One of the keys to designing a fault tolerant network is to eliminate as many single points of failure as possible. Taking advantage of devices that support high-availability configurations can be a big step in that direction. One way to do this is with HSRP (Hot Standby Routing Protocol). HSRP is commonly deployed at the network edge or at critical points of interconnection to provide N+1 redundancy to critical gateway devices and services.

HostMySite's network border is comprised of Cisco 7200 Series routers running a unique combination of HSRP and IBGP. This architecture delivers an edge routing solution that is both fault-tolerant and scalable.

An example of one of our border clusters might look like this:



Full BGP routes are received from all upstream providers and exchanged between the routers via an IBGP session that runs on a private interconnection. This way, though upstream connections are distributed amongst multiple routers, all outbound paths can be leveraged to allow for optimal forwarding.

More information on HSRP can be found here. Feel free to post any questions or comments you might have!

/~darrell

We actually have more than 4 providers, don't we?

That's correct. The diagram represents an example of what one such cluster would look like. Its not necessarily an exact representation of any specific cluster.