IT Notes : Disaster Recovery using DNS Failover

Disaster Recovery Process

• react to a local failure by redirecting all requests to an alternative location
• data replication between primary and alternate locations
• applications able to receive connections at any time in any location

Timeline

• incident identification
• incident diagnosis to determine the cause of slowness or failure. Consider to failover if the diagnosis is undetermined or long outage is expected.

DNS Failover

1. Manual Process by changing site's DNS record
2. After change, web site's DNS details will be change to DR server
3. some time needed for DNS propagation. May speedup by changing the Time to live (TTL) value to 60 Seconds

DNS fallback on restoration

DNS Failover Advantages :

1. Works with any applications - protocol independent
2. Easy to implement

DNS Failover Limitation

1. DNS caching in client application (browser default 15 or 30 minutes timeout )

Disaster Recovery without ip changes : 

A DNS-based solution that can automatically change IP addresses attached to ‘A’ or ‘AAAA’ records when a monitor detects that the primary site is down, and a network-based solution that simply redirects traffic without having to update DNS, similar to a hardware router or load balancer.

The HTTP Redirect service : users are almost instantaneously redirected to the alternate device(s) without an IP change.

1. use HTTP redirect function 301 and 302
2. incoming client requests are redirected to the selected location