5.1.2 Cloud-Based Virtualization
5.1.2 Cloud-Based Virtualization
In the past, most virtualization deployments were implemented on-premise in the local
datacenter. Today, with cloud computing, virtualizations may be implemented on-premise, in the
cloud, or a combination of both.
On-Premise Virtualization
In an on-premise virtualization deployment, hypervisors are implemented on high-end physical
servers in the datacenter. They are connected to the physical production network, which allows
other devices in the organization to communicate with the virtual machines running on them.
Cloud-Based Virtualization
However, there are other ways to implement virtualization. Many organizations are increasingly
moving their virtualization solutions from their local datacenter to cloud-based virtualization
providers.
● Hardware and software costs are dramatically reduced. Instead of purchasing expensive
equipment for the datacenter, the cloud vendor is responsible to provide all of the
hardware and software needed for the deployment. All you need to do is provide a basic
network infrastructure that allows Internet access.
● Administration overhead is reduced. Because most of the virtualization infrastructure is
moved to the cloud, a large staff of administrators isn’t required. Instead, the cloud
vendor provides most of the administration staff required to keep the virtualization
solution running on their end.
● Some computing risks are shifted to the vendor. For example, the service level
agreement you sign with the cloud provider should include some type of guarantee in
terms of availability. This means the cloud provider is responsible for putting
redundancy mechanisms in place to ensure your services and information remain
available in the event of a disaster. In addition, the cloud provider should guarantee that
they have implemented the appropriate security controls to prevent data breaches and
malware infections.
● The datacenter can scale up and down dynamically as needed. In a physical datacenter,
scaling up means buying, installing, configuring, testing, deploying, and monitoring
additional virtualization components. This can be costly in terms of hardware, software,
power, space, and administrative overhead. In addition, if your computing needs scale
back in the future, you must determine what to do with the surplus hardware and
software that are no longer needed. In a cloud-based implementation, you can quickly
and easily purchase additional capacity when you need to scale up. When you need to
scale back down, you simply tell the vendor you don’t need the extra capacity anymore.
● Provisioning speed is enhanced. In a cloud-based deployment, pre-configured VMs can
be deployed within a few minutes. When you’re done, you simply de-provision them.
This deployment model dramatically reduces the amount of hardware, power, and space
required in your on-premise datacenter. It also significantly reduces your administration costs
because a smaller number of in-house administrators are required to manage the deployment.
Instead, the cloud provider is responsible for providing most of the administrative overhead.
However, this type of deployment also results in an almost complete loss of physical control of
your VMs and the information they contain. Their physical location may be unknown. In addition,
you are dependent upon the cloud vendor to protect your VMs from disasters and security
incursions.
This type of deployment costs more in terms of hardware, software, and human resources than
the full deployment because there are hypervisors running on-site that need to be deployed,
monitored, cooled, and managed. However, it also lets the organization have more control over
their VMs.
Using this model allows you to take advantage of the concept of cloud bursting. When the virtual
machines and hypervisors deployed on-premise become overloaded, some of the workloads
can be offloaded to resources provided by a cloud vendor. When demand return to normal,
those workloads are migrated back to on-premise equipment.
There are many other smaller providers that you could choose to use as well. Compare your
organization’s requirements against the products and services these vendors offer and decide
which one is most closely aligned.