform. Minimize the need for backup or spare servers by building in local level blade redundancy or virtualizing nodes. Our edge-ready blades incorporate advanced thermal and mechanical design features to accelerate performance in any environment.

hd-module-banner.jpg

Repurpose to Maximize Investment

Blades can be reconfigured or repurposed across multiple applications and platforms, minimizing the need for spares, streamlining serviceability and maximizing investment. With per-blade weights low as 5lbs, they are easy to store, ship, carry and replace.

RES HD XR6 Blade Specifications

Common Blade Technical Specifications

The following specifications apply to select compute (HDC, HDC-U.2), expansion (HDP) and storage (HDS4, HDS8, HDS8R) blades.

Specifications

Dual Intel® Xeon® scalable processors with 22 cores, 44 threads*
Up to 1TB DDR4 ECC 2666MHz, 16 DIMM memory
2 Disks on module (up to 128GB per DOM)

I/O Options

2 USB 3.0 ports
1 VGA port
4 Ethernet ports (RJ45) with multiple Ethernet options:

  • Single port: 100Gbps (QSFP28)
  • 2 ports: (2)1GBaseT, (2) 25Gbps SFP28, (2) 10GBaseT, or (1) 1GBaseT + (1) 100Gbps (QSFP28)
  • 3 ports: (2) 1GBaseT + (1) 56Gbps (QSFP)
  • 4 ports: (4) 1GBaseT, (4) 10GBaseT, (4) 10G Eth / SFP+, (2) 1GBaseT + (2) SFP28 (25/10Gbps), or (2) 1GBaseT + (2) 56Gbps (QSFP)

1 Serial port


Mechanical
Height: 1.6 inches (41 mm, single slot) or 3.35 inches (85 mm, double slot)
Width: 7 inches (177 mm)
Depth: 20.3 inches (515 mm)
Weight: See individual blade specifications

*Typical customer specifications. Specifications are configuration dependent. Up to 20 cores per processor for 0-50°C thermal profiles. 
**Typical customer specifications. Actual weights are configuration dependent. 

Compute Blades Technical Specifications (HDC)

Compute Blade (HDC)

hdc-computemodule.png

See Common Blade Technical Specifications

Weight**

  • 10.5 lb. (4.76kg)

Expansion

  • Up to 2 PCIe 3.0 x16
    low profile, half-length cards

 

Compute Blade with U.2 NVME Storage (HDC-U.2)

hdc-u.2.png

See Common Blade Technical Specifications

Weight**

  • 10.5 lb. (4.76 kg)

Expansion

  • Up to 1 PCIe 3.0 x16
    low profile, half-length card
    (in lieu of U.2 SSD)

Storage

  • 2 NVMe U.2 drives for up to 64TB

Expansion/Storage Combined

  • Storage: 1NVMe U.2 drives for up to 32TB
  • Expansion: Up to 1 PCIe 3.0 x16
    low profile, half length cards

Reduce Data-Storage Bottlenecks at the Edge

Gain faster insights, support larger data sets required for machine learning and collect more or higher-resolution data from multiple edge devices. Our newest blades pack multiple processors and high-capacity, high-performance NVMe storage so you can accelerate decision-making by capturing, storing and analyzing data at the edge in real time instead of discarding information or sending it to the cloud for processing.

Storage Blades Technical Specifications (HDS)

Storage Blade With 8 Front-Access EDSFF-based Intel E.1L Drives (HDS8R)

hds8r.png

See Common Blade Technical Specifications

Weight**

  • 18 lb. (8.16 kg) with drives

Storage

  • 8 EDSFF E1.L drives
    for up to 256TB of storage

Expansion

  • Up to 2 PCIe 3.0 x16, low profile, half-length cards

Front I/O Expansion

  • Up to 2 PCIe 3.0 x16 low profile

Rear I/O Expansion

  • Up to 2 PCIe 3.0 x16 low profile

 

Storage Blade With 8 SATA or SAS3 Drives (HDSE)

hdse-cmyk-2.png

Weight**

  • 8 lb. (3.63kg) with drives

Requirements

  • Special rear I/O configurations

Expansion

  • Low profile, half-length cards

Storage

  • 8 SATA or SAS3 drives
    for up to 240TB of storage

Storage Blade With 4 Front Access SATA, SAS3 or U.2 Drives (HDS4)

hds4-flat.png

See Common Blade Technical Specifications

Weight**

  • 15 lb. (6.80kg) with drives

Storage

  • 4 front-access SATA, SAS3 or U.2 storage drives
    for up to 120TB of storage

Expansion

  • Low/high profile, half-length cards

Front I/O Expansion

  • Up to 2 PCIe 3.0 x16 + 1 PCIe 3.0 x8

Rear I/O Expansion

  • 2 PCIe 3.0 x16 + 1 PCIe 3.0 x8 or 1 PCIe 3.0 x16 + 3 PCIe 3.0 x8

Storage Blade With 8 SATA or SAS3 Drives (HDS8)

hds8-flat.png

See Common Blade Technical Specifications

Weight**

  • 18 lb. (8.16kg) with drives

Storage

  • 8 SATA or SAS3 drives
    for up to 240TB of storage

Expansion

  • Low profile, half-length cards

Front I/O Expansion

  • Up to 2 PCIe 3.0 x16

Rear I/O Expansion

  • Up to 1 PCIe 3.0 x16 + 2 PCIe 3.0 x8

 

Optimize Big Data Workloads with Parallel Computing

Harness parallel processing to maximize throughput, boost productivity and push the boundaries of compute-heavy AI, machine learning and sensor fusion applications. Our HDP blades tackle demanding workloads with the latest NVIDIA A100 Tensor Core GPU or multiple T4 GPUs.

PCIe Expansion Blade Technical Specifications (HDP)

PCIe Expansion Blade supporting NVIDIA A100 Tensor Core GPU (HDP)

hdp-pcieexpansionmodule.png

See Common Blade Technical Specifications

Weight**

  • 12.5lbs (5.67kg), 15lbs (6.80kg) with NVIDIA card

Expansion

  • Up to 5 PCIe Cards
    • 1 double wide PCIe 3.0 x16 (eg. NVIDIA GPU)
      2 PCIe 3.0 x16, low profile, half-length cards
    • 1 PCIe 3.0 x 8, 1 PCIe 3.0
      low profile, half-length cards
    • 2 PCIe 3.0 x16 low profile, half-length cards and
      2 PCIe 3.0 x8 full-height, full-length cards

Efficiently Move Data in Real-Time

Streamline demanding edge workloads with high-speed, low-latency and bottleneck-free networking through our 1GbE or 100GbE blade switches that integrate data center-proven Mellanox technologies.  

DOWNLOAD NETWORKING BLADE DATASHEET

Networking Blades Technical Specifications (HDN)

1GbE Networking Blade (HDN1)

hdn-mellanoxswitchmodule.png

Weight*

  • 3.08 lbs (1.4 kg)

Ethernet Ports (RJ45)

  • 16 1GbE

Other Ports (RJ45)

  • 2 (serial, ethernet management)

Network

  • Layer 2/3 Marvell Prestera®-DX PonCat3

Processor

  • Intel® i5-5300U vPro Processor

Routing

  • IPv4/IPv6 - IP multicast, VLANs, IETF, IEEE, DSL Forum

Quality of Service

  • QoS IEEE 802.1 priority tagging, DSCP, traffic queues

Management

  • Ethernet (http, telnet, SNMP, and RS232 (CLI)

Functionality

  • 48Gbps performance switching capacity
  • 35.71 Mpps maximum forwarding rate
  • 8K entries MAC address table size, 3.5Mbits Packet Buffer
  • Forwarding mode: store and forward
  • Supports half/full duplex operation at 10/100Mbps
  • Supports port auto-negotiation

 

100GbE Networking Blades (HDN100)

Weight*

  • 4 lbs (1.81 kg)

Switch

  • Integrated Netgear M4300-8X8F

Ethernet Ports

  • 16 10GbE Ports that include:
    8 10GBASE-T (Fast Ethernet, 1G and 10G speeds)
    8 x SFP+ (1G and 10G speeds)

Network

  • Layer 3 static, policy-based
    RIP OSPF VRRP PIM dynamic routing

Other Ports (RJ45)

  • 1GbE port for out-of-band management (OOB, RJ45)
    RS232 for local management

Other Ports

  • Mini-USB port for local management
    USB for storage, logs, configuration or images

Performance

  • Mini-USB port for local management

Cost-Effective and Scalable

Rapidly Launch Technology and Achieve Precise Functionality

Our capability-dense "plug-and-pull" blades can be combined, configured and repurposed in multiple chassis to deliver cost savings, keep pace with growing edge workloads and drastically reduce your overall server footprint.

LEARN ABOUT OUR BLADE SERVERS

DISCOVER MORE

Webinar: Harness More Data with Intel-based Rugged Computing

Handling AI, SIGINT or compute-heavy workloads that require moving data at real-time speeds? Learn...

SEE MORE

Webinar: Accelerate Time to Insight with NVIDIA-based Rugged Computing

Eliminate bottlenecks, process, store and move huge volumes of sensor data at 5G speeds and...

SEE MORE

White Paper: Bringing Commercial Technologies to Defense in the Age of AI

Artificial Intelligence (AI) is rapidly transforming the defense industry by broadening the scope of...

SEE MORE

Webinar: AI at the Tactical Edge: Securing Machine Learning for Multi-Domain Operations

In this webinar, you will learn how Mercury Systems, NVIDIA and SigmaX are securing machine learning...

SEE MORE

White Paper: Tackling AI and Multi-Domain Workloads in Real Time with PCIe Gen 4.0

PCIe Gen 4 technologies increase application efficiency by eliminating bottlenecks and increasing...

SEE MORE

Strengthening Cybersecurity at the Edge: Intel and Mercury talk multilayered security

Taking secure computing to the tactical edge requires a built-in multilayered security approach from...

SEE MORE

Webinar: AI at the Tactical Edge: Protecting Machine Learning with Data Security

In this webinar, learn how Mercury Systems and Intel are working together to design secure AI edge...

SEE MORE

Webinar: Computing in the Wild: How to Build Your Resilient Edge

AI processing capability remote from the data center is required for smarter and more autonomous...

SEE MORE

Innovation That Matters®
BACK TO TOP

Copyright © 2021 Mercury Systems, Inc. All rights reserved.