As organizations grow and evolve, they must implement technology changes to accommodate evolving infrastructure needs, often within complex systems running business-critical applications. Along with this, there frequently is an increased demand to reduce the costs of technology by sharing hardware and software resources, a demand many companies try to meet by establishing virtual environments. Virtualization balances the often underutilized physical machines by consolidating them into a single physical host that runs multiple Virtual Machines (VMs) sharing the four core resources-CPU, memory, disks and network cards-of one physical host.
Effectively managing applications that leverage a virtual infrastructure can be an elusive goal, however. Server virtualization has brought significant efficiencies to IT operations, but also has produced a host of new challenges for IT personnel. These challenges include monitoring the health of virtual machines, analyzing data, alerting staff to problems, and performing various administrative tasks. They are compounded by the fact that the tools used to manage them in a physical environment do not always function as effectively for virtual machines
While the benefits of virtualization are many-increased business continuity, disaster recovery, flexibility and agility, as well as reduced downtime, utility costs and data center space-many organizations don't fully realize them because their virtual infrastructures are hampered by both the new set of challenges and poorly performing applications that fail to meet service levels and the needs of end users.
A further complication for many organizations is the price tag for implementing a virtual environment. The switch from a physical environment to a virtual one involves significant, often hidden, costs which frequently are unanticipated when the initiative is undertaken. In fact, the total overall cost of the virtual environment is most impacted by decisions made before the virtualization initiative is even begun.
The keys to a well-developed virtual environment that contains costs, provides significant ROI and achieves optimal performance are good planning combined with a set of concrete techniques for ensuring your virtualization initiative effectively supports your applications and end user satisfaction objectives.
We'll take a look at cost-effective planning for a virtual environment later in this article, but first, let's discuss the five techniques that will effectively support applications and end users. You can better address the operational challenges of virtualization if you have a good understanding of the virtual infrastructure, and if you manage the relationships and interactions between all components in the environment. Here are the five techniques:
Better understand how the virtual environment impacts the entire application.
Visualizing the entire virtual infrastructure in a single view allows you to clearly see the data centers, data stores, clusters, resource pools, ESX servers, VMs and other resources concurrently in use across the many layers of the environment. Having the ability to correlate events occurring in the virtual infrastructure with the entire application environment will help you determine the root cause of incidents in the virtual environment and reduce mean time to resolution for incidents and problems.
Determine the root cause of an incident or problem before end users are affected.
The key to determining the root cause of an incident or problem is expertise. Thus, IT needs detailed alarms with recognized industry expertise covering best practices, future predictions, deviations from normal activity and specific operational problems. Tools that help you determine a problem exists, convey why it's a problem and recommend ways to resolve the problem are ideal for diagnosing and resolving issues before they impact end users.
Track movement of VMs to understand their potential impact on applications and the business.
Organizations need to be able to track assets, in terms of changes to both configuration and location, so they can assess the impact of those changes on performance and availability of the dependent application, as well as other VMs in the same physical environment. By tracking the movement of VMs, you can better understand their impact on applications and end users, and determine what happened and why.
Contain alarm storms from VMs and physical servers for a prioritized IT response.
By effectively containing alarm storms and turning data into meaningful information, IT can better understand the correlation between infrastructure, host and VM issues. This is an essential tactic for prioritizing problem-resolution efforts and preventing application performance issues that can lead to poor end user experience.
Identify resource contention between VMs to prevent over-commitment of resources.
Organizations need to be able to show the resource impact of moving a VM image from one physical system to another. With this information, they can determine in advance whether planned moves will unfavorably impact resources, applications and end users. Also, performance and utilization data gathered at several levels-infrastructure, host and VM-will help identify contention issues. A clear understanding of core resources and historical trend information, will allow organizations proactively plan for the future and prevent problems.
Avoiding the Hidden Costs of Virtualization
Now let's take a look at how you can implement your virtual environment at a reasonable cost. In many cases, the quick growth of virtual machines in an enterprise occurs without any planning at all and can quickly spiral out of control, so it's important to understand the importance of planning and how decisions you make before you even start building it will impact the total overall cost of your virtual environment. There is a lot to think about when you are planning a move into a virtual infrastructure, and the three most important considerations are server workloads and candidate selection, operations and administration, and the impact on applications and end users.
Determination of server workloads and candidate selection should come first. Remember that a virtual environment should allow you to be more proactive than reactive to avoid any impact on end users. Create the standards for your virtual environment before virtualization begins, formulate "Virtualization First" policies and avoid the temptation to adopt virtualization too quickly. I cannot stress enough how important it is to take your time and go slowly if you are going to start out on the right foot with your virtual environment.
A virtual environment does come with significant costs. Hypervisor software is not free, and you will need new server hardware with a ton of memory. You will need hypervisor management tools and a lot more storage. The more the virtual environment grows, the more storage you need, and this is a primary cost around virtualization. Other financial impacts, especially in the beginning, include the cost of resources, whether internal or consultant-based, and the service and revenue impact of unexpected downtime when your IT staff is not familiar enough with the virtual environment to respond quickly when a problem occurs.
Make the right choices up front to avoid hidden costs and get the best ROI for your new virtual infrastructure. Be realistic about what you want to get from virtualization and the new hardware you will need. Acquire minimum hardware at the beginning, and don't make excessive licensing purchases before you determine whether virtualization is right for you.
Doing it Right
To successfully support applications, minimize any negative effects on end users and avoid spiraling costs, virtualization must be undertaken with great care and planning. By achieving optimal performance of the virtual infrastructure, knowing how and where to contain the costs of implementing a virtual environment, and simultaneously understanding both the environment and the applications that leverage it, organizations can gain the benefits of virtualization while effectively supporting applications and end users.
Effective management of your virtual applications and services now, and as your IT environment changes, is essential to aligning critical applications to business service goals, improving application performance, mitigating business risk and lowering IT operational costs.