VM Sizing Solutions
Most VMs are over-allocated resources when deployed. Virtual environment optimization is critical to get Hyper-V and vSphere sizing correct and free up CPU, memory, and storage resources. With optimization, higher VM/host densities can be achieved to dramatically improve virtualization ROI.
vOPS Overview Video
Inefficient Resource Use Brings Down Data center ROI
Correctly sizing VM resource allocations is key to maintaining a high ROI for virtualized environments. When VMs are provisioned, they are deployed using an estimated amount of CPU, memory and storage resources. Over time, as the usage pattern for a VM is established or changes, it is typical to discover that many VMs were initially over provisioned.
"Right-sizing" VMs requires an analysis of CPU, memory and storage consumption at both peak and average usage over a historical period of time. Most environments will find significant numbers of VMs that require rightsizing. QA, test and development environments are prime right sizing targets.
However, correct VM sizing is not without significant effort. Without an automated way to reclaim CPU, memory and storage, rightsizing the environment may not occur and the capital savings will not be realized. Automated right-sizing needs to occur in a controlled manner allowing for selective resource reclamation by VM, host or cluster and even by resource type. Some resources require a reboot to reclaim whereas others can be reclaimed on the fly. Finally, whether a VM requires a reboot or not, it is important to have the ability to schedule reclamation during off-hours or on a recurring basis.
Right-sizing implemented correctly drives maximum capital cost reduction second only to recycling wasted resources.
Unused Data Files Clutter Storage and Other Hardware
Managing and deploying VMs generates waste. The waste comes from various sources. As templates and snapshots are created and used, the older versions get out of date and are no longer needed. These objects take up valuable storage space. The ease with which VMs are created also causes VMs to be abandoned with the same ease. These “Zombie VMs” run idle with no load on them, but still consume CPU, memory and storage allocations. Finally, while VMs may be deleted from vCenter or Systems Center, their associated images still need to be wiped from storage. Failure to do this, results in abandoned VM images that consume storage, but are lost since there are no pointers to them from vCenter or Systems Center.
Identifying these wasted resources requires analysis of usage patterns, comparison of file lists and structures, and time-sequencing information. Once identified, the ability to analyze these VMs' resource utilization and examine inventory information is needed to determine their true status. For example, is the Zombie really a Zombie, or just a low usage VM? Once investigated, VMs need to be marked for deletion or excluded from future analysis if their usage is still valid. Also, like with resizing VM resources, if the deletion of this waste cannot be enacted automatically, it is likely that it will not occur at all.
Next to correct VM sizing, recycling wasted resources can drive dramatic capital savings especially with storage resources.
Effective Resource Optimization Requires Analytics
Optimizing VM sizing allocations is challenging and requires constant monitoring and data analysis to identify peak loads and calculate optimal allocations. Likewise, tracking unused virtual objects such as idle VMs and snapshots is tedious and involves extensive usage analysis. Analytics of the data provided by Microsoft Systems Center and vCenter is necessary to accurately determine correct resource allocations for VMs, identify Zombie VMs, and recognize data files that have long been unused.
vOPS Optimizer Enables Efficient Resource Utilization and Maintains VM performance
To effectively cut down on overallocated VMs and unused data file waste, several VM performance metrics must be analyzed to note correct allocation amounts on a peak and average basis, and file usage must be monitored to detect unused data files. VKernel vOPS Optimizer, part of the VKernel vOperations Suite, is designed to control VM sprawl by finding and reclaiming over allocated and wasted resources in virtual environments. Using the patent-pending Capacity Analytics EngineTM, Optimizer tracks average and peak resource utilization for all VMs while also identifying abandoned files. Installed as a virtual appliance, Optimizer provides value within 20 minutes from download by:
- Recommending which resources can be safely reclaimed using peak and average historical trending for Hyper-V and VMware VMs
- Visually displaying all VMs and their resource efficiency health on one screen
- Controlled, automated implementation of resource reclamation through the vCenter API
- Uncovering abandoned VM images, snapshots, templates, powered-off VMs and Zombie VMs
- Automatically powering off Zombie VMs through the vCenter API
- Automatically deleting abandoned VM images and auto-merging unused VM snapshots
- Allowing for optimization data integration into external systems such as orchestration engines and systems management frameworks
Optimizing a Hyper-V or VMware environment increases the data center ROI and can constitute up to 20 - 30% worth of reclamation of existing capacity. vOPS Optimizer is simple to download and provides immediate measurable value.
