For years, organizations have struggled with maintaining functional disaster recovery capabilities for their most critical business systems. The capital investment for redundant hardware, software and data center infrastructure, along with the personnel time required to ensure that business applications are adequately protected and recoverable, can be monumental. These same concerns and challenges have persisted even though VMware has enabled data centers to dramatically simplify infrastructure management through server virtualization. As a result, DR-as-a-Service offerings are becoming an increasingly popular way to implement data protection and application recovery functionality.
What is DR-as-a Service (DRaaS)?
In fact, according to “Markets and Markets“, the DR-as-a-Service (DRaaS) market is projected to grow from $640 Million (2013) to $5.77 Billion in 2018. This represents a compounded annual growth rate of 55.2%. DRaaS is a way for business application owners to protect critical data assets by leveraging an internal or external service to replicate data offsite. Rather than investing in all the required redundant hardware, software and offsite data center facilities to build out a dedicated solution, the end-user can merely “subscribe” to the DRaaS offering and essentially pay a monthly metered rate for the service. This opens up advanced data protection and DR capabilities to a much wider audience since there is no large upfront capital investment required.
The cloud will be a key enabler for helping businesses deploy DR solutions since cloud provider facilities can be utilized as replication targets for those organizations lacking an offsite data center. When surveying the market landscape for solutions that can enable organizations to implement DRaaS, IT decision makers should consider four key decision making criteria:
- A technology that works seamlessly with VMware infrastructure
- A solution that provides granular, VM level RPO/RTO’s
- A solution that is Storage Agnostic
- A solution that fully automates the orchestration of data protection and application recovery
Many software replication offerings claim to be “VMware ready” but upon closer inspection, some of these technologies merely run as an agent inside each virtual machine (VM) on the host. The challenge with these solutions is that it requires virtual administrators to install and maintain a separate software instance across all the VMs requiring data replication services in their environment. So for example, if the ESXi host is updated to a new software version, the administrator may have to separately update each VM hosting the replication software to a version that is compatible with the new ESXi operating system release.
This may be less of an issue if there are only a handful of ESXi hosts to manage but in mid to large sized data centers, this could mean having to perform manual software updates across dozens or even hundreds of individual VMs. This creates operational complexity, consumes valuable IT personnel cycles and introduces undue risk to the environment.
Consequently, it can be advantageous for infrastructure planners to implement DR software technologies that provide integration at the VMware hypervisor layer. For example, by directly integrating into the vSphere kernel, virtual administrators can designate which individual VMs per host they wish to protect. Moreover, those solutions which “snap-in” to the vCenter management console, make the replication provisioning and management process simpler and more intuitive.
This hypervisor awareness eliminates the need to install a separate agent across each VM. Instead, through a single installation on the ESXi host, administrators can selectively designate which VMs per host require replication protection. Simplifying DR management reduces the likelihood of human errors and can help improve business agility as it will take less time to provision data protection for critical VMs in the production environment.
But DR data protection rarely is a “one-size-fits-all” approach. For example, less critical business applications may only require a 24 hour recovery point objective (RPO) that can be satisfied by a nightly backup process. On the other hand, a mission critical Oracle database system may not be able to tolerate any downtime or any data loss whatsoever. In these instances, businesses need a solution which can support a high availability “stretch” cluster that synchronously mirrors production data to an offsite location. And then, of course, there are applications which need a level of protection that sit somewhere in between highly available stretch clusters and a once a night backup process – such as point-in-time snapshots that are asynchronously replicated.
The point is, businesses need the flexibility to configure highly granular RPO’s and RTO’s across their entire application environment. When data protection options are too monolithic, end-users may opt to go elsewhere to get their specific needs met. In fact, “Shadow IT”, when end-users go to cloud providers or deploy their own in-house silos of data protection, can become a threat to those IT organizations that don’t take more of a service bureau approach to dispensing data protection solutions. Shadow IT can also be a threat to the organization as a whole, since data protection and disaster recovery can no longer be managed and monitored centrally.
Therefore, a second criteria when evaluating technologies which enable DR-as-a-Service capabilities, is to look for solutions which provide a full range of RPO and RTO data protection functionality. This includes synchronous and asynchronous replication that can be configured on a per VM level basis.
In today’s increasingly virtualized environments, perhaps one of the most important attributes when implementing a DR solution is having the flexibility to utilize any storage resource. Many organizations have SAN and NAS platforms from multiple vendors. In addition, to keep costs in check, it is not uncommon for virtualized infrastructure in remote offices to run on direct attached or internal, server-side storage resources. Moreover, following a business acquisition, an IT organization may suddenly “inherit” multiple forms of storage (SAN, NAS, DAS) from various manufacturers. As a result, there is a pronounced need to utilize DR technologies that are both storage platform and vendor agnostic.
By supporting all types of storage, businesses can universally deploy DR-as-a-Service to protect critical applications, regardless of the underlying storage resource. This allows IT organizations to tier data storage and data storage protection to meet the financial and recovery objectives of their end-user constituency. This can help to increase the overall adoption of IT related services – especially in managed service provider (MSP) environments where cost competition is particularly acute.
Orchestrated DR Provisioning and Recovery
Time-to-market has become increasingly critical given the global nature of today’s business marketplace. And server virtualization is a key enabler for allowing organizations to dramatically compress the time it takes to roll-out strategic business systems. But with increased application deployment speed also potentially comes increased business risk.
For example, if a new revenue generating application is instantiated without the corresponding data protection policies put in place, a business could suffer downtime, lost opportunities and damage to their brand. These types of exposure points can be mitigated, however, if the DR service has built-in orchestration features which automate the data protection provisioning process.
For instance, if the service automatically detects newly created VMs (or VMDKS) and presents these to virtual and storage administrators, proactive data protection measures can be taken to ensure that the applications residing on these VMs can be assigned the appropriate data protection policies. If the newly created VM was a service that was part of a mission critical SAP system, for example, the administrator could assign that VM to an existing consistency group and the data protection policies (i.e. Synchronous/Asynch replication) defined for that group, would immediately be spawned into the environment.
From a recovery perspective, ideally the solution would also allow recovery operations to take place automatically as soon as an outage was detected. This would remove the need for manual intervention and help ensure that recovery time objectives (RTOs) could be met more reliably.
DRaaS offerings can make it possible for businesses of all sizes to protect their most critical business systems affordably and reliably. Through these solutions, businesses no longer need to make an outsized investment in redundant data center infrastructure or assign a legion of highly skilled IT personnel to implement and manage DR capabilities. The key, however, is to provide an offering which can align with the specific recovery service levels and financial objectives of the end-user’s business application. This requires a technology that can integrate seamlessly with VMware infrastructure, provide granular RPO/RTO’s, work with any type of storage resource and enable DR service orchestration and automation. These capabilities can enable businesses to differentiate their DRaaS offerings and as a result, increase end-user adoption.
Technologies like EMC’s RecoverPoint for Virtual Machines offering, is an example of a proven DR solution which provides all of the above capabilities. IT organizations, whether private data center environments or managed service providers, can leverage this offering to provide data protection services that are uniquely customized to meet their end users business needs.
This Article Sponsored by EMC