Four suggestions to avoid "Virtual Machine stagnation" (1)

Many analysts in the industry believe that virtualization and integration of data center servers can significantly reduce costs. Therefore, no matter what industry or scale of enterprises, at least some servers and applications should be virtualized.

However, when enterprises begin to implement virtualization to save costs, they must realize that the second stage of virtualization may require enterprises to spend on new tools, these expenditures may far exceed the cost saved by virtualization projects in the short term. If you do not plan for the second stage, the Enterprise is likely to stop in the first stage. Analysts believe that once the first phase of the project stops, it means that although the virtualization project saves the enterprise money in hardware, the enterprise actually only obtains a small portion of the benefits from its virtualization.

James Staten, chief analyst at Forrester Research, pointed out that when companies find that the cost of building 10 to 20 virtual servers is only equivalent to the price of one physical server, they begin to compete for the first server virtualization project, it is planned that 25% ~ 35% of Physical services are migrated to virtual servers. However, soon these enterprises will encounter "Virtual Machine stagnation" VM stall)-because of some cost and organizational problems that lead to virtual stagnation, the entire virtualization project has been directly affected.

Staten said: "During the migration from a physical server to a virtual server, many enterprises may have migrated 50% of the servers, but their thinking is still similar to using physical servers, they did not think carefully about many issues. For example, the license cost of A virtualization project, the number of virtual machines on a host, the performance and capacity management problems in the virtualization environment, and the workload required for support, and so on."

Chris Wolf, vice president of research at Gartner, pointed out that if you plan virtualization for every workload of each server, but IT does not fundamentally change the way the IT department plans capacity, as well as the way IT allocates computing resources and the support and maintenance time of IT Department staff, IT department can only be filled with a large number of repetitive processes, and with physical-to-virtual P2V migration) as the process continues to expand, the return on investment will also decline.

"If you repeat the previous management methods on a virtual server, your return on investment will go into a vicious circle of rapid decline," Wolf said ." Staten believes that to maintain the healthy development of virtualization projects, it is necessary to make a dual change in organization and technology, and to maintain coordinated development in two aspects at specific stages of migration. The following suggestions will help you effectively avoid "virtualization stagnation" in the four key phases of the virtualization project ".

1. Do not blindly pursue the number of virtual machines

Compared with other stages of the virtualization project, the first stage is an ecstatic stage-visible and effective cost savings.

However, Gary Chen, a research analyst at IDC, believes that this is likely to generate unrealistic expectations for future IT and business department managers.

Chen explained that after the first attempt to virtualize "delicious", many IT departments will estimate future success with the same cost indicator, this means that the IT department only considers the density of virtual machines on physical hosts, but does not focus on future investment in management tools and training for IT managers, to optimize the virtual resource allocation in a new way.

"The ratio of physical servers to virtual servers and the number of machines integrated in an environment makes no sense, although these indicators are usually the capital that many people boast about. We really need to focus on how to achieve higher efficiency, that is, how to achieve higher persistent and peak usage in the entire computing resource pool ." Staten said.

2. Improve IT Management

Staten said that in the next phase of migration, to achieve true cost advantages, you need to know the business department for which a single virtual machine is running the task and the resources required by the virtual machine.

Wolf believes that IT management and support methods must be changed. For IT managers, if a resource inventory list cannot be created, more detailed information is provided than the number of available physical servers, then they cannot intelligently allocate specific virtual machines or workloads to the entire available server group. Enterprises with remote virtualization infrastructure cannot allocate resources to their data center branches.

"When trying to automate the supply of virtual machines and regulate the use, movement, and life of resources, you may find that new management tools are needed ." Staten said.

To avoid "Virtual Machine stagnation", you must change the resource management and allocation methods, and assign the responsibility for managing virtual machines to the system administrator. The administrator can use the virtual machine's business department, or specific applications or other factors to determine how to allocate, rather than distribution based on the server deployment location.

Human resources cannot be efficiently allocated, leading to repetitive work, extra work, and blank responsibilities. All of this will lead to the absence of specific owners of virtual machines, this forms a huge waste of resources.

Staten said: "If an enterprise does not implement lifecycle management, or does not implement any automation for system management and support processes, the process may become a phenomenon of human movement and slow down the progress." The most basic requirement is that you must understand the specific tasks of these virtual machines, or whether they have been assigned specific tasks.

1E software company is an asset management vendor and regularly sponsors research on resource use efficiency. According to the company's CEO Sumir Karayi, many data center administrators reported that about 15% of the servers they maintain did not execute any useful tasks, in other words, these servers are idle.

Karayi said: "The IT department usually only focuses on the normal running time. Their job is to keep the machine running normally, but no one cares about what is running, so they only view the resource usage, rather than the non-work load. If a server is being backed up, patched, and restarted after the patch is installed, the server seems to be working all the time. On the virtual server, these problems are more likely to be ignored, because even if the virtual server does not do any practical work, it just keeps running, it seems that there is not much actual cost for outsiders."