Ethernet Switching & MAC Address Tables

Ethernet Switching Overview

• Explains how Ethernet functions within a switched network
• Aims to clarify relationship between Ethernet sublayers and frame fields
• Describes Ethernet MAC addresses
• Explains how a switch constructs its MAC address table and forwards frames
• Explores switch forwarding methods and port settings for Layer 2 switch ports

Switch Fundamentals

• Layer 2 Ethernet switches forward packets using Layer 2 MAC addresses
• They are not aware of the carried data, such as IPv4, ARP, or IPv6 ND packets
• Forwarding decisions rely solely on Layer 2 Ethernet MAC addresses
• Ethernet switches utilize the MAC address table for forwarding decisions
• Legacy Ethernet hubs differ by repeating bits out of all ports except the incoming port
• When a switch starts, its MAC address table is initially empty
• The MAC address table is also known as a content addressable memory (CAM) table

Switch Learning and Forwarding

• Switches examine the source MAC address to learn new information
• The source MAC address and incoming port are examined for each entering frame
• If source MAC address is new, it is added to MAC address table along with the port number
• If the source MAC address exists, the switch updates the refresh timer for that entry
• Most Ethernet switches maintain an entry in the table for 5 minutes by default
• If source MAC address exists on a different port, it is treated as a new entry
• The entry is replaced using the same MAC address but with the current port number
• If the destination MAC address is a unicast address the switch searches for a match in its MAC address table
• If the destination MAC address exists in the table, traffic is forwarded via the specified port.
• If the destination MAC address is not in the table, the frame is forwarded out all ports except the incoming one, known as an unknown unicast
• If the destination MAC address is a broadcast or multicast, the frame floods out all ports except the incoming one

Filtering Frames

• Switches populates MAC addresses table by examining source MAC addresses in the received frames
• If the destination MAC address exists in the table, frame is filtered and forwarded using a single port

MAC Address Tables on Connected Switches (Video)

• Topic covers how switches construct MAC address tables
• Topic covers how switches forward frames using the content of their MAC address tables

Sending the Frame to the Default Gateway (Video)

• Topic covers what happens when a destination MAC address is not listed in the switch's MAC address table
• Topic covers what happens when a source MAC address is not listed in the switch's MAC address table

Lab - View the Switch MAC Address Table

• Part 1: Build and configure the network
• Part 2: Examine the switch MAC address table
• Downloadable lab guide available from cisco LMS or YJ LMS: 7.3.7-lab---view-the-switch-mac-address-table.pdf

Frame Forwarding Methods on Cisco Switches

• Switches use store-and-forward switching or cut-through switching methods
• Store-and-forward switching receives the entire frame and computes the CRC
• Valid CRC results in the switch looking up the destination address to find the proper outgoing interface
• Cut-through switching forwards the frame before it is fully received
• Reading the destination address of the frame is the bare minimum required to forward
• Store-and-forward switching identifies errors before propagating frames, discarding any detected errors.
• Discarding erroneous frames reduces consumed bandwidth by corrupted data.
• Store-and-forward switching enables quality of service (QoS) analysis
• Congregated networks require frame classifcation for traffic prioritization, like VoIP over web browsing

Cut-Through Switching Variants:

• Switch acts as soon as the data is received, even if the transmission isn't complete
• The switch buffers enough to determine which port to forward the data
• No error checking on the frame
• Fast-forward switching offers lowest latency by immediately forwarding packets after reading the destination address
• Packets may be relayed with errors, later discarded by the destination NIC
• Fast-forward switching is a typical cut-through method
• Fragment-free switching is a compromise to reduce latency, storing and checking the first 64 bytes of the frame for errors before forwarding it.
• By checking first 64 bytes, switch ensures a collision has not occurred before forwarding the frame

Memory Buffering on Switches

• Ethernet switches store frames using buffering during forwarding or if the destination port is busy
• Port-based memory involves queues linked to specific incoming and outgoing ports
• Frames transmit to the outgoing port after successful transmission of all ahead frames
• Possibility of single frame delay despite other frames being transmittable to open ports
• Shared memory involves common buffer shared by all switch ports with dynamically allocated buffer memory
• The frames get linked to the destination, enabling reception and transmission in different ports without queuing
• Shared memory buffering results in larger transmittable frames with fewer drops
• Asymmetric switching is important with different data rates, plus dedicated bandwidth to certain ports like server ports.

Duplex and Speed Settings

• Bandwidth ("speed") and duplex settings are crucial for each individual port on a switch
• Duplex and bandwidth settings need to match the switch port and connected devices
• Full-duplex allows both ends to send and receive simultaneously
• Half-duplex allows only one end to send at a time
• Autonegotiation is on most Ethernet switches and NICs
• Autonegotiation allows two devices to negotiate best speed and duplex capabilities
• Gigabit Ethernet ports operate only in full-duplex

Duplex and Speed Settings Issues

• Duplex mismatch is common on 10/100 Mbps Ethernet links
• Mismatch occurs when one port operates at half-duplex and other at full-duplex
• Reason for mismatches is resetting ports
• It is best to configure both Ethernet switch ports as full-duplex

Auto-MDIX

• Connections once depended on crossover or straight-through cables based on interconnecting devices
• Direct connections between a router and a host require cross-over connection
• Most switch devices support automatic medium-dependent interface crossover (auto-MDIX)
• auto-MDIX automatically detects the cable type and configures the interfaces
• Feature enabled by default on switches running Cisco IOS Release 12.2(18)SE or later but can be disabled
• Use the correct cable type. Do not rely on auto-MDIX
• Auto-MDIX can be re-enabled using the mdix auto interface configuration command

Quiz Questions

• Host will discard a frame with a unicast destination MAC address that does not match its own MAC address
• Switch is the network device that makes forwarding decisions based on the destination MAC address contained in the frame

Module Summary

• Layer 2 Ethernet switches make forwarding decisions based on Layer 2 Ethernet MAC addresses
• Switches dynamically build the MAC address table by examining the source MAC address of the frames received on a port
• Switches forward frames by searching for a match between the destination MAC address in the frame and an entry in the MAC address table
• Switches use store-and-forward or cut-through switching as forwarding methods; cut-through variants are fast-forward and fragment-free
• Use port-based memory and shared memory as memory buffering methods
• Two types of duplex settings used for communications on an Ethernet network: full-duplex and half-duplex