How Virtualization is Taking Costs Out of Business Continuity By George Pradel published: Tuesday, July 17 2007
Traditional backup and disaster recovery solutions are a lot like insurance – users pay real costs for something they hope they never need. The more data, users and applications that organizations chose to protect, the higher “premiums” they’ve traditionally paid in terms of longer backup times, more backup hardware and associated maintenance and support costs. However, virtualization is changing this relationship between cost and coverage.
Virtual backup and business continuity solutions enable organizations to improve on the 1-to-1 ratio between primary and backup systems used for traditional disaster recovery and business continuity systems. Because virtual machines can support more users with fewer resources, they provide a practical way for organizations to extend backup protection to more systems – without requiring a corresponding increase in equipment or support. For IT organizations charged to “do more with less,” virtualization is an excellent way to enhance backup and recovery operations.
VMware set the stage for virtualization to emerge as a reliable business continuity strategy with its inclusion of new recovery and high-availability tools in VMware Infrastructure 3. Many organizations are now rethinking their business continuity plans to see how virtualization can be used to reduce hardware and support requirements, make backup programs more inclusive, and cut costs. Virtualization is also beginning to have a strong impact on small and mid-sized businesses (SMBs), who might not otherwise have the budget to implement wide-scale backup and recovery nor the resources to support it.
There are already several very different options available for virtualizing backup and recovery operations. These include new image-based applications, plus adaptations of traditional LUN and agent-based approaches. It is important to distinguish the specific advantages and limitations of each approach from the general benefits of using virtualization for backup, recovery and business continuity. The following sections will help make these emerging issues clear.
Bringing Convenience to Continuity
Traditional business continuity strategies involve the development of redundant parallel systems intended to ensure capacity should the original system fail. Redundant systems are not intended to be highly utilized, so they add acquisition and support costs that are disproportionate to their utilization. A major advantage of virtualization is in improving utilization of IT assets and to slow infrastructure growth.
Virtualization makes it practical to expand backup and business continuity service to more users because it doesn’t requiring the doubling of infrastructure that is typically used in traditional, physical environments. Now that virtualization is established as a viable option, organizations can rethink their backup strategies to actually consider under-utilized servers that are currently in the environment as resources that can be utilized as hosts for restoration of failed servers. After all, if virtual servers are good enough to support users with data and applications in the first place, why wouldn’t they be good enough to restore them after an interruption?
This aspect approach is doable because applications to be replicated to dissimilar hardware, which frees system administrators from having to ensure primary and backup equipment, maintain the same configurations and licensing. As long as the hardware can support the ESX Server 3, it can virtual machines. This functionality is an important time saver for day-to-day system maintenance and also helps quickly restore data or application availability to users after an interruption. Virtualization also simplifies recoveries, backups and routine management because they can be executed by department-level IT staff without involving the SAN team.
Virtualization can work side-by-side with existing systems, so it can be implemented to extend backup and recovery to operations not currently covered without having to disable or replace the legacy solution. Organizations can take advantage of this approach to evaluate virtualization for backup and recovery without fully committing or switching over.
Traditional LUN- and agent-based replication solutions are available for virtual environments. The approach of these LUN- and agent-based replication solutions is to repeat the model they have created for the physical world and carry it over to the virtual world. While this approach is necessary for a selected group of highly critical virtual machines, it may not be the best or most appropriate solution for ALL virtual machines in the data center. New approaches that utilize an image snapshot of virtual machines offer significant advantages in providing a reliable, yet extremely cost effective backup and recovery solution. Image-based backup can require IT administrators to master Linux scripting -- unfamiliar territory for many of them. However there are new tools emerging that do not require scripting and are actually easy to use by any level administrator.
The following sections outline virtualization backup and recovery approaches, including the requirements, drawbacks and benefits.
LUN-based Replication
This widely used replication strategy from the physical world is being adopted for virtual environments. Like their physical counterparts, virtual LUN replication systems take a snapshot of the complete volume and transfer the snapshot to a remote site. Among the advantages -- if the decision has been made to adopt LUN replication -- is that it will be completely integrated into the environment, and is familiar ground for administrators. Another advantage of LUN-based replication solutions is their ability to process large files, which is why they are popular with data-intensive, high-transaction businesses such as financial institutions and e-commerce sites.
Infrastructure and support requirements are the primary drawbacks to LUN-based virtual replication. Organizations must maintain redundant hardware and ensure the primary and backup machines maintain the same configurations. In addition, most LUN replication modes are proprietary bundles of complementary software that must be purchased, licensed and maintained. Administrators, therefore, must be SAN specialists, and require significant training in the solution. Because LUN systems are resource-intensive, they are often used selectively to protect the most critical information and applications, leaving other information and users with inadequate protection. SAN admins must also ensure adequate bandwidth is available to transfer large backup files.
Agents
Agent-based technology is also resource intensive, primarily because of the time required to install and maintain agents – every virtual machine and server requires an agent. Agent-based technology does provide the benefit of faster execution than LUN systems, because there is usually only a small gap between when data is written to a server and when it is replicated. Agent-based technology is suitable for transactional virtual machines and also provides the benefit of familiarity, because many physical backup and recovery solutions use agent technology.
Because most agent technologies were built for the physical world, agents are not optimized for virtual environments and can cause DLL incompatibilities on virtual machines. Virtual machines need to be rebooted after agents are installed, which reduces availability and increases support time. Scalability is also concern, because organizations may be reluctant to leverage and expand their virtual environments due to the time and expense required to add agents to new virtual machines.
Image-level Replication
Image-level replication software is developed specifically for and takes advantage of virtualization in that entire virtual machines can be encapsulated and treated as files, which reduces resource requirements and simplifies management. Image-level replication requires no agents and is executed from a central server, which can be physical or virtual. For data backup, image-level replication can review the server’s block table to see what has changed since the last backup, and only replicate what has changed. The results are faster execution of both backups and recoveries, and smaller, more network-friendly transfers.
Image-level replication does not require the host and destination server to be the same, so users can take full advantage of virtualization to store multiple file types and applications on a single virtual host. This enables server consolidation and simplifies support, which is typically handled without engaging SAN specialists.
Image-based replication is an emerging approach and users have a limited choice of solution options, especially for Novell and Linux environments. The technology has not been optimized for applications that feature high transaction volumes. Also, virtual machine files are extremely large and require significant bandwidth to transfer them to remote sites, so tools that leverage compression help to mitigate this aspect. Image-based replication is not necessarily the optimum choice for file-level restoration. Though Image replication usually costs less to implement and operate than agent or LUN technology.
Summary
Most organizations are already hard pressed to provide all the backup and fast recovery times that business users want, and are reluctant to increase capacity (even if they can win the budget to do so). This forces businesses to prioritize what information they will back up, putting other data and business users at risk. IT administrators are faced with managing turf wars for available capacity, or managing more backup and recovery infrastructure – or both.
In physical approaches, redundancy and additional capacity are built into systems to back up more data and improve business continuity, but the result is simply more underutilized hardware that is sitting there "just in case" primary systems go down. The approaches used in the physical world, LUN replication and agent-based replication, offer significant advantages but are not optimized for virtual environments. Newer products, specifically created for virtual environments, leverage virtualization and offer a cost-effective yet reliable method for backing up and recovering virtual machines, even in the event of disaster or major failure.
Organizations traditionally virtualize to get more utilization from their existing IT hardware. Using virtualization for backup and business continuity puts a twist on this approach: it helps organizations improve utilization and reduce their IT infrastructure, but also enables them to back up more information and support more users with existing resources. In fact, many SMBs and other organizations who would never consider virtualization for server consolidation can use virtualization to tap underutilized servers for backup and recovery, without having to invest in more hardware.
Virtualization provides a way to reduce redundancy without reducing reliability. It is a practical way to ensure data, applications and IT resources will be available when needed, without having to pay the premiums of excessive hardware and support.
George Pradel is senior systems engineer at Vizioncore. Founded in 2002, Vizioncore Inc. is the leading provider of innovative dynamic backup, performance monitoring, migration and disaster recovery software applications that enhance VMware Infrastructure. Vizioncore's products include esxRanger Professional, esxReplicator, esxMigrator, and esxCharter. For more information visit www.vizioncore.com or contact Pradel at
This e-mail address is being protected from spam bots, you need JavaScript enabled to view it
|
