What Hurricane Michael Taught Me About the Cloud

It wasn’t a big deal, we’ve faced storms before.  We live two hours inland from the Gulf of Mexico in the southwest corner of Georgia. The land between the Gulf and us always beats the storms down to Tropical strength before it gets to us. We know how to ride out a tropical storm.  We were completely unprepared when Michael rolled through our area as a CAT3 hurricane. In its wake we were left without power, internet or cell service and it remained like that for days. The restoration process is still ongoing, but here are some of the things we’ve learned.

  1. Critical services like email and websites belong in the cloud secured in a robust data center. Once cell service is restored, which is the first service to return, you can reconnect with clients and employees via low bandwidth email and web communications.
  2. Line of business applications like accounting software and industry specific back office applications really should still be on premise with a backup in the cloud especially if your business requires local resources to function.  Power is restored quickly to commercial areas of a city. With an on-premise server, once power is restored or a generator is activated you are back in business. Having those services provided offsite is useless if you can’t get to them or if the available bandwidth is too slow to make them functional.  
  3. Be careful about becoming too dependent on the internet. Restoration of internet service is much slower than the restoration of power.  Depending on cell phones data services as a backup doesn’t work if the cell service is disrupted too or if the service is so oversaturated that it can’t handle the demand.  We found cell data service was the first service we lost and even after it was restored it was too slow to useful.  
  4. Disaster recovery plans are essential.  Even though we didn’t plan for how to respond to a Cat 3 hurricane, we were able to adapt our written plan and process to minimize disruptions.  We were able to do a safety check with all of our staff and make sure that they and their families were safe and that they knew what to do about work.  We were able to safely shut down all of our clients critical services before the storm and restore them once power and internet were restored to their locations.  This would have been much more chaotic without our written plans. 

The take away from this real-life test is that cloud only isn’t the answer, hybrid-cloud is.  Put critical low bandwidth services in the cloud. Leave high bandwidth location dependent services on premise and back them up to the cloud.  Having on premise servers allowed us to relocate one client to our office to dispatch their critical outbound technicians after part of their building was destroyed. It also allowed us to get a doctor’s office back to seeing patients at their specialty practice even though their internet connection will be down for weeks.  Hybrid on premise and cloud solutions gave us the flexibility to quickly get our clients back up and running.