Problem Overview

As virtualization takes hold with aproximately 60% of all workloads today virtualized, it has led to “VM sprawl” creating a nearly unmanageable number of virtual servers, still managed as if they were physical PCs. There are over 100 million virtual machines created already with more being created every second.  Each of these VMs has its own copy of the Guest OS, Application, Plug-ins, agents and application specific data creating a huge management burden. With increasing server utilization, and the average density of VMs per physical host hovering around 6 and expected to move up to 10 or more, and the ease at which new VMs can be cloned and provisioned the number of VMs and their management needs will continue to increase exponentially.

Deploying applications into a virtual machine and managing the updates, patches and app versions is tedious because it has to be done for each and every server VM. Costs of management scale with the number of VMs, and with the ever increasing number of server VM’s being managed uniquely this cost will only continue to rise.

CloudVolumes Solution

The efficiency of server virtualization can be radically improved if the VMs were leaner and if there was a way to manage applications with less effort, greater flexibility, and deploy them more rapidly.

CloudVolumes’ unique approach for deploying server applications and workloads into virtual environments optimizes the storage, migration and management of the VMs. This is achieved by separating the VM into individual components – OS, application, and unique server data. Common components can be shared and wasteful duplication avoided. When needed, the components can be “snapped” together to rapidly provision new workloads. Further, since the VMs are leaner and share single copies of applications, the costs of managing and updating the application workloads is significantly lower because updating the single image reflects automatically into all VM’s associated with that workload.

Complex multi-tier virtualized server workloads such as Microsoft SQL, IIS, and SAP can be easily deployed while being agnostic to the type or kind of underlying hypervisor or cloud used. It is also possible to have different versions of the same application available to deploy into the VMs. For example, SQL Server 2008 can be upgraded to SQL Server 2012 instantly, with the app being swapped seamlessly and without requiring a reboot, and just as easily be unassigned in the case of a failed patch or upgrade. This will reduce the risk of performing mass upgrades of an application across a large numbers of server VMs.

When virtual machines and their related workloads need to be migrated from one server to another, it is quick and uses minimal bandwidth because only the workload itself needs to be moved, not the entire virtual machine as with todays approaches such as VMotion. This ability to deploy and migrate quickly can be used to migrate workloads from one hypervisor type to another, or from a private data center to the cloud all without having to convert or migrate the VM – only move what you care about – your apps and data.