As part of our annual disaster recovery protocol, this year we performed a live transition of our entire catalog of public web services - from our primary data center to our alternate data center. This test included www.icann.org, www.iana.org, whois.icann.org, whois.iana.org, and czds.icann.org to name of few of the approximately 85 services involved.
I am pleased to share this year’s test process involved almost no manual intervention steps. This is a marked improvement compared to seven years ago, when many manual intervention steps were necessary. What our test proved is that if our primary data center went dark, nearly all services would recover automatically with almost no downtime.
We attribute this improvement to continued advancements in our service architecture. This has resulted in more resilient coding practices aided by a higher level of release automation, test automation, and configuration management.
By design, new services are now being architected with advanced service-level automation. As a result, while the portfolio continues to increase in size, its robustness also continues to increase.
We hope that we never have to confront a disaster. That said, we have a plan and it’s running smoothly should we need it.
