A different cause for cloud outage has been just discovered: Lightning storm. Amazon and Microsoft cloud services went offline for several hours as a lightning strike hit a power source that support Amazon and Microsoft’s cloud computing infrastructure in Dublin, Ireland on August 8, 2011.
The services are gradually restored several hours after the force majeure. Amazon reports that within 12 hours 60 percent of EC2 instances have been restored. As of Microsoft, the BPOS services in Europe were able to be restored within four hours time.
I would say both are doing a great job in communicating the cloud outages to the users – something that they were lacking in the past.
The cloud outage impacted Amazon and Microsoft cloud services in Europe only, but the business impact of the lightning strike is widespread, as the data centers in Dublin, Ireland also serve non-Europe companies that do business in Europe.
New lessons learned
Technically speaking, the cloud outage should not be happening due to power failure as backup generators should automatically pick up the electrical load that is previously provided by the main power generator.
However, the Dublin lightning strike causes explosion and fire that are large enough to disable part of the system that will activate backup generator plant when required. So the backup generator plant was activated manually – thus the several hours of cloud service outage.
It is said that during a natural disaster (such as the Japan earthquake) the cloud helps businesses to secure their important business data, information and operations. However, we should expect service interruptions to happen during a disaster as – whether you like it or not – cloud computing is powered by data centers and power generators that present physically (thus directly exposed to natural disasters.)
One thing for sure: There will be more research and development for better ways to activate backup system – so, stay tuned, and don’t give up on cloud services just yet!