Evaluate server virtualization

In this article we will give you an overview of evaluating server virtualization and the type of data you need to choose to perform such an assessment.

Here is a series of three sections that cover the server virtualization assessment using the Microsoft Assessment and Planning Toolkit 3.1 (MAP 3.1). Part 1 of this series will introduce you to an overview of server virtualization evaluation and the type of data you need to choose to perform such an assessment. The information presented in this article has acknowledged that you have a server deployed in a Windows environment.

What is server virtualization assessment?

To design and plan an effective virtualization architecture, you need to take the time to perform an assessment of server virtualization. This is a very important step to ensure that your server virtualization project has a smaller, internal scope or contains the entire business. A server virtualization assessment involves problems such as capturing the current configuration, performance, and environment data for the infrastructure you are interested in virtualizing, accessing data for. With a set of requirements and restrictions for virtual infrastructure, creating key reports to support you build a detailed transition plan for the existing infrastructure. virtual infrastructure. The final conclusion about evaluating server virtualization issues, you need to achieve the following issues:

When you complete the evaluation of server virtualization, you will put yourself in a position to be able to define the impact that your virtualization project will have on reducing the number of physical servers. deployed in their environment. These are the information you can use to evaluate the project's return on investment (ROI) and also what you can use to justify your virtualization project.

What data do you need to choose?

There are three main data sets that you need to support to assess the server virtualization problem:

In addition, you need to collect workgroup, legacy domain, Active Directory forest, domain, site and IP subnet for configuring physical location for your environment.

Inventory of hardware and software

Specifying which existing workload is appropriate for virtualization depends on selecting specific hardware and software information for each server within your project. The physical hardware information you need to collect includes the Basic Input / Output System (BIOS) parameters, the processor type, the number of processor cores, physical, numeric and model memory configurations. network interface card, information about storage devices, USB devices, devices with serial and parallel ports, as well as any other hardware components depending on the current workload. Information about the software you need to select includes the operating system, installed applications as well as software updates, hotfixes, and service packs installed on the physical server. In addition, you need to capture a list of services running on each server as well as the configuration information associated with each running service.

Performance metric

When you have compiled hardware and software statistics for physical servers within the project, you need to include performance metrics for each server. In particular, you need to capture processor, memory, network, and disk performance parameters. Performance metrics must be selected over a period of time long enough for you to capture the peaks and the lowest levels associated with the performance of common applications. The general recommendation is to capture data for a minimum of one month to ensure that the capture period contains high-load events. To reduce the impact of capturing performance data on the server, you need to make sure to set the time to practice more than once every 5 minutes. Please note that the performance parameters mentioned above are Windows operating system counters.

The main processor performance metric that you need to aggregate is the percentage of processor time and the percentage of interrupt for each processor. The percentage of processor time reflects the processing power level required by the workload on each server over time. Percentage of interrupt time reflects the amount of time spent processing the device and peripheral related interrupts. This information will be used to point out virtual processors for each virtual machine and workload combination to optimize processor usage on the virtual host.

The memory performance metric you need to capture is the available memory byte parameters and memory pages per second. The available byte parameters of the memory represent the amount of physical memory calculated in bytes available to serve as an assignment for a processing action; In other words, it is the amount of free physical memory. Pages remembering per second reflect the speed at which these pages are read from or written to disk to resolve hardware page errors. This information will be used to distinguish memory for each virtual machine and workload combination to optimize memory on the virtual host.

The network performance metric you need to collect is the total number of bytes per second for each physical network adapter. This information will affect the design of virtual networks and virtual machine connections to optimize the network load across network adapters.

The storage performance metrics that you need to collect are the real and average time values ​​for reading and writing actions to the physical drive in the server. This information will affect the design of storage systems to ensure that there is sufficient capacity for virtual machines.

The Tet holiday

The environmental data you collect as part of evaluating server virtualization issues includes information evidenced by demonstrating the benefits of cost reduction from migration to a facility. virtual infrastructure. The main data that you need to collect is essentially about power and cooling issues, the cost of the space allocated to each physical server is the goal of virtualization. However, besides that, you must not forget the source backup devices, account storage problems and peripheral devices related to each server in cost calculation. Once you have defined virtual candidates, the number of virtual hosts and their associated systems need to be added to the infrastructure, you can estimate the cost savings from Reduce servers and related systems.

How to use data like?

In terms of evaluating virtualization, hardware and software statistics, power metrics are evaluated for a limited set of components in a virtual environment to distinguish virtual candidate workload, potential workload combination. When configured in a manner that will optimize the utilization of virtual host resources, the number of virtual hosts needed. From the perspective of cost, service, and management, we recommend defining two or three standard server configurations to allow you to support virtual workloads of all sizes, large and small. medium. Defining the physical server configuration requires some repetitive actions to optimize the number of virtual hosts.

For starting a set of host configurations, you can evaluate your physical server's workload data with limited performance and hardware issues of the virtual hardware of the Windows Server 2008 Hyper virtual machine. -V as well as the virtual host's physical hardware configuration. Hardware and performance limitations are based on basic memory, disk space, processor, and network components. Some other limitations that may exist here are parallel or serial hardware devices, etc.

During the hardware assessment, you need to evaluate each target physical server for virtual machine hardware limitations. If a physical server targets these limitations, you need to exclude the relevant workload from the virtual candidate area and continue to assess against the next target physical server. If the target physical server meets all restrictions, you can evaluate the relevant workload and its performance metric for predefined performance constraints.

The implementation of server virtualization evaluation by manual method will be very difficult and is not possible to perform with a large number of physical workloads. However, in this regard, the MAP 3.1 toolkit can help you automatically collect performance, software, and hardware data for your existing environment, as well as access to data for virtual restrictions. predefined, create reports of virtualized candidates for workloads.

Conclude

In Part 2 of this series, I will talk about some of the Microsoft Assessment and Planning 3.1 toolkit features that can assist you in performing server virtualization evaluations.