I want to create a scenario where the existing DNS (having the domain configured) is going to delegate the subdomain to BIG-IP DNS.
I want to know the traffic flow (considering CNAME, A and NS Records are created on the existing DNS) when the existing DNS reply to LDNS, and the LDNS will send the query to listener IP of BIG-IP DNS. Out of the 2 Active-Active Systems, to which BIG-IP DNS, the LDNS will send the query for subdomain.
Yes, they will be in sync group and are sharing the different metric. But for 2 BIG-IP DNS present in 2 separate Data Center, which may be in 2 separate continents...How could the metric value would be same for different load-balancing algorithm.
I am new to BIG-IP DNS and try to understand the traffic flow.
> Out of the 2 Active-Active Systems, to which BIG-IP DNS, the LDNS will send the query for subdomain.
That depends on the LDNS. You cannot control that.
Your main DNS server replies to a lookup for gtm.yourdomain.com by saying
that the names-servers for gtm.yourdomain.com are ns1.gtm.yourdomain.com and ns2.gtm.yourdomain.com, and by the way, I'll give you the IP addresses for ns1.gtm.yourdomain.com and ns2.gtm.yourdomain.com (glue records) as well.
The LDNS now has two name servers and IP addresses for them - it then selects one (or maybe both) and sends the next step in the recursive lookup to get the authoritative SOA from the child domain name-severs and the authoritative name-servers from SOA record.
But what the LDNS does then is entirely implementation dependent. It may select one nameserver randomly every time. If it caches the result, it may use a round-robin scheme. It may query both and deliver the fastest returning result. It might always use the first one. But however it does it, it is out of your control.
For a good summary of how this works and the possible (and actual) issues, look at