Understanding Ethernet and Switch Operations

Ethernet was developed in the 1970s by Digital Equipment Corporation (DEC), Intel, and Xerox. Later, the IEEE defined new standards for Ethernet called Ethernet 802.3. 802.3 is the standard that is in use today.

Ethernet 

Ethernet is one of the most widely used LAN standards. As Figure 3-1 shows, Ethernet operates at Layers 1 and 2 of the OSI model.

Figure 3-1 Physical and Data Link Layers

The physical layer (Layer 1) defines cabling, connection specifications, and topology.

The data link layer (Layer 2) has the following functions:

• Provides physical addressing
• Provides support for connection-oriented and connectionless services
• Provides frame sequencing and flow control

One sublayer performs data-link functions: the MAC sublayer. Figure 3-2 shows the Media Access Control (MAC) sublayer (802.3). The MAC sublayer is responsible for how data is sent over the wire. The MAC address is a 48-bit address expressed as 12 hex digits.

Figure 3-2 MAC Sublayer

The MAC sublayer defines the following:

• Physical addressing
• Network topology
• Line discipline
• Error notification
• Orderly delivery of frames
• Optional flow control

Ethernet LAN Connection Media

The term Ethernet encompasses several LAN implementations. Physical layer implementations vary, and all support various cabling structures. The following four main

categories of Ethernet exist:

• Ethernet (DIX) and IEEE 802.3: Operate at 10 Mbps over coaxial cable, unshielded twisted-pair (UTP) cable, or fiber. The standards are referred to as 10BASE2, 10BASE5, 10BASE-T, and 10BASE-F.
• Fast Ethernet or 100-Mbps Ethernet: Operates over UTP or fiber.
• Gigabit Ethernet: An 802.3 extension that operates over fiber and copper at 1000 Mbps, or 1 gigabit per second (Gbps).
• 10-Gigabit Ethernet: Defined in 802.3ae, runs in full-duplex mode only, over fiber.

Network Media Types

Network media refers to the physical path that signals take across a network. The most common types of media are as follows:

Twisted-pair cable: Used for telephony and most Ethernet networks. Each pair makes up a circuit that can transmit signals. The pairs are twisted to prevent interference (crosstalk). The two categories of twisted-pair cables are unshielded twisted-pair (UTP) and shielded twisted-pair (STP). UTP cable is usually connected to equipment with an RJ-45 connector. UTP (see Figure 3-3) has a small diameter that can be an advantage when space for cabling is at a minimum. It is prone to electrical noise and interference because of the lack of shielding. Examples of categories of UTP cable exist: CAT 1, CAT 2, CAT 3, CAT 4, CAT 5, CAT 5e, CAT 6, CAT 6a, CAT 7, and so on

Figure 3-3 UTP

Fiber-optic cable: Allows the transmission of light signals. This offers better support in bandwidth over other types of cables. The two types of fiber-optic cables are multimode and single-mode, defined as follows:

Multimode: With this type of fiber, several modes (or wavelengths) propagate down the fiber, each taking a slightly different path. Multimode fiber is used primarily in systems with transmission distances less than 2 km.

Single-mode: This type of fiber has only one mode in which light can propagate. Single-mode fiber is typically used for long-distance and high-bandwidth applications.

UTP Implementation

An RJ-45 connector is used with UTP cabling. Figure 3-4 shows an RJ-45 connector and its pin connections, following the T568B standards.

Figure 3-4 RJ-45 Connector

The two types of Ethernet cables are straight-through and crossover. Straightthrough cables are typically used to connect different devices (data terminal equipment [DTE] to data communications equipment [DCE]), such as switch-to-router connections. Figure 3-5 shows the pins for a straight-through cable.

Figure 3-5 Straight-Through Wiring

Crossover Ethernet cables are typically used to connect similar devices (DTE to DTE or DCE to DCE), such as switch-to-switch connections. Exceptions to this rule are switch-to-hub connections or router-to-PC connections, which use a crossover cable. Figure 3-6 shows the pins for a crossover cable.

Figure 3-6 Crossover Wiring

Role of CSMA/CD in Ethernet

All stations on an Ethernet segment are connected to the same media. Therefore, all devices receive all signals. When devices send signals at the same time, a collision occurs. A scheme is needed to detect and compensate for collisions. Ethernet uses a method called carrier sense multiple access collision detect (CSMA/CD) to detect and limit collisions.

In CSMA/CD, many stations can transmit on the Ethernet media, and no station has priority over any other. Before a station transmits, it listens to the network (carrier sense) to make sure that no other station is transmitting. If no other station is transmitting, the station transmits across the media. If a collision occurs, the transmitting stations detect the collision and run a backoff algorithm. The backoff algorithm computes a random time that each station waits before retransmitting.

Ethernet LAN Traffic

Three major types of network traffic exist on a LAN:

• Unicasts: The most common type of LAN traffic. A unicast frame is a frame intended for only one host.
• Broadcasts: Intended for all hosts. Stations view broadcast frames as public service announcements. All stations receive and process broadcast frames.
• Multicasts: Traffic in which one transmitter tries to reach only a subset, or group, of the entire segment.

Ethernet Addresses

The Ethernet address, or MAC address, is the Layer 2 address of the network adapter of the network device. Typically burned into the adapter, the MAC address is usually displayed in a hexadecimal format such as 00-0d-65-ac-50-7f. As shown in Figure 3-7, the MAC address is 48 bits and consists of the following two components:

Organizational Unique Identifier (OUI): 24 bits. This is IEEE assigned and identifies the manufacturer of the card. Vendor-assigned: 24 bits. Uniquely identifies the Ethernet hardware.

Figure 3-7 MAC Addresses

Switching Operation

Ethernet switches perform four major functions when processing packets: learning, forwarding, filtering, and flooding.

Switches perform these functions by the following methods:

• MAC address learning: Switches learn the MAC addresses of all devices on the Layer 2 network. These addresses are stored in a MAC address table.
• Forwarding and filtering: Switches determine which port a frame must be sent out to reach its destination. If the address is known, the frame is sent only on that port, filtering other ports from receiving the frame. If it’s unknown, the frame is flooded to all ports except the one it originated from.
• Flooding: Switches flood all unknown frames, broadcasts, and some multicasts to all ports on the switch except the one it originated from.

A switch uses its MAC address table when forwarding frames to devices. When a switch is first powered on, it has an empty MAC address table. With an empty MAC address table, the switch must learn the MAC addresses of attached devices. This learning process is outlined as follows using Figure 3-8: 

1. Initially, the switch MAC address table is empty

Figure 3-8 Frame Forwarding by a Switch

2. Station A with the MAC address 0260.8c01.1111 sends a frame to station C. When the switch receives this frame, it does the following:

a. Because the MAC table is empty, the switch must flood the frame to all other ports (except E0, the interface the frame was received).

b. The switch notes the source address of the originating device and associates it with port E0 in its MAC address table entry.

3. The switch continues to learn addresses in this manner, continually updating the table. As the MAC table becomes more complete, the switching becomes more efficient, because frames are forwarded to specific ports rather than being flooded out all ports.