One of the hallmarks of business is pragmatism. Effectiveness, rather than dogmatic purity, is prized. That’s demonstrated in the way that businesses are approaching cloud implementations. More and more companies are implementing cloud computing in a hybrid fashion, mixing public cloud, private cloud, and on-premises applications and services to build application infrastructures that combine the best aspects of each architecture into an optimal whole.
With all of the different components, one of the critical issues is being able to manage the total application infrastructure as a single entity. That seems obvious, but the next piece of the puzzle might be less so. To manage the entire infrastructure as a single system, consider treating every piece of the infrastructure as a cloud.
Managing hybrids as single systems
In the early days of cloud computing, the tendency was to treat every cloud like a wandering on-premises system. It was understandable but didn’t get the greatest benefit from the system because it didn’t demand the most from the components. That’s the real danger in looking at any part of the overall architecture as anything but a cloud: When you compromise the performance of any component, you compromise the performance of the whole.
It is certainly easier to treat local, on-premises servers as cloud resources when:
- Physical servers are host to virtual servers
- The virtual servers are managed by a framework that places them within the application infrastructure along with public and private cloud resources
Components from each type of resource are spun up and spun down on demand.
The first question to ask a vendor is whether the collection of components used in a server is compatible with the application management platform that your organization has selected for its infrastructure.
Seeing local storage as a cloud resource
In addition to the basic server components, local storage should be seen as a cloud resource and managed as such for the best results. If all of this is starting to sound like a highly converged, if not hyper-converged, system then you’re catching on.
In many respects, the optimal approach to a hybrid cloud infrastructure is essentially the same as that for a hyper-converged on-premises system. This way, each component is treated as a resource that can be deployed or withdrawn automatically as the application requires.
Application performance is not the only aspect in which cloud management techniques make the most sense for a hybrid environment. A management framework that controls every component as a cloud-like service minimizes the number of “seams” between applications and services that can be exploited as a vulnerability by an attacker.
Many public and private cloud services offer their own management consoles to customers. But it makes sense to employ a single, unified management framework for an entire application infrastructure. Once that decision is made, treating local assets the same way you treat cloud assets can maximize local server effectiveness and minimize long-term cost and vulnerability.