Hypervisor server diversification is becoming a key initiative for many businesses. In an effort to lower licensing costs by introducing competition for their hypervisor footprint, many organizations may be unwittingly taking on the potential burden of added infrastructure management, higher costs and increased backup complexity of their virtualized environment. On the other hand, if data centers had the ability to transparently migrate data from one hypervisor to another, many of these issues would be eliminated. As a result, Any-to-Any Hypervisor recovery is becoming a critical requirement for data centers looking to drive down costs and increase flexibility.
Virtual Silos
Currently, since there is limited interoperability between the differing hypervisor platforms, organizations may have to maintain separate silos of infrastructure to manage virtual machine (VM) workloads. Key capabilities like migrating VM workloads, via VMotion type technologies, across different hypervisor vendors are often limited to like for like systems. Recovery efforts may also be hindered as backed up VM images will need to be restored on to a system compatible with their hypervisor configuration “fingerprint”.
In short, the value of infrastructure virtualization is diminished when data centers need to silo out server infrastructure resources according to the hypervisor in use. A VMware hypervisor, for example, may be highly resource constrained while a Hyper-V or KVM hypervisor in the same data center may have excess resource capacity. Since these systems are incompatible, however, there are limited ways to perform VM recoveries or to migrate existing VM workloads across hypervisor platforms. In addition to contributing to higher operating costs, this lack of flexibility can delay recovery efforts and potentially increase turnaround times for deploying new business applications.
Core Provider Functionality
While enterprise data centers, cloud service providers (CSPs) and managed service providers (MSPs) need to creatively find ways to drive down their costs to remain competitive, it is critically important for them to do so in a way that doesn’t compromise their ability to protect and restore critical data or hinder their ability to remain agile.
To maximize the value of all infrastructure assets, data center managers need the flexibility to perform data recoveries on any platform to any platform, regardless of the underlying hypervisor operating system or physical machine configuration. Additionally, since most organizations are still in the throes of migrating from physical to virtual server (P2V) infrastructure, they need solutions which can protect both physical and virtual servers while also facilitating P2V migrations in the data center and/or out to cloud based infrastructure.
The ideal backup architecture would also be capable of supporting virtual to physical (V2P) data movement. If problems are encountered when application workloads are migrated from physical to virtual machines, it is critical to have a fail back mechanism in place which enables administrators to quickly restore machine image(s) back to a physical server so that the issue can be investigated without causing extended downtime to the effected application.
Ideally, backup applications should provide the ability to deploy agent software on to a physical server and/or as a virtual appliance running within a VM. Data centers running workloads on physical or virtual servers could then leverage a common backup platform to conduct “off host” backups of physical servers and the VMs running within hypervisors. Some backup applications may claim to support both physical and virtual machines but sometimes require installing backup agents on to the guest VMs themselves. This adds management complexity as these agents may have to be periodically upgraded and/or patched and it also creates additional contention for system resources allocated to the VM.
As importantly, the software should also integrate with the API sets of hypervisor operating system backup applications like VMware’s vStorage APIs for data protection (VADP) to leverage native features like change block tracking (CBT) and combine it with data deduplication to enhance backup and recovery efficiencies. Lastly, the system should be nimble enough to enable virtual system administrators to manage backup processes directly from the VADP console should they wish to consolidate backup management.
These core capabilities are essential for enabling organizations to transform their legacy data center environments to next generation computing infrastructures.
Agnostic Hypervisor Recovery
Perhaps one of the most important features to implement in a virtualized deployment is the ability to perform an any-to-any hypervisor recovery. As discussed previously, many IT planners are seeking out ways to reduce their total cost of ownership (TCO) in the data center. By adopting heterogenous hypervisor technologies, businesses can improve their negotiating leverage and drive down hypervisor licensing costs.
The challenge with mixed environments, however, is increased complexity and lack of flexibility in utilizing hardware infrastructure resources. For example, with many backup applications, it is not possible to perform a recovery of a VMware virtual machine image on to a Hyper-V configured hypervisor host. Likewise, the reverse is true, Hyper-V VMs cannot always be recovered into VMware hosts. The same holds true for Xen, KVM and Parallels hypervisor offerings. True, heterogenous, any-to-any data hypervisor recovery should provide support for all of these environments. Don’t assume because a vendor has a backup application for these platforms, they can recover data between heterogenous hypervisors. In most cases, they are treated like distinct islands of data.
Companies like Acronis, however, have the ability to capture an image of a VM’s data and perform a rapid recovery to any hypervisor host – regardless if the VM was formerly on a different hypervisor system. This capability provides maximum flexibility for data centers to utilize all available server infrastructure to improve application recovery response times and to meet recovery service level agreements (SLAs).
Furthermore, virtual administrators can leverage this feature to quickly create and start new VMs, by using backed up VM images as templates, and restoring those images into any hypervisor host. Likewise, VM migrations can be performed across heterogenous hypervisor platforms by conducting a standard snapshot backup of a VM from, for example, a VMware hypervisor and restoring that VM image into a Hyper-V virtual machine.
Physical to Virtual Workflow
It is also important for the backup application to be physical and virtual server agnostic. Solutions which focus solely on virtualized infrastructure are not capable of protecting data on physical server assets nor can they aid with promoting P2V migrations. Backup offerings like those from Acronis can run as a virtual appliance within a hypervisor and tie directly in, via API’s, to hypervisor backup frameworks, like VADP, to efficiently backup and restore data.
The software can also be deployed as an agent on to a physical host to provide comprehensive physical and virtual server data protection within the data center and across geographically dispersed private and public cloud infrastructure. Furthermore, P2V migrations can be more readily facilitated by “restoring” backed up physical server images into any hybrid VM configured locally or in the cloud. Likewise, virtual to virtual migrations (V2V) can be performed between local systems or into or between cloud infrastructure assets.
These features also enhance disaster recovery capabilities as data movement and restore operations become in effect, virtualized. To a certain degree, the hardware infrastructure becomes abstracted, allowing DR administrators to perform local or remote, cloud based recoveries on any available server infrastructure platform – physical or virtual, with zero dependencies on the hypervisor configuration.
Conclusion
Large data centers and particularly CSPs/MSPs, need advanced backup and recovery capabilities to increase utilization of their hypervisor infrastructure to drive down costs and improve their ability to quickly respond to business demands. Solutions which help promote P2V migrations will enable data center service providers to more seamlessly migrate customer business application workloads into their facilities, resulting in greater adoption of their service offerings.
Furthermore, by using a framework like Acronis’ to backup and recover data regardless of the underlying physical or virtual server platform, data center infrastructure managers will be able to more quickly transform their computing environments to meet customer demands and attain profitability.
Acronis is a client of Storage Switzerland
As a 22 year IT veteran, Colm has worked in a variety of capacities ranging from technical support of critical OLTP environments to consultative sales and marketing for system integrators and manufacturers. His focus in the enterprise storage, backup and disaster recovery solutions space extends from mainframe and distributed computing environments across a wide range of industries.
StorageSwiss.com - The Home of Storage Switzerland 