This vendor-written tech primer has been edited by Network World to eliminate product promotion, but readers should note it will likely favor the submitter's approach.
A common misconception in IT is that virtualization and enterprise-class disaster recovery just don't mix. Either you get the flexibility of the virtual data center or the robust capabilities and short recovery time objectives (RTOs) and recovery point objectives (RPOs) that come with great DR, but never both at the same time.
The result is that organizations keep mission-critical applications on physical infrastructure, or move these apps to virtualized environments but fail to protect them effectively. Both approaches lead to a technology gap. The logical solution to close this gap though, is to bring robust, effective DR out of the storage stack and implement it at the hypervisor level.
[ MORE: 7 reasons to move DR to the cloud ]
Hypervisor-based disaster recovery is more than just protecting data -- it's about protecting both data and applications within the virtualized infrastructure by ensuring complete application recovery, protecting the IT infrastructure from logical errors and enabling frequent testing.
Traditional array-based DR can achieve these DR objectives, but because replication takes place outside the virtualized infrastructure, it undermines the inherent benefits of virtualization. From the perspective of IT, existing array-based solutions are tough to adapt to the scalability, mobility and flexibility that enterprise-class applications require when running on a virtual infrastructure. Furthermore, the method of grouping LUNs together used in array-based replication can lead to consistency issues and still ties replication, and ultimately DR, to physical storage volumes.
There's no reason for DR to be left out of the virtualization equation, but you might be asking, "Is replacing my array-based replication worth the time and money?" To answer that, ask yourself these questions about your existing DR solutions:
" Is it "virtual-ready"? With array-based solutions, replicating a complete application running on eight VMs, for example, requires the IT team to first track down and consolidate storage locations and then define replication policies -- a complicated and expensive process.
" Can it support mission-critical application needs? Many replication solutions don't support consistency grouping across hosts and storage, which can affect write order fidelity of the entire application, making recovery a slow and difficult process. Furthermore, RTOs and RPOs can be far longer than required.
" Can it provide mobility and migration? Hypervisor-based replication can move workloads between data centers while still maintaining full transaction integrity.
" Is it application-aware? The essence of the virtualized infrastructure is application awareness. It is the enabler for the high level of IT service that users demand. Traditional DR does not replicate at the application level, and doesn't allow IT to configure the replication policy for an entire application at one time.
" What's the learning curve? The key to successful adoption and implementation of any enterprise tool is overall ease of use. Array-based DR usually requires significant setup time, training and ongoing support in order to use it effectively.
" Is the solution scalable? As the virtual infrastructure grows (as it is wont to do in most enterprise environments) organizations need DR to grow along with it, without the need to purchase, install and configure additional proprietary software. Managing hundreds and thousands of VMs is not trivial, and traditional DR can often increase the challenge of managing them.
Impact on the bottom line
Once these questions are answered, you should look at the overall effect that hypervisor-based replication can have on the business. First, it reduces IT hardware costs. Since the DR happens between hypervisors, storage types and brands no longer matter. Organizations can leverage existing, underutilized storage assets as the replication target, or even implement lower-cost storage at the replication site. Additionally, since the system is protecting VMs and not entire LUNs, overall storage and bandwidth requirements are greatly reduced. LUNs can be utilized for multiple applications and organizations can replicate only what is necessary.
The second big impact on the business is that IT operations become streamlined. With DR at the hypervisor level, there's less day-to-day management, and there's no need for back-and-forth communication between the storage and virtualization teams. The tedious task of consolidating LUNs or creating new ones is eliminated, and the server and storage infrastructure don't require significant changes to implement the technology.
Lastly, and perhaps most importantly, organizations get powerful DR capabilities for their virtualized mission-critical applications.
Enterprises understand that taking full advantage of virtualization leads to more flexibility, better allocation of resources, less management of IT assets and overall cost savings. It simply does not make sense for DR to be left out of the virtualization equation.
Levonai is vice president of marketing and products for Zerto. He spearheads marketing and product management, supporting the corporate vision by leading the go-to-market strategy.
Read more about data center in Network World's Data Center section.
This story, "Understanding Hypervisor-based Disaster Recovery" was originally published by Network World.