Tags


Latest Posts


Latest Comments


Authors

Fail to plan, plan to fail

stefan-haase.jpg

Posted by |

Power cuts at data centres have been in the headlines a lot recently. A couple of weeks ago, Amazon’s generators fell foul of extreme weather conditions in America, taking high-end clients like NetFlix, Instagram and Pinterest with it.

This week, it was London’s turn to experience a power-failure-related data centre outage. An uninterruptible power supply (UPS) problem saw some ISPs unable to provide connectivity to customers for over five hours.

The first thing to say about any of this is that accidents can – and do – happen. Data centres and the Cloud are transforming both the pace and flexibility of how we do business, but like all technologies, without fault tolerance and strong design, they remain vulnerable. But blaming the weather and/or faulty equipment is small consolation for your customers. It’s the contingency planning that you have in place that makes the difference.

Proper planning and investment in the right technology are key. Generators and UPS equipment are only as good as the quality of the maintenance plans supporting them – age, poor maintenance and over-stressing equipment in an effort to keep costs down can all result in ‘Plan B’ existing in name only.

Ill-advised cost-cutting measures are also seeing some data centres pushing capacity beyond sustainable levels; the latest design principles promote fixed capacity, ensuring that providers aren’t constantly looking to add extensions to their equipment, placing an excessive burden on resources.

High resilience data centres offer N+1 resilience for all electrical equipment. And there’s no substitute for planning – regular, rigorous testing in failover mode, testing both UPS and generator equipment should be standard practice. Waiting until there’s a substation failure to find out whether or not your generators pass muster is just asking for trouble. Similarly, staff training and regular testing of plant and switching procedures can mitigate the role that human error all-too-often plays in data centre failure.

Times are hard, but under-investment in equipment and qualified data centre staff puts customer services at risk. Continued investment and contingency planning aren’t optional extras when it comes to quality data centre provision.

If you’d like to take a tour of a state-of-the-art data centre, why not join us at our new Reading facility? We’re hosting an Open Day on July 19th.

Comments

 

Post a comment

Comment submitted! Comments needs approval before being displayed.