The cloud has been invariably called an expanded data tier, an extension of the virtual environment and an entirely new, infinitely scalable data center infrastructure.
But no matter how you describe it, the fact is that the cloud represents a wealth of newly provisioned resources waiting to be filled with all the data and applications currently housed in your bricks-and-mortar data center. And in that sense, the cloud requires a lot in the way of plain old data migration − the same task that already gives you fits when it involves a new server deployment or storage array, except more of it.
Unfortunately, migrating data and applications to the cloud requires a different process from traditional hardware-to-hardware solutions. The cloud is a much more fluid environment − like a big bowl of soup on a wobbly bistro table. Once data hits the bowl, it's awfully hard to keep track of it.
That's one of the reasons the front office is still not entirely sold on the cloud even as IT salivates over all that additional capacity. According to a survey by Symantec, less than half of CEOS are willing to entrust mission-critical applications to the cloud, citing reliability, security and performance as their top concerns. The cloud represents significant improvements in both capital and operating costs, particularly as enterprises struggle to maintain control of increasingly large and complex data loads. However, the transition from traditional data center infrastructure to the cloud can be extremely disruptive if not handled properly.
One of the chief concerns is the fact that the cloud is built on virtual resources, as opposed to bare metal access. This means a fair amount of preparation is needed to ensure data availability.
"The enterprise administrator has limited options for migrating existing workloads to the cloud and must rebuild them package by package," says James Strayer, vice president of product management and marketing at system portability specialist Racemi. "This increases their costs and their risk by introducing manual effort and multiple failure points, creating a barrier to entry for enterprises. Some cloud providers are looking to help by providing support for importing virtual machines, but that assumes that, number one, you are virtualized and, number two, you are running the same virtual platform as the cloud provider."
Racemi addresses this problem through an image-based provisioning system on its DynaCenter automation platform. The idea is to migrate the entire server stack from the data center to the cloud while at the same time drawing out the necessary drivers, tools and other components needed to operate within the target resource. In this way, it makes no difference whether the data or application is being redeployed into a physical, virtual or cloud environment
Oddly, it seems that migration to the public cloud will emerge as the lesser of two evils compared to migration to a private cloud where the responsibility for ensuring compatible environments is borne completely by the enterprise.
"If you decide to go to an external provider and have done the diligence to justify the move, then you will relinquish the responsibility for the infrastructure, in which case you will not be concerned about data migrations," says Marie-Pierre Belanger, vice president of product management at SANpulse. "If you are moving it to a private cloud, the need to build, measure, report and break down on-demand infrastructures requires versatility and quick and easy configuration management with analytics that gauge utilization to be able to charge back the respective business units."
Naturally, this will require a fairly sophisticated level of automation, considering the very fluid nature of the cloud.
"Migrations are notorious for being a manually intensive and very complex process that is subject to errors as spreadsheets are created, maintained and changed," Belanger says. "Because organizations are not static, discovery and mapping of information takes place while it is active and subject to change. For this reason, migrations are perhaps one of the most risky operations a storage professional will experience."
Source : itbusinessedge
