Story image

Thinking outside of the rack: Designing and planning by pod

30 Jun 17

In olden times, the supporting infrastructure for data centers and IT rooms would be built out all at once…based on some vague understanding (guess?) of what future IT needs would be 5 or more years down the road.

This obviously led to overbuilt and/or stranded capacity as one resource got used up before others. Some of this behavior was driven by budgeting practices, perhaps…“this is my one shot, so I’m going to try to get as much as I can possibly get” - type of mentality.

But as modular power and cooling systems and architectures began to emerge, building in more of a “pay-as-you-go” approach came into being. The IT rack rather than a server became the core unit of organization that designers use to estimate the requirements of the facility. 

Knowing average and peak rack power densities along with the number of racks not only allows for efficient facility sizing, but also determines branch circuit ratings and the heat rejection needed for each rack. Racks have typically been the unit of deployment in data centers. 

They mark the location of physical servers and their VMs. In DCIM systems, dependencies on power paths and cooling resources are mapped to the rack.

However, for larger data centers, IT deployments often come in larger and larger chunk sizes.

These “pods” of racks are deployed together and typically share common resources such as air containment systems, a PDU, network switch, or perhaps are grouped because they share a business function or serve a particular client or tenant. 

But unlike the simple, highly standardized 19” rack, there is no industry standard for IT pod design and deployment. White Paper 260, “Specifying Data Center IT Pod Architectures” proposes a high-level standardized spec that can be used to define IT pod requirements and help to compare and contrast design alternatives quickly.

The paper explains these 3 main drivers of an IT Pod’s design in detail: the choice of the electrical feed (power), the physical space available (i.e., number of racks), and the average rack power density required. 

When commonly available voltages and breaker sizes are considered, optimum pod configurations emerge that make planning and design easier. 

Standardizing pod designs and limiting the number of configurations can help make pod-level deployments simpler and faster. Organizing IT racks into pods makes it easier to vary power and cooling redundancies and architectures based on the specific business needs within an IT room or hall. 

A free-standing pod frame system can decrease data center project time by reducing construction time, allowing off-site integration of IT racks and IT gear. 

In addition, it simplifies maintenance by making it easier to roll racks in and out since containment and services are mounted to the frame instead of the rack. 

Article by Patrick Donovan, Schneider Electric Data Center Blog Network

Data centre cybersecurity actions that most people overlook
Schneider’s Steven Carlini discusses ways to improve data centre cybersecurity that most people don’t think of until it’s too late.
Alibaba Cloud showcases commitment to Hong Kong
The company’s service capability in Hong Kong has doubled since it established its first data centre in the city in 2014.
5 tips to reduce data centre transceiver costs
Keysight Technologies' Nicole Faubert shares her advice on how organisations can significantly reduce test time and cost of next-generation transceivers.
The new world of edge data centre management
Schneider Electric’s Kim Povlsen debates whether the data centre as we know it today will soon cease to exist.
Can it be trusted? Huawei’s founder speaks out
Ren Zhengfei spoke candidly in a recent media roundtable about security, 5G, his daughter’s detainment, the USA, and the West’s perception of Huawei.
SUSE partners with Intel and SAP to accelerate IT transformation
SUSE announced support for Intel Optane DC persistent memory with SAP HANA.
Inspur uses L11 rack level integration to deploy 10,000 nodes in 8 hours
Inspur recently delivered a shipment of rack scale servers of more than 10,000 nodes to the Baidu Beijing Shunyi data center within 8 hours.
How HCI helps enterprises stay on top of data regulations
Increasing data protection requirements will supposedly drive the demand for Hyper-Converged Infrastructure solutions across the globe.