Keith Kessinger, Assistant Site Editor
After a virtual infrastructure rollout, it's important to protect your investment by monitoring virtualization performance. There are striking differences between server resource management in traditional and virtual infrastructures.
Virtualization creates a layer of abstraction between the OS-level software and physical hardware. As a result, the physical hardware becomes a shared resource pool for the guest virtual machines. In highly available infrastructures, the complexity increases with the addition of live migration and failover technologies. Any weak points will cause an infrastructure-wide strain.
Virtualization administrators have greater responsibility to monitor and manage the virtual environment's resources -- such as CPU, memory and I/O traffic. Failure to continually optimize host resources will lead to virtualization performance degradation.
The tips and articles in this virtualization performance and server resource management guide provide advice on how to maintain a virtual infrastructure's resource pool.
TABLE OF CONTENTS
FAQ: How to select optimal memory, CPU and storage for virtualization
Managing a virtual environment's host resources is a balancing act. In many cases, simply purchasing the top-of-the-line hardware is not the optimal setup. This FAQ runs down the common questions about selecting hardware for virtualization.
Selecting CPU, processors and memory for virtualized environments
Most server processors support virtualization. When purchasing a new server, however, it's important to understand the current CPU, processors and memory technologies. Taking the cheap route may lead to compatibility issues and a the inability to use some features.
Capacity planning for virtualization FAQ
Capacity planning is an age-old practice in corporate data centers. It allows IT departments to forecast server resource loads and equipment purchases. In a virtual infrastructure, capacity planning is especially important, because virtualization can waste resources and create bottlenecks if not managed properly.
Optimizing server resource use with systems management tools
Virtualization professionals should always look to fine-tune their virtual infrastructures. Despite virtualization's obvious benefits, if it's implemented incorrectly, there can also be serious negative consequences. Systems management tools can help your organization meet specific virtualization performance benchmarks.
In virtual infrastructure, CPUs must process workloads from the guest machines that reside within the server. If the CPU gets overloaded with requests, virtualization performance will suffer. It's important to understand how CPUs interact with the hypervisor, as well as how to properly monitor its usage.
Memory and CPU allocation in Xen environments: Optimizing performance
In Xen-based environments, dynamically adjusting memory and CPU allocations is crucial. Admins can maximize virtualization performance through meticulous CPU and memory management.
Get started with Intel VT
For years, Intel has supported virtualization. Intel Virtualization Technology (Intel VT) solved many hardware-related issues that plagued the early iterations of virtualization software. When purchasing new hardware for virtualization, make sure that the Intel chip supports Intel VT.
How AMD Virtualization (AMD-V) technology works
The Intel-AMD processor war extends to the virtualization world, and there is a lot of conflicting information at which company provides better virtualization performance. AMD Virtualization technology helps AMD-equipped servers perform better in virtual environments.
Processor and memory horsepower are just a part of the virtualization performance equation. Suitable networking and I/O hardware must be installed to handle virtual infrastructure communication. There are several ways to approach I/O traffic management.
Network I/O virtualization and 10 GbE
Virtualization exposes weak points in a data center infrastructure. Bandwidth and I/O deficiencies, for instance, can stall virtualization deployments. One way to solve these problems is with network I/O virtualization and 10 Gigabit Ethernet.
Benefiting from I/O virtualization
I/O virtualization creates a layer of abstraction for network traffic within a physical network infrastructure. In theory, this technology allows for easier traffic management. It also improves flexibility, utilization and provisioning.
Maximizing I/O virtualization
I/O virtualization further simplifies a virtual environment. It reduces the amount of physical equipment that's required, in addition to streamlining the provisioning process. Before you can reap the benefits of I/O virtualization, you must master virtual network interface cards.
Network consolidation and virtualization solve management problems
Network consolidation through virtualization can improve efficiency and reduce cabling. But I/O virtualization creates security concerns. This technology transmits every traffic protocol -- from the most sensitive to the most insignificant -- through the same pipes. As a result, your information can be vulnerable to network security breaches.
Admins can use specific tools and commands to help with Microsoft Hyper-V server resource management. The following tips and articles directly pertain to Hyper-V management and performance.
Optimizing Hyper-V performance: Monitoring tips
Provisioning new VMs in Hyper-V is a breeze, but this simplicity can lead to bad habits. It's important to stay vigilant and optimize your Hyper-V performance, especially when it comes to CPU, memory, networking and storage configurations.
Sizing the right hardware for high-availability Hyper-V clusters
The mantra "bigger is better" doesn't always apply to server purchases. Advanced Hyper-V features, such as high availability and failover clusters, can actually reduce your server resource requirements. The key is to size the right hardware for your Hyper-V clusters and guard against some inherent problems.
Hyper-V and PerfMon: Even more useful counters
The PerfMon tool aggregates important Hyper-V statistics that provide an accurate assessment of your virtual environment's health. Use the PerfMon counters to troubleshoot problem areas and keep your environment running efficiently.
This was first published in September 2010