DRaaS - Rethinking disaster recovery for the cloud
Given the financial and reputational threats of unplanned data center outages, cloud-based Disaster Recovery (DR) is as an approach than can minimize friction and ensure innovation scalability. It is time for CIOs to reimagine how they deploy IT staff from a reactive model focused on DR and Business Continuity, to one that is preventative, proactive and designed for organizations to be able to navigate through an adverse event. Now is an opportune juncture to implement a strategy of cloud-based DR.
Many Organizations Are Unprepared
Everyone knows that they need a DR plan and the accompanying technologies in place to support it, however more than 75% of small businesses (Nationwide Insurance Survey, 2016) have no documented DR strategy or plan. Making the assumption that your organization is immune to data breaches and/or natural disasters can be an extremely costly mistake that you never recover from.
According to the Federal Emergency Management Agency (FEMA), more than 40% of businesses never reopen after a disaster, and of those that do, only 29 percent were still operating after two years. As for those that lose their information technology for nine days or more after an adverse event, they enter bankruptcy within a year. Given these unsustainable costs, the question is not whether you need a DR plan, but what plan is best for your organization
Even those with a DR plan might have poor insight into whether they can fully recover from an extended outage, because their DR plan is rarely tested or scores low marks in preparedness. When factored against IDC’s recent research, which shows that companies lose an average of $84,000 per hour of downtime, costs can be prohibitive.
According to a Disaster Recovery Preparedness Council Survey:
- 73% of survey participants worldwide scored ratings of either D DR Recovery plan
- 40% admitted that the disaster recovery plan they currently have did not prove useful
In addition, DR is a frequently a requirement of regulatory bodies or even insurance companies. Given how IT has become the lifeblood of modern businesses it only makes sense to ensure that operations can continue as smoothly as possible after some sort of disruption.
For many years, the answer to DR was to send daily backups offsite and have a prearranged site ready to restore those tapes. There are some companies that rose in prominence because they could provide like-for-like hardware to be used in the event of a disaster at their facilities, but at a premium expense. Other companies would buy hardware as an insurance policy that they would keep at a second location for the same purpose. While both of these options made sense at the time, increases in data size and reductions in bandwidth costs popularized the use of replication software to keep a second site closer to the current state of production. This was certainly an improvement, but still required an expensive second site with gear that largely just gathered dust.
Cloud-based DR, a Paradigm Shift
There have been some significant transformations in backup and disaster recovery over the last couple of years. Just like almost every other aspect of IT, the cloud is making its presence felt by driving IT infrastructure and business transformation. Integrating the cloud into a backup and DR strategy involves turning to a flexible and scalable cloud solution to better manage these processes. Cloud providers have a much greater amount of compute and storage capacity, all of which is metered and paid for on a daily or hourly basis. By replicating to a cloud, you no longer have to provision, manage or maintain a second site of your own. You are able to use either the public Internet or some form of dedicated connectivity to a provider who can serve as a target for your replication.
Most of the time, DR processes are just about sending data to be stored in a second location; cloud providers only charge for that storage to reduce ongoing expenses. During DR tests (or an actual DR event), the cloud provider is able to rapidly provision compute resources based on the replicated data. You pay for that compute only during the hours that the test or event is occurring, and when you're done, it disappears back into the service provider's general pool of resources and you are back to only paying for storage. This model allows you to more tightly tie your expenses to actual needs, while also letting someone else worry about keeping the DR infrastructure up to date and current. Cloud-based DR is much better than the old model of dedicated gear and a full-time location.
Designed for disruption
Continuous disruption is a function of today’s IT landscape. Organizations can’t count on static technology investments to continue returning value if they can’t adapt in an era of constant change. Cloud-based DR, just like other technology and systems built today, will inevitably be disrupted tomorrow, so it is important to understand how by migrating to a cloud-based approach, organizations can achieve scale and seamlessly accommodate different application workloads.
At its core, cloud-based DR solutions are about quickly and efficiently moving data to a second site with access to compute resources. When properly architected, the second site can easily be isolated from production environments, meaning that you have access to a second, updated copy of your production environment whenever you want.
So, what else could you do with that?
- The simplest and most straightforward advantage is the ability to execute a DR test whenever you want without impacting production. Instead of limiting yourself to running tests in the middle of the night or weekends, why not do them on a Tuesday afternoon? Why limit yourself to only two tests a year? Why not 4, 6, or even 12 times a year? The more you test, the more confident you are that your DR plan works, and the more proof of success you can show to your boss or an auditor.
- If it's easy to spin up a copy of production, why not use that as the basis of a QA or dev/test environment? Most organizations want to have a proper process for testing software changes, but the complexity and costs of building and maintaining separate environments can be overwhelming. The cloud allows for environments to be spun up and scaled whenever they are needed; the addition of replicated production data makes it even more appealing. You can quickly build a QA environment before a major upgrade, test it a few times, and then delete it when done. No more excuses for "testing in production".
- The cloud provides wonderful economics around dev/test, but many of those same attributes make it attractive for production workloads as well. A frequently cited reason for not moving to the cloud is simply not knowing how to do it in an efficient and effective manner. The same technology that moves workloads to the cloud can be used for migration; a migration is really just a failover that never ends. The non-intrusive testing aspect of cloud DR, coupled with the ability to fail over individual or groups of workloads, means that a migration plan can be created and tested multiple times to ensure that it can be performed in low-risk phases, instead of all in one shot.
- Of course not all clouds are created equally, and some are better suited to particular workloads than others. You want to be able to pick and choose cloud platforms based on specific business needs, but moving workloads between clouds underpinned by different vendors can require navigating a wealth of native tools. If you use a DR tool that supports the major public clouds, as well as VMware and Microsoft hypervisors, you can not only use one cloud for DR and another for other needs (keeping your eggs in separate baskets), but also move workloads between clouds based on evolving business requirements. This ability to keep your workloads mobile allows you to reconfigure and redeploy services with maximum agility.
Move beyond traditional DR
Moving to a cloud-based DR approach is not an all-or-nothing proposition. Few organizations can actually afford to simply walk away from their existing infrastructure investments. In addition, not all systems or platforms are necessarily good candidates to be migrated to the cloud. Cloud services are fundamentally modular and flexible, making implementation readily adaptable to any organizations strategic IT vision. There are some additional steps that can be taken and expanded over time.
- Re-allocate budgets from traditional DR infrastructure to a DR as a Service (DRaaS) solution in the cloud. This can help introduce the financial, security, and operational benefits of the cloud, while freeing up resources to be reinvested in other strategic initiatives.
- Migrate to cloud-based DR for your most critical digital services. Incorporating always-on principles into the requirements, design, implementation, and operational processes for your most critical applications and data can immediately improve cybersecurity risk posture and seed a cloud-based approach for the organization, given that public cloud platforms have shown themselves to be more secure than most on-premises data center environments.
The ability to quickly bring up a copy of a production environment in case of some sort of disaster is a longstanding requirement, but the evolution from offsite tapes to simple replication, to cloud DR has brought capabilities to bear that were unthinkable 20 years ago. The same tools and technologies that have been developed to meet these needs can also be used for a variety of other IT initiatives, particularly around workload mobility and making the most of the cloud marketplace. Don't think of cloud DR as the end of single business need, but rather look beyond to what else it can provide your organization.
As a leading provider of managed cloud services, Navisite has watched the cloud and DR markets evolve over time and have selected tools and technologies to help our customers take advantage. Navisite has helped customers migrate to the cloud since 2011 when we launched our first multi-tenant, VMware-based cloud in the US and UK. Thanks to our close partnership with VMware and the diligence of our outstanding R&D and engineering teams, our hosted clouds continue to grow in both size and functionality to meet the needs of an increasingly diverse cloud marketplace.
In 2016, we launched support for the Microsoft Azure cloud allowing the same high level of managed services to be provided by the Navisite team on clouds both within our data centers and with one of the industry leading public clouds. By supporting separate clouds from separate providers, we are able to provide our customers with solutions tailored to their specific needs instead of forcing them into a pre-defined box.
In support of our cloud DR services, Navisite has partnered with Zerto since 2015. Zerto originally came to market with a uniquely high-performance and enterprise-focused DR solution, and since then has grown to be one of the few solutions that can truly check all of the necessary boxes for cloud DR. Zerto can replicate workloads from VMware to Azure to Hyper-V to AWS all within the same simple software stack and keep sites up to date within seconds of each other. Zerto's power lies in the ability to keep sites up to date within seconds of each other while also allowing for isolated failover to meet all of the use cases described above. Zerto is one of Navisite's key partners in making a multi-cloud strategy work.