Many data centre owners and operators lay claim to certain levels of resilience and redundancy when it comes to their data centre infrastructure. However, only a few have taken the responsibility of obtaining objective and formal certification from recognised data centre certification authorities.
This is according to Lee Smith, director: data centre professional services and training at data centre training firm DS, speaking today at the ITWeb Data Centre Summit 2017. Discussing whether there is value to be derived from formal certification of data centres and their operations' capability, Smith pointed out certification ensures verification of a data centre's design, construction and functional resilience.
"Most certifications focus on the manner in which a data centre was designed and built. Very few also certify the operational capability of their data centres i.e. people, process, policies, documentation etc. Certification proves the site is able to perform as expected in the event of planned and unplanned incidents or outages and therefore should be conducted by independent, unbiased and recognised data centre industry authorities," he advised.
Data centre certifications, explained Smith, consists of either the Uptime Institute certification or the TIA (EPI) certification requirements, which uses ratings 1, 2, 3, 4 and also deals with data centres, electrical, mechanical, telecommunications and architectural certifications.
He asked why and when certification is important.
"If your business is a data centre-as-a-service, you should consider certification. Site certification has its merits but a business case is usually required - not everybody has the requisite in-house skills. You should also consider certification if clients request for it, as it provides a competitive advantage and will also provide market differentiation. Certification also assists businesses in getting an independent unbiased opinion on their data centre strategy."
Certification prices could range from US$90 000 - US$250 000,dependent on the certification authorityused, data centre size, redundancy levels and complexity, noted Smith.
"Organisations should also consider the costs of the professional teaminvolved in the design and implementation stages of the certification. There could also be costs involved due to non-compliance, design changes and subsequent delays involving people or teams who have done this successfully before."
Around 22% to 48% of data centre failures are due to human error, with root causes of unplanned outages being UPS system failure (24% -25%) cyber crime (22%), water, heat or computer room air conditioning failure (11%-15%), he revealed.
"There is no point in having a highly redundant or resilient data centre facility and then failing during the ICT procurement, installation and operational stages. A resilient data centre does not also imply resilient ICT and people, so it's important for organisations to consider the impact and adapt or respond to it as things change," he concluded.
Share