VLAN Basics
Inter−Switch Link (ISL) protocol was designed to allow VLAN traffic to flow from one Cisco device to
another. The protocol adds a header that uniquely identifies the source and destinations of the data as well as
the VLAN the data is a member of. If data from one VLAN needs to be forwarded to another VLAN, it
requires some type of Layer 3 routing.
Layer 3 routing can be provided by any number of modules known as internal route processors. The internal
route processors available from Cisco for Cisco switches are the Route Switch Feature Card (RSFC), NetFlow
Feature Card (NFFC), Multilayer Switch Feature Card (MSFC), Multilayer Switching Module (MSM), and
Route Switch Module (RSM). Layer 3 routing for VLANs can also be provided by some Cisco routers that
support ISL, such as the Cisco 4000 series and the Cisco 7000 series.
Spanning Tree Protocol (STP), which can be applied to each individual VLAN, keeps the network from
forming bridging loops when a packet can reach a given destination multiple ways. This means you can
provide multiple ways to get data from point A in your network to point B, thereby providing redundancy in
case one link fails. STP blocks the redundant ports so only one path exists for data in the network.
VLANs allow you to use these links to load balance data. By assigning different VLANs to each link, data
from one VLAN can use one link and another VLAN can use the second, redundant link. A VLAN would use
the other link only during a link failure in the network; in this case the VLANs assigned to the lost link would
converge and use the link that was still available.
A Properly Switched Network
Let’s take a look at how a properly switched network should look. This network implements the switches
using a hierarchical model, as shown in Figure 5.1. Notice that you don’t need a high−speed link to every
workstation in order to create an efficient network, even when using high−end applications such as graphical
CAD applications. In the figure, you see 10Mbps links to each workstation, a Fast Ethernet trunk to the switch
containing the servers, and 100Mbps links to each server. This way, the amount of bandwidth entering from
the 10Mbps switch will not overwhelm all the server links, and you create an efficiently switched network
without bottlenecks.
Figure 5.1: An example of a properly switched network.
Note Switching technology complements routing technology, and both have their place in the networks of
today.
Using Layer 2 switches to create individual collision domain segments for each node residing on a switch port
increases the number of nodes that can reside on an Ethernet segment. This increase means that larger
networks can be built, and the number of users and devices will not overload the network with more
broadcasts and packets than each device on the network can handle (and still maintain a consistent level of
90