Before you plunge headlong into a Cloud DR model, here are some best practices that we recommend.
When it comes to your business applications, you can’t be more careful. Business applications and systems are the lifeblood and data is the oil for your organization, which is why CIOs need to take a closer look at the new and emerging DR practices that can help them recoup their business in the event of a disaster.
Interestingly, a large number of CIOs are waking up to the possibility of using Cloud DR. According to IDC, cloud-based DR products witnessed a steady 20.4% growth in 2016. And rightly so, Cloud DR has emerged as a viable option for enterprises that are reeling under a shortage of IT resources. Obviously, adopting DR in the cloud reduces the TCO phenomenally as it eliminates the need to have additional datacenter real estate to park and idle the DR-related IT infrastructure and resources.
However, the challenge is that there is no single template for successful migration of DR on cloud. When it comes to the adoption of DR in cloud, the proverb ‘different strokes for different folks' holds sway. And there are plenty of reasons why.
Every organization is distinct in the way it runs its applications and systems. There is no one size fits all approach here. The relevance and prioritization of the applications to its business environment also vary across different levels. So understandably, their DR in cloud strategy should also be tailor-made to their specific business requirements and industry regulations.
But, before you plunge headlong into a DR in cloud model, here are some best practices you must observe.
It all starts with getting your priorities straight. The first step for CIOs is to prioritize their asset, systems, and applications. You need to perform a detailed risk analysis and prioritize the criticality of IT assets—databases, systems, and applications—in terms of how important or ancillary are such assets to your business availability, and document each and every process and procedure. While this might sound too simple, the prioritization helps determine snapshots of the configuration, infrastructure, processes and technical interdependencies that would be replicated as DR setup on cloud.
Successful recovery is dependent on the ability of application ecosystem to recover and perform in cases of disaster. Clearly, apt attention needs to be paid to determine whether your applications are architected to failover correctly, leveraging virtualization and cloud functionalities. In case your applications aren't designed to failover from its existing VMs to a new VM that spins up in a case of a failure along with the inherent processes, the applications wouldn't survive the failover despite an elaborate DR setup.
Usually, the prevalent industry advice is to right provision IT resources as a case of either over or under provisioning isn’t considered as productive and efficient. However, while prepping for DR, it may make more practical sense to overprovision virtual machines, just a tad. In an instance of a disaster, the cost of lost business opportunity or reputational loss may mount significantly if you are to await an idle bare metal to come up in a physical environment or a VM to be provisioned and spun up reactively on cloud. Depending on the risk assessment, workload priorities, and Recovery Time and Point Objectives (RTOs/RPOs), an enterprise can determine the extent of VMs that can be running as hot DR spares on cloud. In case of an eventuality where the planned DR VMs failed to spin up, the workloads can be failed-over to the extra backup.
It is of profound importance to regularly check if the processes, procedures, /and failover tactics have been configured correctly to achieve the desired results of ensuring redundancy and seamless failover. However, DR testing is a complex process in itself. And, that deters many organizations to conduct more frequent testing of their DR setup. Several companies relegate their testing to the bare minimum needed – once or twice a year. However, Cloud DR does reduce the inherent complexity of testing the DR setup and allows more frequent test runs. With Cloud DR, organizations can test at regular intervals how the DR setup would behave in several scenarios, such as single system loss, multi-node failure, complete data loss etc., without causing significant disruption of their production environment.
DR strategy has a profound impact on the entire organization. Disasters are inevitable, and if IT organization within an enterprise isn’t geared up to failover seamlessly, the enterprise wouldn’t be able to enjoy business continuity and would suffer losses in terms of reputation and revenue. Hence, CIOs can’t afford to relegate DR planning with an internal team or a service provider, and not be aligned and directly involved with DR approach and strategy.
One of the crucial components of a Cloud DR strategy is bandwidth management, both from usage as well as cost perspectives. Cloud DR can be quite bandwidth intensive, if not planned well. CIOs need to ascertain how the service providers approach bandwidth management, as organizations need adequate bandwidth to allow data transfer either to backup or restore. Certain cloud service providers charge for standard bandwidth consumption as well as excess usage beyond the monthly or annual cap. The other major problem that arises out of an inept bandwidth strategy is unstructured and unscheduled usage of bandwidth across an entire organization while the backup or restoration of data is being carried out. The strategy should definitely provision for Quality of Service (QoS) to right prioritize bandwidth-intensive services and bandwidth scheduling capabilities to manage the peak and off-peak traffic better. Data deduplication will also help manage the bandwidth usage much better.
Organizations, increasingly, are witnessing mobility emerging as the central element in their IT strategies. A number of employees are preferring to use mobile devices to access, process and share enterprise data, irrespective of the verticals their organizations belong to, to be more productive, timely, and effective. CIOs aren't able to block or avoid the advent of mobility within their enterprises and the flow of enterprise data across such devices. Hence, it becomes even more imperative for the DR strategy to align with the enterprise mobility or BYOD strategy. In addition to forming the acceptable use policy and deploying mobile data management platforms to better manage the personal and professional silos on such devices, IT teams need to ensure seamless connect of mobile devices with Cloud DR to regularly back up enterprise data and restore when required. Here, the strategies need to articulate the approach and framework quite clearly to avoid backing up of personal data on such devices to avoid any privacy related concern whatsoever and unnecessary wastage of DR resources. While no one can predict the time and extent of natural disasters, these rough and ready best practices can surely help you plan a fool-proof Cloud DR strategy and keep your business afloat in times of crises.