Disaster Recovery and High Availability

May 18, 2020
Charles Kaplan

In these uncertain times, you can be certain that everyone is relying on technology more than ever before. Technology is how we can continue to do business and how we can connect with others while Social Distancing. The world has slowed down, but without technology, it would stop. Today, no one can afford a technology outage. 

Disaster Recovery (DR) is the ability to recover after a technology failure. Disaster Recovery is like insurance that you use only when the unanticipated occurs in order to get you back on your feet. But there is a deductible, a heavy deductible that has to be paid. That deductible is the loss of business during the time it takes to recover and the possible loss of data that hadn’t yet been backed up when the disaster occurred. DR hurts, but it gets you back on your feet.

High Availability (HA) is the ability to keep an application running through both planned and unplanned technology outages. Planned outages occur daily. Your application needs to be down to end-users when you run end-of-day processing or when you run your backups or for planned maintenance and upgrades. Unplanned outages rarely occur, but they are due to technology failures, infrastructure failures or natural disasters. With HA, your application is always available to end-users.

The difference between DR and HA is that with HA there is a spare server with current data that is ready to take over when the production server fails. The downtime is the time to switch over to the spare server and is typically 30 minutes or less. On the other hand, with DR, another server has to be provisioned and restored in order to get the application running again. DR is typically measured in days. HA is measured in minutes while DR is measured in days.

I had two customers who recently experienced server outages. One customer had an HA solution and the other customer was in the midst of implementing an HA solution but was not there yet. Both customers were on currently supported hardware and software platforms and were in good shape to be recovered. In both cases, it took the manufacturer’s service technicians two days to repair the server.

The first customer was back up and running their 150 retail outlets and warehouses in 30 minutes on their backup server. The other customer had to wait for the server to be repaired, the system to be restored and tested and was back up in 4 days. Thankfully, two of the days were over a weekend, when the business is closed, and it was a lighter than normal holiday week. HA = 30-minute outage, DR = 4-day outage.

It sounds expensive to have a spare server sitting around waiting for a disaster to occur, but the loss of business from a disaster can be far greater. The spare server can be in the same room as the production server or in another one of your locations or even in a managed data center. And you don’t even have to make an investment in a spare server, you can subscribe to a HA cloud server. And the spare server is available for other uses like application development and testing, reporting and data warehouse.

Some of us are busier than ever right now which is all the more reason to have an HA solution. You can outsource the entire HA project and get it up and running with minimal involvement of your staff. Others find themselves with time on their hands making this the perfect time to invest in HA to be prepared for when things get back to normal.