The movement to private clouds represents an important trend in application delivery as server virtualization gives...
way to more robust cloud environments. Even so, a fair amount of confusion still exists about the differences between these two technologies.
Cloud computing has become an integral part of many organizations' operations; IT departments implement private cloud architecture in their own data centers or on platforms such as Amazon Web Services or Microsoft Azure.
From server virtualization to the cloud
Server virtualization has become the mainstay of application delivery in the enterprise. By abstracting physical compute resources, virtualization makes it possible to use hardware more effectively, isolate workloads, secure applications and manage operations. In fact, the use of server virtualization to run business applications has become so widespread that bare-metal deployments are now the exception rather than the rule for running enterprise workloads.
The continued evolution of the data center has led to a certain level of confusion about the differences between virtualization and cloud computing, even though they remain two distinct technologies. Cloud computing uses virtualization, but conceptually, they are as similar as a car is to its wheels. The wheels provide a way to interface with the road, but it's the entire vehicle that makes it possible to maximize the wheels' benefits.
Many organizations turn to private cloud architecture because it provides more control and flexibility over public clouds, as well as greater data security. The individual organization designs and deploys the cloud specifically for its own needs. Organizations can implement the platform in their own data centers or host it with a third-party private cloud provider to realize many of the same benefits available to the public cloud, such as automation, self-service provisioning, metered usage and rapid scaling.
Private cloud architecture
At a conceptual level, the infrastructure that makes up the private cloud is essentially the same as the public cloud. The differences come down to where the infrastructure is deployed and who controls it. As with the public cloud, the private cloud abstracts and pools resources to provide scalable, on-demand services that are logically separated from the underlying physical components. Admins can provision resources automatically or they can make resources available to business users through self-service tools.
This is a much different model from server virtualization, which focuses on VMs. Although both virtualization and cloud computing abstract hardware resources, virtualization provides dedicated resources through one or more VMs, whereas cloud computing moves beyond the boundaries of the VM to provide resources as pooled services across all systems.
Another way to look at this is that server virtualization takes one server and turns it into multiple environments. Cloud computing, on the other hand, takes multiple compute, storage and networking systems and turns them into consolidated resource pools.
Although virtualization can exist without the private cloud, it's rare for a private cloud architecture to exist without virtualization. Most private clouds rely heavily on virtualization to help better use resources and isolate workloads. In fact, virtualization provides the foundation on which most private clouds are built. The software that manages and orchestrates the cloud environment works directly with the virtualization software to provide an integrated platform.
Private cloud management
VMware offers its vCloud Suite of products to virtualize resources and build private clouds. The suite includes vSphere to virtualize compute resources, NSX to virtualize network resources and vSAN to create shared storage pools. In addition to these components, vCloud Suite includes vRealize, a cloud management platform natively integrated with vSphere, NSX and vSAN. In addition to managing the cloud infrastructure, vRealize automates operations, allocates resources and monitors systems, using virtualization extensively to provide a unified cloud platform.
OpenStack, an open source cloud OS and platform that provides the components necessary to deploy, manage and orchestrate a complete private cloud architecture, is another option. At the center of the OpenStack ecosystem is Nova, a cloud management platform that works with one or more hypervisors to provision compute instances and manage resource pools. The OpenStack project also includes components to implement and manage storage, networking, directory services and application messaging.
In both vCloud and OpenStack, server virtualization plays a pivotal role in delivering cloud services. For most private clouds, virtualization is essential to delivering an effective cloud infrastructure. Although cloud computing doesn't require virtualization, it can be instrumental in enabling the cloud software to more effectively use and pool resources.
Server virtualization might be enough to meet some organizations' application delivery needs, and they have no reason to take on the additional complexities that come with the cloud environment. Server virtualization is a tried-and-true method that has proven an effective strategy for many organizations. However, for those organizations that want to extend their data centers beyond the limitations of the VM-centric model and increase automation and on-demand capabilities, a private cloud architecture could be the ideal environment if they're willing to commit the additional resources.