Five ways to improve your virtualization ROI
Virtualization enables more flexible allocation of physical infrastructure resources, empowering IT to better leverage those resources and respond to constantly changing business needs. Some organizations achieve greater success than others when it comes to return on investment (ROI), however. Following the best practices outlined here will help improve data center economics by enabling smarter use of existing capacity, lower virtualization total cost of ownership and broader deployment of virtualization across the business.
1. Optimally allocating virtual resources to specific application workloads.
The benefits of virtualization arise from accurately aligning IT capacity with business need. To maximize those benefits, IT needs clear visibility into how application workloads are using their allocated capacity. This is not just a matter of watching an application’s utilization of its assigned virtual machines (VMs). IT staff need granular visibility into how specific application processes (lookups, retrievals, Web services and so on) are utilizing specific resources, especially processor time and machine input/output.
The reallocation of resources in response to shifting demand must also be easy. Wherever possible, reallocation should be automated based on appropriate workload and utilization triggers. An end-to-end virtualization stack allows IT staff to continually monitor the utilization of all virtual resources by all application processes, and easily allocate or de-allocate those resources as appropriate. The staff can also create rules for automating resource allocation in real time to handle sudden demand spikes and execute outage-driven failovers.
2. Unifying management from physical machine to application workload.
Management of enterprise IT environments already tends to be somewhat siloed. A virtualization layer can add yet another silo to these operations. As a result, IT organizations can find troubleshooting in virtualized environments difficult. The solution is to create a common view into the entire infrastructure stack that directly maps physical resources, virtual resources, OS processes and application processes to each other, so anyone working on one aspect of a performance issue can quickly see how it correlates to the other layers in the stack.
3. Choosing the right virtualization platform based on price, performance and manageability.
With several different hypervisor platforms available, IT organizations must choose solutions that allow the broadest and best possible use of virtualization. Decision makers should be particularly careful not to feel locked in to whatever hypervisor platform was used for an initial rollout. The hypervisor market has evolved rapidly in the past few years, and a platform that was useful for early adoption may not be the best one for a new or broader deployment. Today’s management tools also make it relatively easy to manage multi-hypervisor environments.
4. Documenting virtualization benefits for application owners.
At many organizations, data center managers must convince business application owners to agree to virtualization. To effectively champion virtualization internally, IT must do more than deliver adequate service levels; it must offer real-time visibility into the virtual resources allocated to critical application processes. Only with such visibility can application owners gain the same comfort level with virtualization that they had with conventional, dedicated infrastructure.
5. Accessing ready guidance for workload virtualization.
To scale up virtualization rollouts, most IT organizations must call on external knowledge resources. These resources can provide the guidance necessary to make the provisioning of VMs and the allocation of VMs to application workloads more predictable and less time-consuming. Through transfer of expertise, this knowledge can become integrated into in-house “runbooks” so that deployment of virtual resources becomes faster and easier over time. These external resources can also enable IT to keep abreast of changes in the market.
The goal is for IT to achieve the same comfort level with virtual deployment as it has with physical infrastructure. Once this point is reached, IT will be able to freely leverage virtualization across the enterprise wherever and whenever appropriate, without the constraints that currently limit deployment and ROI.
Selecting a hypervisor? Think about:
Price: The more expensive the hypervisor platform, the more its rollout will be constrained by budget limitations.
Performance: The right platform will provide optimum performance for targeted workloads, while consuming little processing overhead.
Manageability: A hypervisor that is easy for IT to administer and configure with existing skills offers advantages over one requiring specialized expertise.
--Bob Leck is an enterprise technologist at Dell, specializing in virtualization and cloud.