Virtual servers are not right for every situation. Virtualization works best for applications that utilize less than half of a server’s resources. It generally doesn’t work for big transactional database applications. But once you’ve decided to deploy virtualization, a strong implementation methodology will help ensure a smooth deployment.
Server virtualization is a simple concept based on the fact that most servers are grossly underutilized, representing a huge waste of IT resources. Virtualization technology uses a hypervisor—software that transparently doles out server resources and prevents resource conflicts—to enable multiple applications to share the same physical server, which greatly increases utilization. Although the concept is simple, there are some challenges when it comes to implementation that should be addressed:
1. Workload capacity and performance planning
With one application per lightly utilized server, capacity and performance planning aren’t usually an issue. Virtualization technology, however, greatly increases the utilization. Now you have to consider workload capacity and performance. Not every application should run on a virtual server. You need to look at things like present and future usage and workload patterns.
2. Systems architecture and design
Considerations here range from overall budget issues to system component selection decisions to assessing current and future business needs and mapping them to the IT infrastructure. How much CPU? How much storage capacity? How much network bandwidth, and the related cost of each component? These are all questions that must be assessed.
3. Operational processes and procedures
Virtualization technology changes with way people do their jobs and the management tools they use. Make plans that account for how the virtual servers are going to be monitored and managed. Also, pay attention to how the new IT environment will be tuned and optimized to improve performance or reduce costs or both. Mentoring and skills transfer should also be considered.
4. Security
Identify and address system vulnerabilities at the very start. Any vulnerability will be magnified in a virtual environment with each virtual instance becoming a new potential vulnerability or threat. This entails everything from security patching to access control.
5. Storage allocation and management
Take care to estimate and allocate storage so as not to exceed available capacity or constrain performance. Consider everything from transaction rates to capacity usage to application booting.
The key is to consider these issues before implementing the virtualization strategy. Avoid pitfalls like skimping on CPU horsepower or underestimating storage usage, which can hinder the virtualization effort. A smart, methodical implementation will help you realize the many benefits of virtual servers.
