Load balancing is key to bring better availability to new applications topologies… it could be a nightmare is not doing well… datacenters bring high level of redundancy… however, nobody offers 100% protection… split your traffic between two datacenters and active/active redundancy and you will sleep better…
well done Jesus!
What if we are asked to deliver load balancing in a L3 environment? Do you have any idea of how to achieve it? We for sure are familiar with VRRP or HSRP but… real active-active load balancing?
Well, VARP (Virtual-ARP) is a first-hop redundancy mechanism that allows multiple switches to not only share a virtual IP but a virtual MAC address too!.
In most leaf/spine deployments, redundancy at the spine is required in order to deliver high availability.
Modern, two layered architectures use loop-free topologies, using MLAG-like mechanisms in a L2 network and ECMP (Equal-Cost multipath) in a L3 one.
As a First-hop redundancy mechanism, VRRP and HSRP are commonly used to provide active-standby availability, but network doesn´t performs at its best because one link is always unused.
With VARP, each switch is configured with the same virtual IP in any given vlan. That IP acts (just as with VRRP)…
View original post 203 more words