Japan is one of the worlds most prepared nations when it comes to disaster management. Everyone from government ministers down to the school children take this preparation seriously. Their buildings and infrastructure are designed to survive a range of natural disasters and they regularly carry out drills to ensure everyone is ready.
However, on the 11th March 2011, they suffered one of the worst natural disasters in their history. A mega-thrust earthquake with a magnitude of 9.03 located south east of the Oshika peninsula created a Tsunami which took the lives of over 15,000 people.
The earthquake also caused the loss of power at the Fukushima nuclear plant and the tsunami took out the backup generators which caused the second worst nuclear disaster in history.
So how did this happen in a country more prepared against disaster than any other?
In the inquiry that followed the disaster, the Fukushima plant operators explained that they had contingencies in place for an earthquake and a tsunami. They also had backup generators to support the cooling systems in the event of a mains power failure.
Of course, what they were not prepared for was for all of this to happen in a single event.
Hurricane Sandy was another example of a one in a million chance event for which there was no way to prepare. Several major data centre operators in New York city were taken offline by Sandy and their customers suffered extended loss of services in the aftermath.
It is very easy to look back on these events and see where things could have been done differently but it highlights a very real risk for data centre operators.
You cannot create a disaster proof data centre because there will always be a risk somewhere in the design. It doesn't matter how much money is spent on resilience, there will always be something that you cannot anticipate.
You can reduce the risk by duplicating your data centre somewhere else and replicating all of your services but this is an extremely expensive solution to the problem.
Why spend money on something that you don't need for 99.999% of the time?
More companies are moving towards cloud services as a solution to this problem. Their services are spread across multiple cloud providers which are then brought together into a single virtual data centre.
In the event of an incident at one site, the critical workloads can be moved out of harms way.
Large scale data centres are now becoming a liability by putting all of the critical services in one place which will inevitably succumb to disaster one way or another.
Cloud technologies offer a cheaper, more flexible solution to the problems of disaster management.
However, on the 11th March 2011, they suffered one of the worst natural disasters in their history. A mega-thrust earthquake with a magnitude of 9.03 located south east of the Oshika peninsula created a Tsunami which took the lives of over 15,000 people.
The earthquake also caused the loss of power at the Fukushima nuclear plant and the tsunami took out the backup generators which caused the second worst nuclear disaster in history.
So how did this happen in a country more prepared against disaster than any other?
In the inquiry that followed the disaster, the Fukushima plant operators explained that they had contingencies in place for an earthquake and a tsunami. They also had backup generators to support the cooling systems in the event of a mains power failure.
Of course, what they were not prepared for was for all of this to happen in a single event.
Hurricane Sandy was another example of a one in a million chance event for which there was no way to prepare. Several major data centre operators in New York city were taken offline by Sandy and their customers suffered extended loss of services in the aftermath.
It is very easy to look back on these events and see where things could have been done differently but it highlights a very real risk for data centre operators.
You cannot create a disaster proof data centre because there will always be a risk somewhere in the design. It doesn't matter how much money is spent on resilience, there will always be something that you cannot anticipate.
You can reduce the risk by duplicating your data centre somewhere else and replicating all of your services but this is an extremely expensive solution to the problem.
Why spend money on something that you don't need for 99.999% of the time?
More companies are moving towards cloud services as a solution to this problem. Their services are spread across multiple cloud providers which are then brought together into a single virtual data centre.
In the event of an incident at one site, the critical workloads can be moved out of harms way.
Large scale data centres are now becoming a liability by putting all of the critical services in one place which will inevitably succumb to disaster one way or another.
Cloud technologies offer a cheaper, more flexible solution to the problems of disaster management.