Customers in Switch data centers are provided a number of different options for Internet connectivity. These range from single drops, to multiple redundant drops. Redundancy to the customer is provided either by BGP (Border Gateway Protocol) or HSRP (Hot Standby Routing Protocol). Below are a few examples.
BGP, Border Gateway Protocol, is currently the most appropriate routing protocol to be placed between different systems. In this example a firm demarcation is created between the systems while avoiding a single point of failure. Only a handful of routes need be exchanged between the systems keeping complexity at a minimum.
HSRP, Hot Standby Routing Protocol, is a mechanism to provide a redundant gateway for connected devices. In this example, the redundant gateway is provided to all servers. This example still contains a single point of failure in the customer switch. Failure of this device will render the HSRP gateway inaccessible.
In this next example, we use a seemingly small – but highly important modification from Example A (above). Here, a router is placed behind the switch. The result is a clear and vital improvement, a demarcation is declared between the two systems; however, failure of either device will render the HSRP gateway inaccessible.
Here, we see a rudimentary connection scenario at work. In this model, the Switch router interface acts as a gateway for the connected servers. However, it should also be noted that as there is only a single connection, this model is clearly subject to both planned and unplanned outages.
This connection is the simplest way to take advantage of the Switch blended IP product. The positioning of a router is the preferred method as it creates a firm demarcation. This demarcation allows for expansion much more easily as well. However, like the A example, it is also the most susceptible to interruption from maintenance or failure.