A testament to the ever adapting pioneers that they are, Cisco has developed the first gateways redundancy protocol: Hot Standby Router (HSRP). HSRP allows for default gateways to be failed over to another router, based on a priority that can rise or fall contingent upon interface tracking.
The Internet Engineering Task Force (IETF) created a standard that is almost identical: Virtual Router Redundancy Protocol (VRRP), as identified in RFC 2338. The only real differentiator is the terminology. If you have non-Cisco routers or are pairing between Cisco and another vendor then you are using VRRP.
Here is an example of the old days:[iframe src=”http://www.integrateddatastorage.com/wp-content/uploads/2011/10/Nick-Blog-Pic-12222.jpg” width=”535″ height=”525″]
Next in the long line of gateway redundancy protols came HSRP, which allows for failover of the default gateway. The only way to load balance was by creating two different HSRP groups: multiple HSRP (MSHRP), using different IP addresses for the default gateways. Hence you would have to configure Dynamic Host Configuration Protocol (DHCP) pools that give two separate gateway addresses for the SAME IP range. Sound painful, right?
Let’s look at general HSRP operation. For example: you could have Router 1 and Router 2 running HSRP which would both be tracking their WAN links. Below is normal HSRP operation: the router on the left is actively forwarding traffic as the default gateway, and the one on the right is waiting for it fail or lose its WAN link. Notice that the top router is doing absolutely nothing, aside from looking pretty.[iframe src=”http://www.integrateddatastorage.com/wp-content/uploads/2011/10/Nick-Blog-Pic-2-Rev.jpg” width=”605″ height=”450″]
Now, the WAN link fails and the other router takes over.[iframe src=”http://www.integrateddatastorage.com/wp-content/uploads/2011/10/Nick-Blog-Pic-3-Final.jpg” width=”605″ height=”440″]
When the link goes down the other router takes over forwarding traffic. It is a time tested strategy, but if you have two routers why not utilize both?
Introducing another Cisco first: Global Load Balancing Protocol (GLBP). GLBP introduces two router roles:
- The Active Virtual Gateway (AVG): responsible for giving out the Media Access Control (MAC) address to the other routers as well as responding to clients Address Resolution Protocol (ARP) requests.
- The Active Virtual Forwarded (AVF).
The AVG generally gives out the MAC address in a round robin fashion (though there are other choices). Some clients get MAC for Router 1 and some recieve ONE IP address.[iframe src=”http://www.integrateddatastorage.com/wp-content/uploads/2011/10/Nick-Blog-Fourth.jpg” width=”605″ height=”490″]
Normal Operation:[iframe src=”http://www.integrateddatastorage.com/wp-content/uploads/2011/10/Nick-Blog-Pic-5.jpg” width=”625″ height=”525″]
Now, I’m sure you are wondering what happens on a link failure or router loss.
Since there are only two routers in these scenarios, the AVG would take over for the MAC address, making the failover absolutely seamless. The router on the right would lose it’s link and report that it is no longer able to forward traffic. Ok, it might be a little more complicated than that, but you get the gist.[iframe src=”http://www.integrateddatastorage.com/wp-content/uploads/2011/10/Nick-blog-Pic-6-Again.jpg” width=”625″ height=”525″]
GLBP is a great solution for load balance and it offers your users seamless failover of their default gateway upon the failure of a router.
Perhaps the IETF will make this a standard too!