Clearly virtual environments have unique challenges and requirements that are not present in traditional physical environments, and trying to manage a virtual environment like a physical one creates problems.
When a server becomes a virtual machine (VM), it shouldn’t be managed like a physical server since the guest operating system on a physical server has direct and exclusive access to the physical server hardware. In virtualisation, a VM has virtual hardware that is emulated by a hypervisor that controls access to the underlying physical hardware.
Virtualisation is a layer inserted in between the hardware and the guest operating system. This is all transparent to the guest operating system, which is unaware that it does not have exclusive access to the hardware.
Corresponding with the five pain points, there are five key areas that a specialist virtualisation management solution should cover to ensure the effective management of a virtual environment. Addressing these will eliminate any potential threats that may disrupt or impact the proper operation of VMs. These key areas are:
1. Performance Management – Good performance is critical in a virtual environment where resources are shared by many VMs. When these machines are fighting for limited host resources, there must be enough resources available and their usage has to be prioritised.
A world-class solution can help identify problem areas that can affect performance, and determine which resource areas are experiencing contentions. By using dashboards, automated alerts and reports targeting the key areas that can impact performance, such a solution helps to ensure that a virtual environment performs well and stays healthy.
2. Capacity Planning – Maintaining sufficient capacity and planning for future growth is a non-stop job in a virtual environment. Knowing what resources to add when they are needed ensures there is always enough capacity without wasting money on unnecessary resources. A top management solution will automatically notify the network manager if there is a pending resource shortage.
It can also help with host failover capacity planning to ensure there is always enough spare capacity to handle host failures. Such a solution can also perform advanced what-if analysis for creating models that will impact resource availability. Because all VMs are not created equally, understanding what applications and services they support is critical for setting the right allocation priorities and making smart capacity management decisions.
3. Identify VM Sprawl – Because it’s not always easy to spot VM Sprawl, monitoring VM lifecycles and growth patterns is critical to identifying it. The answer lies in applying the same best practice configuration management policies to the virtual environment that you would in a physical setup. Looking at a picture of an environment from month-to-month instead of day-to-day will help growth, since virtual environments seldom shrink and most grow continually.
The best management solutions help monitor growth trends that show where a virtual environment has been and where it is going. They will enable the network manager to identify un-used and zombie VMs, orphaned files as well as VMs that have more resources than they need. Ultimately, the management solution will help to prevent an environment from being overwhelmed by VMs.
4. Chargeback and Showback – Using this feature in a virtual environment helps to control its growth, while tracking resource consumption. Since host resources all have an associated cost, it can be extremely difficult to determine which VMs are using what resources. Chargeback breaks out individual VM resource usage to show exactly what resources each VM is using.
A dedicated virtualisation management solution can provide a business perspective into a virtual environment, allowing the network manager to understand the business context of resource usage. Such a solution can publish resource usage reports by business consumer to show what is driving IT costs. Additionally, chargeback helps fight VM sprawl and ensure that all resource usage is truly relevant.
5. Storage I/O Bottleneck Identification – A storage I/O bottleneck can be one of the biggest threats to performance in a virtual environment, as well as a challenging issue to identify and resolve. Bottlenecks result from many causes: from simple configuration settings to improper architecture designs. Many result from improperly designed clusters and managed resources.
A good management solution can help identify potential bottlenecks before they occur, giving detailed visibility into a storage subsystem and allowing the network manager to identify bottlenecks and deal with them. Using detailed storage metrics such as latency, throughput and IOPs, will show how storage is performing and exactly what areas need addressing.