When AWS or Azure or Any Onpremise server goes down
===================================================
While Designing BCP and DR strategy, Enterprises spend lot of time planning how the Data shall be synchronised and application servers shall come up. While that is the major chunk of the action, But the biggest RTO influencer is not considered at all by the planners and that is the final move to production step which requires DNS mapping to be moved.
DNS mapping normally takes anywhere between 24-48 hrs depending upon the DNS provider and the min any provider gives is a 4-6 hrs window which with Caching at each ISP level becomes another 12-24 hrs of min caching.

What it means
============
It means that although the servers shall come up, the users shall still keep getting downtime and shall not hit production URLs until DNS propagation is completed.So, the application is actually not up across users and worldwide propagation shall still take time.
Normally BCP RTO for a customer facing production application is planned to be as less as possible and most of always on nicely architected application are able to bring this less than 4 Hrs but still it doesnt mean that the application shall be up as resolution shall take surely more time than this.

Solution
========
DNS propagation is the most important part as far as DR or BCP is planned. We offer as low as 1 Sec TTL DNS using Anycast across multiple servers across the globe. And this is implemented withought any downtime to current application. The implementation Cost is less than $5 per year per server or $49 for upto 10 servers.