Capitoline 2015 review of data centre failures published

Data centre outages analysed by cause and time impact over a five year period for cloud providers, colocation data centres and corporate data centres.
LONDON - July 13, 2015 - PRLog -- Capitoline Ltd, the independent data centre consultancy, maintains a rolling five year statistical analysis of catastrophic data centre failures around the world and has just published the 2015 review to subscribing customers.

219 complete loss of service failures were analysed across all the continents. Lead consultant Barry Elliott commented, “Average data centre outage was 12.5 hours per incident with the span being from 20 minutes to eight days. That excludes the six sites that were lost completely to fire or flood.”

Capitoline estimates that this equates to over $151 million in lost business and other costs to the companies involved.

Elliott added, “IT and communications problems are the biggest source of outages but if we extract that then power remains the most common reason of catastrophic failures accounting for 43% of all facilities-related outages.  Environmental issues such as storm, flood and earthquake came in at 6% but sometimes power and environmental issues can overlap where an event such as an electrical storm in turn provokes an electrical outage. 9% of catastrophic failures never had their root cause made public but it would be fair to conclude that at least a third of these would also be power related.”

Perhaps the biggest surprise was the rise in impact of serious fire events which account for 21% of catastrophic failures yet are responsible for 41% of all outage time once IT issues are removed. With so much forced air cooling in computer rooms, fire remains a significant risk and the poor quality of some fire detection and suppression systems has been revealed. Even more revealing is the lack of recovery plans practiced by most data centres that should allow for rapid restoration of services after what might have been a relatively minor fire event. Capitoline’s other research into data centre management practices and techniques shows that lack of fire management and recovery procedures is the area most likely to be lacking in the required suite of data centre operational management practices.

Capitoline has concluded that very few outages could be put down to unpreventable ‘acts of God’ and the vast majority could have been prevented with better thought towards location, design and operational management.

“Auditing designs, existing data centres and management procedures and the proper training of staff to major standards such as TIA 942 and the new European EN50600 standard would have saved these companies a great deal of time, effort, money and loss of reputation.”

Capitoline is an independent engineering consultancy specialising in data centre auditing, design and training and is the market leader in data centre auditing with over 100,000 m2 of floor space audited.

The full report is available at http://www.capitoline.org/why-data-centres-fail/

Author

Barry Elliott

Contact belliott@capitoline.org

Website http://www.capitoline.org

Contact
Capitoline Ltd
***@capitoline.org
End



Like PRLog?
9K2K1K
Click to Share