If you came this far you probably have seen acronyms such as VRRP (Virtual Router Redundancy Protocol) and perhaps HSRP (Hot Standby Router Protocol). They all share the same denominator: first hop redundancy. Speaking plainly, first hop redundancy can be achieved by a series of techniques which might include well-known protocols (focus of this discussion), virtual chassis (such as the one implemented by Juniper), clustered hardware (mostly found on firewalls modules and load balancers), and so on. Gateway Load Balancing Protocol is a proprietary alternative – developed by Cisco – for protocol like VRRP, HSRP, and CARP.
The main advantage GLBP offers is (as its name suggests) the traffic load balancing within a pool of GLBP aware devices, where all routers will actively forward traffic. Both VRRP and HSRP provides you with an active-backup architecture, which depending on the set-up might represent a waste of forwarding capacity, making engineers wonder: “Why must I use only one if I could be using both?“
Load balancing, or more specifically: load sharing, however, can also be accomplished with VRRP or HSRP by simply using multiple instances of themselves. For example: VLANs 10, 20 and 30 might have the highest HSRP priority set to Router-A, and VLANs 40 and 50 to Router-B. This kind of set-up can easily become complex and difficult to maintain, which makes it an unusual practice, not to mention that the amount of traffic being forwarded per second by each of these VLANs is unpredictable, resulting in a highly discrepant load in each of the routers. Continue reading “Networking: Gateway Load Balancing Protocol (GLBP)”