by Andrew Mallaband August 15th, 2013
One of the top priorities of one of our IaaS provider customers is the ability to right-size their subscribers’ virtual machines as part of the migration process from legacy VMware managed service platforms to its IaaS platform. In the legacy environments, the service provider currently has hundreds of bespoke configurations of virtual machines. In the IaaS environment, there are a total of 13 standard virtual configurations that are offered to subscribers.
As part of their initiative, the service provider wanted to determine how to transform subscribers’ legacy virtual machine configurations to the most appropriate configuration of the 13 standard options on offer for IaaS—based on the actual workload demand. Their subscribers commonly under- or over-provision virtual machine resource capacity because they have limited insight into the true workload demands of newly-deployed virtual machines. Through this capability the service provider believes it can help its subscribers maximize the ROI from the ongoing investment in their IaaS offering, while also reducing its internal costs through standardization of the service it provides to subscribers.
When the service provider evaluated VMTurbo they soon realized that our solution provided a unique capability to simplify the challenge they faced. VMTurbo provides the ability to define “templates” that reflect the standard virtual machine configurations offered as part of an IaaS service.
Having profiled the workload demands of virtual machines, VMTurbo recommended actions to right-size virtual machines with non-standard configurations to the most appropriate standard configuration (as defined by the service provider) to meet the demands of the guest applications. The report below shows an example of a recommendation to increase vCPU resource allocation based on a standard virtual machine configuration template.
To implement virtual machine right-sizing recommendations, VMTurbo provides options to execute prescribed changes either manually or through automation. Maintenance windows can also be scheduled where outstanding right-sizing recommendations are executed automatically, without service disruption.
VMTurbo also provides a safety net that minimizes the risk of performance issues when making these recommendations, as it also considers the available host and storage resources that each virtual machine consumes. In a situation where the underlying resources are already constrained, VMTurbo will first recommend to move virtual machines to less utilized hosts or datastores, creating the headroom for the changes in virtual machine resource allocation. If placement is optimal but the available host and datastore capacity is still constrained, then VMTurbo will recommend provisioning more host or datastore capacity to meet the demand.
Having successfully managed the transition of a number of existing subscribers from its tailor-made managed services to its IaaS platform, the service provider is now looking at how the same capability can be exploited to onboard new subscribers from its in-house managed virtual infrastructure.
IaaS providers that want to help subscribers and prospective subscribers maximize the ROI from an IaaS offering while differentiating its service should evaluate VMTurbo.