Active-Active data center

[nknarang (Naresh N)]

Hi Folks,

I’m looking into various design considerations to build out active-active data center. Some of these are:

1. Low latency load balancing and sync of databases.
2. Geo-enabled load balancing.
3. Replication aware monitoring.
4. Localized E-W traffic using distributed default gateway. - This is first time I heard about this one. Does anyone know how it works and how to implement?

I’m sure there are lot of things to take into account and I’m only getting started. If anyone knows of a design document I’ll appreciate that.

Thanks
Naresh

[lagapidis (Lazarus Agapidis)]

Hello Naresh

The concepts you are listing here are quite broad and include many different technologies and architectures. Such topics are not typically covered within the core content of the fundamental CCNA/CCNP/CCIE curriculum, and are thus not covered in any depth on the site.

However, doing some Internet searching myself, I’ve come up with the following links that may give you a good start on some what you list above:

• No Latency Load Balancer in Data Centers - Cisco Blogs
• Cisco GSS Global Server Load-Balancing Configuration Guide
• Cisco VXLAN EVPN Multifabric Design White Paper
• Cisco’s Overlay Transport Virtualization

Some datacenter-related content can also be found at the following post, that deals with one aspect of distributed datacenters:

VRRP (Virtual Router Redundancy Protocol)

Hello Giovanni The trombone effect is something that we see with first hop redundancy protocols such as HSRP and VRRP when you distribute layer 2 across multiple datacenters in remote locations. As in the example you shared, this effect results in suboptimal switching of traffic. The solution to this is the use of FHRP isolation which is applied in Cisco’s Overlay Transport Virtualisation (OTV) infrastructure. FHRP isolation allows for the containment of the election process within each site…

I hope this has been helpful!

Laz