Utilizing an innovative utilitarian approach, the Penguin Computing® Tundra® Extreme Scale (ES) Open Bridge Rack is a breakthrough rack standard designed specifically as a data center solution for OCP. The rack is designed to maximize both efficiency and density for all of the user’s equipment, ensuring that components achieve a maximum lifespan. Each rack is divided into 48mm (1.889in) tall sections or zones, called an OpenU, providing better thermal design. The outer chassis measures 28in while the inner casing is 24in wide, allowing larger space efficiency and utilization when compared to the current industry standard 19in rack.
Penguin Computing offers the Open Bridge rack with two power zones as a solution for HPC power needs, providing up to 52.8 kW total with N+1 redundancy. We can provide various redundancy and power profiles depending on customer demand. A single power zone configuration is available for flexible power requirements including data center workloads and other demanding environments.
Single Power Zone:
- In-rack battery backup supported
- Up to 26.4 kW with N+1 redundancy (277V/480V)
- 208V, 230V, 277/480V AC 3-phase input
- 40 OpenU height
Two Power Zones:
- Up to 52.8kW (26.4kW per zone) with N+1 redundancy (277/408V)
- 208V, 230V, 277/480V AC 3-phase input
- 40 OpenU height
Rack Level Integration: Because Efficiency Adds Up
If your organization deploys servers on a very large scale in a service-based infrastructure,your platform requirements are different from typical medium- and small-scale deployments. We know how to build highly efficient solutions that are specific to your actual workload and datacenter environment. We also offer highly energy-efficient solutions according to the Open Compute specifications. Talk to us to find out how we can design a highly efficient custom solution for your data center.
- Efficiency is key – Even seemingly minor efficiency gains can significantly improve your datacenter’s TCO, such as the reduction of overall power consumption through a consolidation of fans.
- Servers can be ‘vanity free’ – Components that are taken for granted in conventional environments, such as chassis covers, can be eliminated to free up capital for more compute or storage capacity.
- One server is rarely a single point of failure – Typical large scale applications offer resiliency and redundancy features at the application level and your operation relies much less on specific servers. Redundancy features such as redundant power supplies can be eliminated saving capital ex as well as operational expense.
Custom Solution Example
For a well known social networking and micro-blogging service, Penguin Computing™ developed a green data center solution that was radically optimized for overall efficiency, decreasing the overall power consumption per rack by 23% compared to a traditional rack of 1U rack-mounted systems. To maximize compute density, flexibility, Power Usage Effectiveness (PUE) and manageability the system features:
- Central power – 6 central, hot-pluggable, platinum power-supply units connected to a 3-phase AC power source for balanced power consumption; 12V DC power delivery to each server through a central power bus bar
- Central cooling – Cooling through N+1 redundant, hot-pluggable fans enclosed in fan modules. Compared to a traditional rack configuration this custom configuration populated with 42 full-width servers only requires 18 fans, compared to 210 for a traditional server solution, resulting in power savings of 75% for cooling per server compared to a traditional rack of 1U rack-mounted systems.
- Rack-integrated UPS – Built-in central UPS module(s) in a space-saving 1U form factor
- Cables access from front of rack – All server maintenance tasks can be performed from the ‘cold aisle’ front of the system
- Flexible configuration options – The rack can be configured to different requirements. A variety of full- and half-width options for server as well as storage nodes with a storage density of up to 12 drives per 1U are available.
