In the wake of the Japan earthquake and tsunami, Nirvanix, Equinix, and other cloud companies are offering a helping hand to companies disrupted by -- or concerned about -- natural disasters.
Amazon Web Services isn't making public offers of assistance, but behind the scenes it has told members of its Japanese user group, JAWS-UG, that its resources are available, Amazon spokesmen acknowledged. There were reports that AWS offered guarantees of free scalable compute power to the Red Cross in its efforts to air information and to both the Red Cross and the impromptu Save Japan fund at the South by Southwest Conference in their efforts to collect donations. Amazon spokesmen wouldn't comment, saying the company seeks no publicity as a result of the disaster. By coincidence, it announced that it had opened data centers in the Tokyo area in early March and spokesmen confirmed that they remain in operation. The region became EC2's fifth data center concentration, after Amazon US-East in Northern Virginia, Amazon US-West in Washington State, its European centers in Dublin, Ireland, and an existing Asian data center in Singapore.
Under some circumstances, virtual appliances or virtual machine images of existing workloads can be created in the enterprise data center and stored in a cloud data center. In the event of a failure of the former, the virtual machines serve as recovery mechanisms that can be reactivated in the cloud.
In a conversation with SunGard's Indu Kodukula at Cloud Connect 2011, a UBM TechWeb event, he told me that 90% of SunGard customers are now either using -- or at least experimenting with -- the virtual machine re-activation at a remote site as their disaster recovery mechanism. Previous disaster recovery consisted of near look-alike hardware installations in separate locations, with identical software stacks. Keeping the two sites in synch was difficult, and a full test of the system was awkward to implement, since it meant shutting down production systems in the hope that they could be quickly reactivated in the recovery location.
"Every one of our customers is running (as a test) at least one production system in the cloud," said Kodukula, which at the moment meant a SunGard data center facility running VMware ESX Server virtual machines. In the future, the practice may become widespread across all virtual machine users, tapping into different cloud facilities, as customers choose.
One other note: Telehouse America, a supplier of managed IT services in the United States and Japan, says its 21 data centers in Japan, including its Sendai facility in the tsunami affected area, are all operational. As the three-hour mandatory blackouts hit its data centers, they will resort to their backup power generators, which have a 24-hour fuel supply with the ability to refuel as needed. The Amazon and Equinix data centers in all likelihood operate on a similar basis.
Disasters will continue to occur. Indeed, my home area of San Francisco and the Silicon Valley, allegedly on a 100-year cycle for big earthquakes, is overdue. Its nerves are on edge after three large quakes have struck on the ring of fire around the Pacific Rim. They include Chile on Feb. 27, 2010 (8.8 magnitude), Christchurch, New Zealand, on Feb. 22 (7.1), and Japan on March 11 (9.0). The Great San Francisco earthquake of 1906 has been estimated as a 7.9 event. (This list neglects the 7.0 quake that hit Haiti on Jan. 12, 2010, not on the Pacific Rim.)
Before the next disaster strikes, think about how reliance on a second data center in the cloud might provide a guarantee of continued operation, if the unthinkable happens and your primary systems are disabled or destroyed. And think about extending a helping hand to the next guy, if you're left standing up after a disaster.
Multicloud Infrastructure & Application ManagementEnterprise cloud adoption has evolved to the point where hybrid public/private cloud designs and use of multiple providers is common. Who among us has mastered provisioning resources in different clouds; allocating the right resources to each application; assigning applications to the "best" cloud provider based on performance or reliability requirements.