DatacenterDynamics by Country

Country missing? Please select your nearest region...

In the aftermath of last week’s 8.9-magnitude earthquake that struck near Japan’s northeastern shore and triggered a devastating tsunami, there have been few reports of major data center outages. The disaster did, however, cause substantial damage to infrastructure data centers rely on to continue service delivery: power and networks.

Power plants in the affected region were reportedly damaged as were multiple undersea cable network systems. Multiple utilities have implemented rolling black-out plans, during which data centers can be powered by back-up diesel generators. Network-infrastructure damage has slowed connectivity for some, but because the infrastructure is redundant, no complete loss of connectivity has been reported.

Global colocation provider Equinix released a statement on Sunday, saying its two data centers in Tokyo were operating as usual. Kei Furuta, managing director of Equinix Japan, said the facilities were running on grid power but there were concerns about power shutdowns in the Kanto region.

“We have fueled the generators at our Tokyo data centers to their full capacity, which will provide emergency back-up power in the event of any power disruption,” Furuta said. “We can keep our Tokyo data centers up and running as normal.”

While on-site fuel supply can keep data centers running during utility power outages, there is a concern about reliable delivery of fuel when on-site supplies run out. A source that spoke on condition of anonymity told DatacenterDynamics that at least one large telecoms provider had reported problems accessing diesel fuel in the Tokyo area.

Rolling blackouts
Furuta said that the country’s Ministry of Economy, Trade and Industry (METI) was working to prioritize energy supply to data centers, considering them critical to information and telecommunication.

METI released a statement on Sunday, saying power plants operated by Tokyo Electric Power Company (TEPCO) and Tohoku Electric Power Company had been “seriously damaged.” The statement went on, saying restoration of adequate capacity in the following several days was unlikely, with high possibility of a large-scale blackout.

To avoid large-scale outages, TEPCO implemented a rolling-blackout plan on Tuesday, 15 March. The plan consisted of multiple-hour blackout blocks in five regions. The power provider made similar implementation plans for every following day through Friday, 18 March. Tohoku Electric is reportedly taking similar measures.

Click here for detailed implementation plans of TEPCO rolling blackouts this week

Like Equinix, Amazon Web Services, which only recently announced expansion of its physical infrastructure into Tokyo, also released a statement saying functionality of its cloud-delivery platform had been unscathed by the disaster. “The AWS Tokyo Region is not affected by the earthquakes or tsunami,” AWS representative wrote in an emailed statement.

IBM spokesman Douglas Sheldon said that as of Monday, the company's data centers in Japan had not experienced any "serious major outages". He declined to comment on the company's plans for ensuring adequate generator fuel supply in case of prolonged blackouts.

Commenting on reports of damaged undersea cables, Equinix’s Furuta said instead of going from Japan to US directly, network traffic could be diverted to the EU before being sent to the US.

“There is degraded network but connectivity can still be guaranteed in the worst-case scenarios in a properly planned network,” he continued. “Since the 2006 earthquake incident in Taiwan, carriers have built more and more redundancy into the sub-marine cable system. So far, there is less impact than we have seen in the 2006 incident.”

He was referring to a 7.1-magnitute earthquake that struck off Taiwan’s coast in December 2006, damaging undersea cables and causing major communications disruptions across East Asia, according to report by The Register.

Martin Levy, director of IPv6 strategy at Hurricane Electric, provider of colocation and connectivity services that operates a global network, said one path on the company’s network in Asia was down, but the network was fully operational because of path redundancy.

"We are fully operational; but not as optimally as we would like," Levy wrote in an email. "Such is the nature of cable breaks and why we invest in redundancy. It seems that no operator has escaped this one."