Change the DNS record to point to your configuration
Before you can fully activate your AMD property configuration in the production ("live") network, you need to change the existing DNS record for your site or application to be a CNAME record that points to the Edge hostname.
Once you modify the DNS record, it's typically only a matter of hours until your content it is served via the Edge network. The table that follows illustrates some DNS record examples:
|Generic DNS record||
||Where “220.127.116.11” is the IP address of your origin server|
|Non-secure record pointing to Edge network||
|A secure, HTTPS record pointing to Edge network||
In these examples,
www.example.com .edgekey.net are
Edge hostnames for your application.
The next few tables reveal some examples of DNS records required to deliver a web
site or application from the content delivery network. You need to replace
<example> with the actual
hostname(s) of your content to determine the DNS records that should be created.
|Example: Non-secure “www” Site|
|Example: Secure “www” Site|
|Origin-server Hostname||origin-www.<example>.com IN A <IP of origin servers>|
|Production Hostname||www.<example>.com . IN CNAME www.example.com.edgekey.net|
|Example: Non-secure Image Hostname|
How long does the DNS change take?
This depends on the time to live (TTL) set on the existing DNS record for your site. This is commonly set to one day, which means it would take up to 24 hours before your end users are directed to the Edge network.
To shorten the cut-over to the Edge network, you can reduce your DNS TTL in advance of the change, and increase it to the normal TTL after the change.
You can switch your application to the Edge network at any time after completing the activation steps and testing. No additional activation or monitoring is required.
If you notice a problem after switching your content to the Edge network:
- Roll back the DNS change to point your website or application back to your servers.
- Report the problem to Akamai.
- This will allow you and Akamai to identify the problem in a controlled environment without affecting live end users.