The importance of cabling infrastructure design in the data centre of tomorrow

By David Kozischek, Ray Barnes and Keith Sullivan

The data centre is the data processing, routing and storage centre in the network. If the data centre is to the network what the brain is to the body, then our own neural network is to our brain what the cabling infrastructure is to the data centre. Increased efficiency is becoming more and more important as data-intense applications come online. The workload that the brain of the network is required to manage is accelerating. With this, data centre rack space, cable pathways, cooling, installation, maintenance and risk avoidance become increasingly important considerations when planning and designing the data centre. Today’s data centre cabling infrastructure must be designed to offer a competitive advantage and a lower cost of ownership immediately. As a result, the evolution path to higher speeds such as 40G and 100G data rates, as well as next-generation applications such as cloud computing and virtualization, must be ensured early in the data centre cabling infrastructure design and planning process.

“Given the challenges and issues facing the data centre today, data centre designers should install a cabling infrastructure that is ultra-dense, scalable, flexible and reliable –then they can rest assured that their data centres will evolve and grow to meet both today’s challenges and the ones to come.”
-Keith Sullivan, Marketing Director EMEA, Corning Cable Systems

A poorly designed or installed network cabling infrastructure can degrade network performance. Still, when network cabling infrastructure only represents 5 percent of the total cost of the data centre, some designers may adopt the view that there are more important decisions to be made than which network infrastructure products to choose. In today's data centre, change comes rapidly with respect to software, computers and servers -every three to five years. However, the same is not true for the network cabling infrastructure, which is expected to last for 15 years or more. For this reason, proper product selection when building out the brain of the network is critical.

Essentially, there are five persistent and major issues facing the growing and evolving data centres of today: scaling; moves, adds and changes; cable management; network reliability and powering and cooling.

Scalability

The data centre has always been the major hub for data storage and processing in the network. Now, as cloud computing and virtualization become more than vapour-ware, there will be an increase in the data transportation, storage and processing requirements demanded of the data centre. Understanding these new applications and the system requirements they drive, particularly with respect to transmission speeds, is important to ensure that the data centre can evolve and scale to support new platforms and data rates.

The most successful data centres going forward will be those that incorporate a high degree of density, scalability and flexibility into their network cabling infrastructure.

Moves, adds and changes (MACs)

Network churn is a key concern for network designers and administrators. Workstation relocations and network reconfigurations all increase the strain on the network. Making MACs easy and manageable is at the core of the structured cabling approach within a data centre. During MACs, those products that allow for clear labelling and traceability of a circuit will lead to quicker MACs and a reduced chance of downtime related to misidentification. Additionally, highly-scalable, swappable and modular system components will ensure that MACs are completed quickly and efficiently with minimal chance of disruptions to adjacent circuits.

Cable management

Today, data centres are growing and maturing, and cable management is becoming an important consideration. Proper cable management allows the fast identification of all system connections both at the physical layer and at the documentation level; good cable management also improves the maintenance of hardware equipment and the installation of new one; and finally, structured cabling management facilitates efficient cooling of the data centre.  

Reliability

Perhaps one of the biggest concerns facing the data centre administrator is system downtime. System downtime or degradation can lead to huge yearly losses. For instance, PayPal processes every second close to 2,000 US Dollars in payments. A one-hour downtime represents up to 7.2 million US Dollars in lost transactions. PayPal's several hours outage in the summer of 2009 ate up tens of millions of US Dollars.

Broadly speaking, there are three major causes of system downtime: equipment failure, poor design and human error. Having adequate business continuity plans and disaster recovery strategies in place; selecting first-class suppliers that offer good quality equipment; and   using technologies like RAID (Redundant Array of Independent Disks) will help minimize equipment failure. A good design of the data centre (modular, standardized) means planning ahead factors like the data centre growth path for, floor layout, electrical system design, mechanical and design, energy consumption, power back up, etc.

Continuous moves, adds and changes (MACs) are more than often the cause for human errors. A good design will reduce the need for MACs, but will not totally eliminate them as the data centre expands. A properly designed cabling infrastructure will minimize the chance for human error and thus network downtime. One aspect of data centre design that will help mitigate the risk of downtime is the structured cabling approach to deploying network cabling infrastructure. This approach allows for clear administration points in the network that allow for quick and efficient re-routing of network traffic in the event of an outage or degradation of service.

Powering and Cooling

Data centres have doubled their power consumption over the last five years. While IT equipment uses the bulk of power required by the data centre, 25 percent of the power budget is used for cooling in the data centre.

Customers have identified powering and cooling as critical issues in the design and operation of a data centre. In order to minimize the impact of network cabling infrastructure on this major cost, data centre solutions needs to be designed to free up cabinet and cable tray space to help improve airflow, thus reducing cooling costs.

Summary

Finally, given the challenges and issues facing the data centre today, data centre designers should install a cabling infrastructure that is ultra-dense, scalable, flexible and reliable. Ensuring that the cabling infrastructure chosen will meet these requirements, data centre designers can rest assured that their data centres will evolve and grow to meet both today's challenges and the challenges that are just over the horizon.