Definition

The traditional architectural model for IT infrastructures has been physical. For example, there is a one-to-one relationship between applications and the servers upon which they execute. The purely physical approach presents a pair of challenges. First, as the business grows, so typically do the number of applications, servers, databases, and storage devices, along with the underlying network. Second, because of the one-to-one relationship, physical resources often are vastly underutilized. Most independent studies have shown that CPU utilization, for instance, lies somewhere in the range of five to fifteen percent, resulting in significant unnecessary capital and operating expense.

Virtualization deployment transitions the IT infrastructure from one that is purely physical to a more logical architecture model. Virtualization involves the creation of a virtual, or logical, version of a resource, be it a hardware platform, operating system, storage device, or network element. Physical resources still are required to host the guest logical resources. However, because each physical resource is multi-tasked or otherwise logically connected with its peers, far fewer physical resources are necessary.

Applications

Almost any resource in the IT infrastructure is a candidate for a virtualization deployment. The types of resources where virtualization is becoming increasingly widespread include:

Hardware – The virtualization deployment of hardware devices like servers is perhaps most common. A single physical server is the host for a number of guest virtual machines, each with its own applications and operating environment. The host and the guest platforms are separated by software or firmware known as a hypervisor, which handles any communications between physical and logical elements

There are three types of hardware virtualization deployment:

  • Full Deployment – The hardware of the virtual machine is completely simulated, allowing the guest operating system to run unmodified.
  • Partial Deployment – Only some of the virtual machine’s hardware is simulated, so the guest operating system may have to be modified before it can execute.
  • Para Deployment – None of the virtual machine’s hardware is simulated, but guest software is fully isolated, so it can run if specifically modified for this environment.

Software – A single operating system instance can host multiple virtual operating systems. Similarly, application virtualization is achieved by introducing a layer of abstraction between the hosted applications and the underlying operating system.

Memory, Storage, and Database – In each case, distributed physical elements, which may have limited availability, are logically consolidated to appear as a larger, contiguous, single element with greater power to promote efficient execution of activities.

Network – Network virtualization deployment can involve the creation of a virtualized network addressing space within or across subnets, or the deployment of virtual network elements, such as switches, which represent an aggregation of multiple physical network elements.

Key Considerations

Virtualization deployment promotes resource consolidation, which can help distributed enterprises eliminate remote locations and/or simplify their datacenters. Consolidation brings a variety of benefits, including:

  • Reduced Capital Expenditures – Fewer physical resources means less upfront investment in items like servers, storage devices, switches, and routers.
  • Reduced Operating Expenditures – With virtualization, ongoing maintenance costs drop. Because less space is required, overhead rates drop. And fewer physical resources consume less energy.
  • Improved IT Efficiency – With a smaller physical footprint, IT can react more quickly when problems arise, assume a more proactive posture to prevent issues from occurring, and focus on the highest priority tasks of the enterprise.

In addition to these advantages, virtualization also eases scalability and disaster recovery concerns. Because enterprises are no longer tied to the one-to-one physical IT infrastructure model, they can respond more rapidly to dynamic business conditions, whether driven by growth or unanticipated resource availability, by initiating a new virtual resource.

Although virtualization requires fewer physical resources, in many ways, the operating environment becomes more complex, which can make it more difficult to monitor and manage. For example, virtualization can bring CPU utilization up to approximately 80 percent, so traffic volumes and bandwidth utilization rise, latency also may rise, and traffic management is less straightforward.

Enterprises can capture all of the upsides of virtualization and negate the potential pitfalls by deploying a performance management solution that:

  • Offers the breadth to account for physical networks, servers, and applications throughout the enterprise.
  • Has the depth to examine individual transactions, be they within an application or between network elements, in the physical and logical domains.
  • Collects information in real-time in support of trending and thresholding, and stores information over time in support of historical analyses and pursuit of intermittent anomalies.
  • Accounts for different classes of service, qualities of service, and service level agreements, which become more critical when virtualization is implemented.