Getting physical

The process of obtaining high efficiencies from data centre structured cabling involves a number of steps and IT managers in the region are well-advised to keep all of them in mind. Sathya Mithra Ashok finds out more.

Tags: CableExtreme Networks IncorporationInfrastructureReichle and De-MassariUnited Arab Emirates
  • E-Mail
Getting physical Many Middle East enterprises continue to look at short term solutions. - Andrew Sedman, technical director MEA for R&M.
More pics ›
By  Sathya Mithra Ashok Published  October 10, 2009 Network Middle East Logo

It is a fact of life. Achieving optimal performance from the physical layer is essential not only for ensuring equipment efficiency inside a data centre, but also for the overall health of the organisational network.

However, getting the most out of the physical layer is easier said than done, and the process often starts with the choice of the solution that is to be deployed in a data centre.

“When choosing cabling solutions for the data centre, IT managers should keep in mind requirements, architecture and the media. Requirements include apps that users plan to implement today and those that they would migrate to in five years time. Data centre architecture also greatly impacts cable selection. First, there is network switching (top of rack switching, distributed zone switching or a centralised switch architecture).

Then we need to understand where the cabling pathways run. Requirements and architecture drive the type of media selected and define the best option between field installed versus factory terminated systems,” says Dave Harney, group product manager at Molex.

Gautier Humbert, sales manager in the Gulf region for structured cabling solutions at Ortronics and Legrand adds: “There is a common misunderstanding that end-users should always go for the highest performance solution. Actually, the single most important factor to keep in mind is flexibility.

Choosing a reliable product is the basis. Many solutions are out there, but only a few are really applicable to data centres. Once the brands are selected, it comes down to design and to products that allow the data centre to adapt to the changing needs of an organisation. Basically, IT managers have to choose an adapted product, and consider a design that will allow easy modifications in the near future.”

Besides looking for pure functionality in a solution, vendors recommend that IT managers look for quality, adaptability of the systems and to choose a company that is more of a partner than a supplier.

“A data centre can be designed in many different ways depending on size and redundancy. But sometimes a single item can make the difference. For example, a pre-terminated solution can save tremendous time in installation and MACs (moves, adds and changes),” warned Humbert.

Apart from the basic guidelines to keep in mind, there is always the perennial question of whether to go for copper or fibre inside the data centre. Nobody denies that there has to be a mix of both solutions for optimal performance inside an organisational data centre. However, not many can agree on how much and what kind of a mix makes for the ideal solution.

“Although fibre is a more future-proof solution with regard to capacity, cost is higher and handling fibre requires more attention to prevent micro-bends and damage. Copper, on the other hand, is of lower cost and is easier to handle. When distances are larger than 100 metres, there is no other alternative than using fibre.

Therefore, copper is probably the best solution to be used in the data centre rack/row, whereas fibre is required in the data centre backbone,” stated Aziz Ala’ali, regional director, Middle East and Africa at Extreme Networks.

“Selecting between copper and fibre is a balancing act between applications support, the number or connections you need to support in an area and available budget.  Fibre will be the best cable type in high density environments where the user expects a quick migration from 1Gbps networks to 100Gbps networks.  But as of today, copper network transceivers are less expensive than optical transceivers, and so copper still has a place in most data centres,” explained Harney.

Andrew Sedman, technical director MEA for R&M says: “Typically from the main distribution area (MDA) to the zone distribution area (ZDA), purely fibre is used due to high traffic demands. Then from there out to the equipment distribution areas (EDAs) the choice can be made between copper and fibre.

Much has been said about using only fibre throughout the data centre as this can reduce the power requirements by up to 40%, but through design (i.e. keeping copper runs relatively short) similar savings can be achieved negating the ‘green philosophy’ concept of removing all copper cabling.”

As Humbert puts it, all data centres will continue to require a mix of both copper and fibre solutions, and IT managers should keep that in mind when deploying the physical layer in their various data centres.

Add a Comment

Your display name This field is mandatory

Your e-mail address This field is mandatory (Your e-mail address won't be published)

Security code