Failover vs. Failback: Two Disaster Recovery Methods

Our online applications exist in two separate clouds with load balancers - if one goes down the load balancers send all traffic to the active cloud. All automatically. So far in 4 years of operation we have never had a complate outage (touch wood - don't like tempting fate). There is little point having a complete secondary system sitting doing nothing until the primary fails - better to load balance across both instances and make use of the extra performance.
 
Back