Switches, in the traditional sense, operate at layer two of the OSI stack. The first multilayer switches were called layer-3 switches because they added the ability to route between VLANs. These days, switches can do just about anything a router can do, including protocol testing, and manipulation all the way up to layer seven. Thus, we now refer to switches that operate above layer two as multilayer switches.
The core benefit of the multilayer switch is the ability to route between VLANs. This is possible through the addition of virtual interfaces within the switch. These virtual interfaces are tied to VLANs, and are called switched virtual interfaces (SVIs).
Figure 16-1 shows an illustration of the principles behind routing within a switch. First, you assign ports to VLANs. Then, you create SVIs, which allow IP addresses to be assigned to the VLANs. The virtual interface becomes a virtual router interface, thus allowing the VLANs to be routed.
Figure 16-1. VLANs routed from within a switch
Most multilayer switches today do not have visible routers. The router is contained within the circuitry of the switch itself, or in the supervisor (i.e., the CPU) of a modular switch. Older switch designs, like the Cisco 4000 chassis switch, had a layer-3 module that was added to make the switch multilayer-capable. Such modules are no longer needed, since layer-3 functionality ...