For an IP multinetted interface, one of the IP networks on the interface acts as the transit network for the traffic that is routed by this interface. The transit network is the primary subnet for the interface. The remaining multinetted subnets, called the secondary subnets, must be stub networks. This restriction is required because it is not possible to associate the source of the incoming routed traffic to a particular network. IP routing happens between the different subnets of the same VLAN (one arm routing) and also between subnets of different VLANs.
Figure 210 shows a multinetted VLAN named multi. VLAN multi has three IP subnets so three IP addresses have been configured for the VLAN. One of the subnets is the primary subnet and can be connected to any transit network (for example, the Internet). The remaining two subnets are stub networks, and multiple hosts such as management stations (such as user PCs and file servers) can be connected to them. You should not put any additional routing or switching devices in the secondary subnets to avoid routing loops. In Figure 210 the subnets are on separate physical segments, however, multinetting can also support hosts from different IP subnets on the same physical segment.
When multinetting is configured on a VLAN, the switch can be reached using any of the subnet addresses (primary or secondary) assigned to VLAN. This means that you can perform operations like ping, Telnet, Trivial File Transfer Protocol (TFTP), Secure Shell 2 (SSH2), and others to the switch from a host residing in either the primary or the secondary subnet of the VLAN. Other host functions (such as traceroute) are also supported on the secondary interface of a VLAN.