Forum Discussion

Michael_Ozorows's avatar
Michael_Ozorows
Icon for Nimbostratus rankNimbostratus
Apr 10, 2017

IS GTM the way to go?

Hello,

 

We have a large F5 deployment, and I inherited this project. Working to learn my way around F5.

 

That said, I have a request to create 2 VIP's. Easy. But they want 1 VIP per data center. Again easy. They want this to be internally accessible only (so no external NAT) and want it so that if Data center 1 goes offline, the VIP in Data Center 2 picks up and begins load balancing the traffic.

 

I believe GTM is the way to go, but not really sure how to accomplish this ask. Could anyone please point me in the right direction?

 

Thank you

 

1 Reply

  • Kevin_K_51432's avatar
    Kevin_K_51432
    Historic F5 Account

    Greetings, This sounds very much like something GTM (now DNS) could handle. DNS' specialty is steering traffic by resolving DNS queries using a Wide-IP.

    Wide-IPs are made up of pools that are made up of virtual servers. Wide-IPs also use load balancing algorithms to determine which pool to use.

    So, in the case you describe, it sounds like your VIP failover configuration might resemble:

            Wide-IP (www.example.com)
            |  |
            |  |
            +--+----Pool_1 = 1.1.1.1 (data center 1)
               |
               +------------Pool_2 = 2.2.2.2 (data center 2)
    

    You would select global availability as the wide-ip load balancing algorithm which essentially says; hand out IP addresses (1.1.1.1) from Pool_1, until it becomes unavailable, then use Pool_2 IP address (2.2.2.2).

    DNS also has topology based load balancing, so, hand out IP 1.1.1.1 if a query comes from this 10.10.10.1 and 2.2.2.2 if a query comes from 20.20.20.2.

    Here's a very good introduction on the topic:

    https://support.f5.com/kb/en-us/products/big-ip-dns/manuals/product/bigip-dns-load-balancing-12-1-0/1.html