Forum Discussion

kbk491's avatar
kbk491
Icon for Altostratus rankAltostratus
Mar 26, 2024

Traffic failover design between 2 data centers

Good morning,

 

We are designing our network so that during a disaster recovery traffic fails over from main data center to the secondary one.

 

We have a a pair of load balancers (f5 big ip ltm i5600) in active standby in the main data center and during a situation when this data center is down we want to failover to the secondary data center.

 

I did some research and it looks like an f5gtm will be needed?

 

If so will it be needed at both sites or 1 site? I’m assuming both sites right?

 

And will it be a single gtm on each site or a pair of gtms on each site?

 

As far as the secondary data center is concerned only 1 ltm should be enough right? Since it’s only for disaster recovery 1 should be enough I assume? Instead of having 2 ltms like the primary data center.


Also is there a template I can refer to for such a design?

 

 Thank you

  • I did some research and it looks like an f5gtm will be needed?

    Right.

    If so will it be needed at both sites or 1 site? I’m assuming both sites right?

    Yes, both, making your DNS redundant.

    And will it be a single gtm on each site or a pair of gtms on each site?

    Depends on your pocket, one on each site will work but with more nodes comes more resilience.

    As far as the secondary data center is concerned only 1 ltm should be enough right?

    Yes, in practice DR sites have less capacities than main sites, though some clients (wealthy?) chose to make F5 in DR sites redundant too.

    Also is there a template I can refer to for such a design?

    F5 provides a DNS reference architecture, but I actually find the KB articles and guides more advanced technically.