How the pieces fit
- Efficient Cloud Instances run the application in a primary data centre, for example Equinix DB3 in Dublin. Their disks replicate every five minutes to a second site you choose, for example Templus in Barcelona.
- NorainaDNS serves the application’s domain. Its records are failover records: the primary answer is the address in the first site, with a health check that reaches the application directly; the secondary answer is the address in the second site.
- When the first site fails, the replica is started in the second site, and NorainaDNS answers with the secondary address as soon as the primary health check fails.
- Efficient Cloud Backup keeps immutable copies outside our platform, for the failures replication cannot fix, such as deleted data or ransomware.
Stateless parts of the application, such as web front ends, can run in both sites at once with weighted records, so only the stateful part waits for the failover.
Failing back
Once the replica has taken writes, the original site has to be resynchronised before traffic returns to it. We agree with you how failback is triggered, so the health check alone never sends users back to stale data.
What you get
- Production and replica in any two of our five sites: London, Dublin, Paris (two sites) and Barcelona.
- Recovery point of five minutes and recovery time of fifteen minutes from the decision to fail over.
- DNS that follows the health of each site, managed from the console, the API or Terraform.
- Failover tests on a schedule with a signed report, as evidence for business continuity controls in ISO 27001, ENS, NIS2 and DORA.
Cost
NorainaDNS starts at $24 a month, with $8 a month for each additional health check; see the NorainaDNS pricing. Instances and backup are quoted for your configuration: tell us what you run and we send a quote the same week.