In the beginning of LANs the Layer 1 media was shared between all devices in said LAN. They all received and transmitted on the same wires, at the same time. This is whats known as half-duplex (only send or received, not both).
In this configuration devices would sometimes send traffic at the same time, which causes interference on Layer 1. The resulting interference can be enough to disrupt the communications.
The shared “area” in which devices are connected to is known as a “collision domain”.
After some time switches were introduced which allowed the collision domain to be limited to just between the device and the switch (unless you are doing unholy things).
This is much in the same way as routers limit the size of the broadcast domain, but for L2 then of course.
VLANs
The introduction of VLANs was done in the 802.1Q standard på IEEE and allowed administrators to segregate L2 traffic even further. This was possible before VLANs, but would in such cases require more hardware.
VLANs add 32 bits to the header of the frame containing the following fields:
- Tag protocol identifier (TPID): Says it’s a dot1q frame.
- Priority code point (PCP): This is used for QoS between switches and will be covered when i get to that part of the study material.
- Drop eligible indicator (DEI): This is a 1 bit field used as a boolean to tell the switch that the packet can be dropped in case of congestion.
- VLAN identifier (VLAN ID): This is a 12 bit field used as the actual unique identifier of the VLAN.
12 bits gives us 4094 VLANs. However, some of these are “special” VLANs.
- VLAN0 is used for 802.1P (again QoS on L2).
- VLAN1 is the default VLAN.
- VLAN1002-1005 are special Cisco VLANs.
Forwarding is done using the MAC-address table, which contains a database used to resolve which port a given host resides on. This data is stored in the Content Addressable Memory (CAM), which is especially optimized storage for the type of search operations that are run against the MAC-address table.
Leave a Reply