Of all the applications that are unique to the cloud, none is more intriguing to stressed-out data center managers than cloud bursting. The comfort of knowing that, no matter what hits you, there will always be enough resources to handle the load is truly invaluable.
But before we get too caught up in visions of fully automated, dynamic cloud environments in which the infrastructure takes care of itself, note that bursting technology is still in its infancy, and it requires fairly sophisticated technology and processes to bring all these moving parts together.
The first piece of the puzzle, of course, is the hybrid cloud, which itself requires a thorough conversion of legacy infrastructure into a private cloud. For many organizations, that will take some time, although the ball is definitely moving in that direction. Research and Markets predicts the global hybrid cloud market will expand more than 30 percent per year to 2018 at least—and cloud bursting is only one driver in this trend. IT executives are also salivating over the prospect of critical application support in the cloud, as well as streamlined infrastructure, improved resource utilization and, of course, lower operating costs.
When it comes to bursting, however, Microsoft’s Brad Anderson says the hybrid cloud addresses one of the key requirements of proper functionality: system compatibility. With both local and third-party infrastructure on a common footing, the data hand-off should go smoothly. Of course, application components running on cloud resources will still need access to secure data on home systems, so you’ll need to make sure your dev/ops platform is compatible with your cloud. As well, the issue of conflicting APIs, policies, UIs and other components must be considered along with the fact that all of this must be brought under the scope of an overarching automation system—unless you plan on paying technicians to stand by in case traffic suddenly spikes.
Let’s not overlook the importance of a robust load balancer either, says Adam Lawson, CEO of IT architectural and engineering firm Aqorn. Remember, when you burst onto the cloud, it’s not like you are given hard assets of your very own. So at some point, those resources and the data they contain will have to be recouped, which could lead to performance issues if not handled properly. A load balancer helps to, well, balance the data load with available resources, effectively reducing the need to burst large amounts of data except in the most extreme of circumstances. In the bargain, the enterprise is able to maintain a more steady cost structure across local and cloud resources.
Still, the biggest danger in implementing a cloud bursting strategy is converting a capacity problem into a complexity problem, according to infrastructure specialist Nick Hardiman. Even with modern solutions in place, problems remain with systems communication, application architectures that are not designed for bursting, and the need to maintain adequate network infrastructure. And at the very least, it will require a fair amount of retraining and professional support to ensure affected systems and data loads are properly handed off from one set of resources to another. At some point, organizations need to determine whether simply building extra capacity into legacy architectures is the cheaper way to go.
If current trends continue, however, the cost equation will strongly favor the cloud going forward. If retaining staff and rewriting applications is what it takes to produce a more streamlined and more adaptive data infrastructure, then that is likely to be money well spent. And as noted, bursting is only one of the advantages of adopting a cloud architecture.
Peace of mind, after all, is a priceless commodity.