For as long as businesses have depended on datacenters to deliver their services, disaster recovery has been a key topic in the IT world. Sometimes it is an afterthought for many small to mid-size organizations, though for larger enterprises having a solid business continuity plan (BCP) is a must, and often mandated by regulatory compliance.
Regardless of company size, running a secondary datacenter does not scale well. This scenario is more work than you might expect. Doubling infrastructure in another physical location is just the tip of the spear. More interesting is that many companies still do it this way. Datacenter operations teams have become skilled in building their own failover plans between multiple physical sites, but that strategy is unnecessarily painful.
Disaster Recovery as a Service
This leads us to Disaster Recovery as a Service (DRaaS): the modern way to run your datacenter from an alternate physical location. In this case, you replicate your servers & data to your own secondary datacenter or to a service provider’s hosting environment. In the event of a disaster, you still need to solve for the failover of network traffic. It is also worth mentioning that replication traffic is going to require a dedicated circuit with enough capacity to handle your peak activity periods (measured in disk data churn).
When it comes to failover testing you should be running drills at least annually. Some organizations may decide (or be required) to run Disaster Recovery (DR) drills more often. It is also critical to spend time discussing various failure modes and agreeing on what the most likely failures would be for your situation. In my experience, the focus has been on hurricanes or severe weather potentially taking a datacenter off the grid for weeks. You will need to run through the likely scenarios and come to agreement on logistics, staffing plans and other considerations for business continuity.
Disaster Recovery Solutions
The top DR solutions today include VMware Site Recovery, Acronis Disaster Recovery Service, Sungard Availability Service, and Azure Site Recovery, among others. Each has been well reviewed and proven by the experts to produce acceptably low Recovery Time Objectives and Recovery Point Objectives. Most organizations prefer solutions that are well supported but also allow for self-service deployment and minimal commitment, and happily Azure Site Recovery fits this description perfectly since it is truly a pay-as-you-go cloud service.
In normal operations, Azure Site Recovery has three main costs:
- replication bandwidth
- storage cost in Azure for your replicated disks
- minimal compute time needed during your regular DR test drills
There is no additional contract, and if you are already signed up for an Azure subscription you can start right away. It is even possible to run a proof-of-concept with all the required resources using free Azure credits included with a Visual Studio subscription. If you remember to shut down/clean-up test Virtual Machines (VMs) after you’re done, it is a very cheap way to take Azure Site Recovery (ASR) for a test drive!
Finally, let us look at how ASR fits into a longer-term cloud journey. Although many organizations have achieved “full cloud” hosting, there are still plenty of hybrid cloud and even “full on-premises” environments. ASR can protect on-premises workloads running on Hyper-V, VMware and even physical servers in your datacenter. VMs running Windows or Linux are supported and Microsoft has done extensive testing for various workloads.
If you are already running workloads on Azure VMs or plan to be, ASR can support replication for these scenarios as well. In fact, the effort required to replicate Azure VMs is even less than for on-premises workloads (no appliance required). One lesser-known fact about ASR is that many organizations have used it to migrate to Azure by simply failing over VMs to an enterprise landing zone ready for production.
Key Considerations for Disaster Recovery
If you are currently reviewing your business continuity plans and are ready for a new disaster recovery solution, ask yourself the following:
- What is the timeframe for datacenter hardware procurement, depreciation and renewal? If you are coming up on a hardware refresh it makes sense to consider the cloud as your secondary site.
- Do you already have an existing footprint and connectivity into Azure? This makes it almost zero friction to get started with ASR.
- How frequently does your company need to test the preparedness of the team and conduct DR drills? The more frequently you are required to do this the more emphasis I’d place on a self-service option such as ASR.
Business Continuity, Resilience and Security
The Cognizant Microsoft Business Group can help build resilience into your applications, infrastructure, processes and teams with platform native automation tooling, systematic monitoring and event correlation. Get in touch today to prepare your business for the unexpected — with focus, simplicity and scale.
About the Author
Saul Dolgin is an Azure Infrastructure Consultant at Cognizant Microsoft Business Group. His passion for all things cloud and desire to never stop learning is evident in the multiple certifications he’s earned. Outside of work, Saul enjoys spending time with his family — his wife, their two sons and a yorkie — snorkeling, biking and doing anything involving the beach.