A lot is lost when the systems that help us conduct our businesses go down. Time, money, product, reputation (your company’s and your own) are all at stake. What makes this particularly egregious is that we can’t control when “the cloud”, public or private, goes down. What we can control is how it impacts us. The decision you make today about resiliency and recovery will determine your next job. Are you on the market because your peers in the industry are impressed with the job you did or because the CEO just fired you? Having hopefully got your attention, let’s get some terminology straight.
Many in IT today conflate disaster recovery (DR) with high availability (HA). While they’re very similar the difference between them is quite important. A disaster recovery is about massive geographic separation. HA is about surviving (or at least minimizing the impact of single points of failure.)
When it comes to public cloud you have a lot more options available for high availability. For most private data centers capital expenditure, “capex” realities prohibit companies from truly building the level of resiliency these applications deserve. This is the biggest differentiation between the public cloud and the private cloud. Simply put, capex refers to the funds a company invests in acquiring, upgrading, or maintaining physical assets with the aim of generating future benefits. Public cloud companies, like AWS and Azure, spend billions of dollars a year on capex. Most private clouds have much smaller amounts of capex, usually ranging from $1-10 million dollars.
That difference in scale comes into play regarding the number of data centers a company has at its disposal, which in turn contributes to the speed by which its particular backup method – high availability or disaster recovery – can restore during an outage.
In this article, you’ll learn the difference between disaster recovery and high availability.
High availability
Within the modern hyperscale cloud context, high availability refers to the ability to survive the outage of a single data center. That single data center, or a couple of data centers, that are located within the cloud are referred to as availability zones on Amazon or Azure, or availability domains on OCI.
A region generally is an extended metropolitan area which contains multiple data centers. As much isolation is built into their locations as possible so that it is statistically improbable that if data center goes down, another one will go down. What this means for the customer is that while a data center itself might go down, the region can remain up. Subsequently, JD Edwards (or any traditional Enterprise application for that matter) will take that outage, but the customer can architect it in such a way that it comes back up very quickly, with damage and downtime minimized as much as possible.
High availability works by replicating data from one availability zone or domain into the other. The replication acts as a standby server. The database can be from Oracle or SQL – both technologies support this type of replication by and large. The key factor is that the customer doesn’t have to do a database recovery in the event of an outage, because the replicated database already exists in that other availability zone or domain. The backup servers can be pre-staged and ready to go in the event of an outage, or, because these servers are smaller, a restoration process can be done very quickly.
Disaster recovery
A similar concept applies to disaster recovery as with high availability when it comes to utilizing geographical separation for data centers. Clouds that operate with disaster recovery generally delineate their regions based on governmental FEMA regions. Production, or the hosting site, runs on one coast, and the disaster recovery runs on the other coast.
The key difference between HA and DR is that with disaster recovery, the pre-staged server(s) in the backup location are not present. A full restoration needs to be completed in an urgent manner, resulting in aggressive recovery time objectives that can be achieved for customers.
Something to keep in mind with disaster recovery in a private cloud is that there is a 1:1 ratio from a cost standpoint. If you have a production server in your hosting region, you need to have the corresponding capacity for that production server in your DR region.
Leyla Shokoohe
Leyla Shokoohe is an award-winning journalist with over a decade of experience, specializing in workplace and journalistic storytelling and marketing. As content manager at ERP Suites, she writes articles that help customers understand every step of their individual ERP journey.
