CompTIA Network+ N10-006 Cert Guide (2015)

Chapter 4. Ethernet Technology

After completion of this chapter, you will be able to answer the following questions:

What are the characteristics of Ethernet networks, in terms of media access, collisions domains, broadcast domains, and distance/speed limitations of various Ethernet standards?

What functions are performed by Ethernet switch features, such as VLANs, trunks, Spanning Tree Protocol, link aggregation, Power over Ethernet, port monitoring, user authentication, and first-hop redundancy?

Odds are, when you are working with local-area networks (LANs), you are working with Ethernet as the Layer 1 technology. Back in the mid 1990s, there was tremendous competition between technologies such as Ethernet, Token Ring, and Fiber Distributed Data Interface (FDDI). Today, however, we can see that Ethernet is the clear winner of those Layer 1 wars.

Of course, over the years, Ethernet has evolved. Several Ethernet standards exist in modern LANs, with a variety of distance and speed limitations. This chapter begins by reviewing the fundamentals of Ethernet networks, including a collection of Ethernet speeds and feeds.

Chapter 3, “Network Components,” introduced you to Ethernet switches. Because these switches are such an integral part of LANs, this chapter delves into many of the features offered by some Ethernet switches.

Foundation Topics

Principles of Ethernet

The genesis of Ethernet was 1972, when this technology was developed by Xerox Corporation. The original intent was to create a technology to allow computers to connect with laser printers. A quick survey of most any corporate network reveals that Ethernet rose well beyond its humble beginnings, with Ethernet being used to interconnect such devices as computers, printers, wireless access points, servers, switches, routers, video-game systems, and more. This section discusses early Ethernet implementations and limitations and references up-to-date Ethernet throughput and distance specifications.

Ethernet Origins

In the network-industry literature, you might come upon the term IEEE 802.3 (where IEEE refers to the Institute of Electrical and Electronics Engineers standards body). In general, you can use the term IEEE 802.3 interchangeably with the term Ethernet. However, be aware that these technologies have some subtle distinctions. For example, an Ethernet frame is a fixed-length frame, whereas an 802.3 frame length can vary.

A popular implementation of Ethernet, in the early days, was called 10BASE5. The 10 in 10BASE5 referred to network throughput, specifically 10 Mbps (that is, 10 million [mega] bits per second). The BASE in 10BASE5 referred to baseband, as opposed to broadband, as discussed in Chapter 2, “The OSI Reference Model.” Finally, the 5 in 10BASE5 indicated the distance limitation of 500 meters. The cable used in 10BASE5 networks, as shown in Figure 4-1, was a larger diameter than most types of media. In fact, this network type became known as thicknet.

Figure 4-1 10BASE5 Cable

Another early Ethernet implementation was 10BASE2. From the previous analysis of 10BASE5, you might conclude that 10BASE2 was a 10-Mbps baseband technology with a distance limitation of 200 meters. That is almost correct. However, 10BASE2’s actual distance limitation was 185 meters. The cabling used in 10BASE2 networks was significantly thinner and therefore less expensive than 10BASE5 cabling. As a result, 10BASE2 cabling, as shown in Figure 4-2, was known as thinnet or cheapernet.

Figure 4-2 Coaxial Cable Used for 10BASE2

10BASE5 and 10BASE2 networks are rarely, if ever, seen today. Other than their 10-Mbps bandwidth limitation, the cabling used by these legacy technologies quickly faded in popularity with the advent of unshielded twisted-pair cabling (UTP), as discussed in Chapter 2. The 10-Mbps version of Ethernet that relied on UTP cabling, an example of which is provided in Figure 4-3, is known as 10BASE-T, where the T in 10BASE-T refers to twisted-pair cabling.

Figure 4-3 UTP Cable Used for 10BASE-T

Carrier Sense Multiple Access Collision Detect

Ethernet was based on the philosophy that all networked devices should be eligible, at any time, to transmit on a network. This school of thought is in direct opposition to technologies such as Token Ring, which boasted a deterministic media access approach. Specifically, Token Ring networks passed a token around a network in a circular fashion, from one networked device to the next. Only when a networked device was in possession of that token was it eligible to transmit on the network.

Recall from Chapter 1, “Computer Network Fundamentals,” the concept of a bus topology. An example of a bus topology is a long cable (such as thicknet or thinnet) running the length of a building, with various networked devices tapping into that cable to gain access to the network.

Consider Figure 4-4, which depicts an Ethernet network using a shared bus topology.

Figure 4-4 Ethernet Network Using a Shared Bus Topology

In this topology, all devices are directly connected to the network and are free to transmit at any time, if they have reason to believe no other transmission currently exists on the wire. Ethernet permits only a single frame to be on a network segment at any one time. So, before a device in this network transmits, it listens to the wire to see if there is currently any traffic being transmitted. If no traffic is detected, the networked device transmits its data. However, what if two devices simultaneously had data to transmit? If they both listen to the wire at the same time, they could simultaneously, and erroneously, conclude that it is safe to send their data. However, when both devices simultaneously send their data, a collision occurs. A collision, as depicted in Figure 4-5, results in data corruption.

Figure 4-5 Collision on an Ethernet Segment

Fortunately, Ethernet was designed with a mechanism to detect collisions and allow the devices whose transmissions collided to retransmit their data at different times. Specifically, after the devices notice that a collision occurred, they independently set a random back off timer. Each device waits for this random amount of time to elapse before again attempting to transmit. Here’s the logic: Because each device almost certainly picked a different amount of time to back off from transmitting, their transmissions should not collide the next time these devices transmit, as illustrated in Figure 4-6.

Figure 4-6 Recovering from a Collision with Random Backoff Timers

The procedure used by Ethernet to determine whether it is safe to transmit, detect collisions, and retransmit if necessary is called carrier sense multiple access collision detect (CSMA/CD).

Let’s break CSMA/CD down into its constituent components:

Carrier sense: A device attached to an Ethernet network can listen to the wire, prior to transmitting, to make sure that a frame is not currently being transmitted on the network segment.

Multiple access: Unlike a deterministic method of network access (for example, the method used by Token Ring), all Ethernet devices simultaneously have access to an Ethernet segment.

Collision detect: If a collision occurs (perhaps because two devices were simultaneously listening to the network and simultaneously concluded that it was safe to transmit), Ethernet devices can detect that collision and set random back off timers. After each device’s random timer expires, each device again attempts to transmit its data.

Even with Ethernet’s CSMA/CD feature, Ethernet segments still suffer from scalability limitations. Specifically, the likelihood of collisions increases as the number of devices on a shared Ethernet segment increases. CSMA/CA refers to using collision avoidance, which is common in wireless networks.

Regarding wired Ethernet, devices on a shared Ethernet segment are said to belong to the same collision domain. One example of a shared Ethernet segment is a 10BASE5 or 10BASE2 network with multiple devices attaching to the same cable. On that cable, only one device can transmit at any one time. Therefore, all devices attached to the thicknet or thinnet cable are in the same collision domain.

Similarly, devices connected to an Ethernet hub are, as shown in Figure 4-7, in the same collision domain. As described in Chapter 3, a hub is considered to be a Layer 1 device and does not make forwarding decisions. Instead, a hub takes bits in on one port and sends them out all the other hub ports except the one on which the bits were received.

Figure 4-7 Shared Ethernet Hub: One Collision Domain

Ethernet switches, an example of which is presented in Figure 4-8, dramatically increase the scalability of Ethernet networks by creating multiple collision domains. In fact, every port on an Ethernet switch is in its own collision domain.

Figure 4-8 Ethernet Switch: One Collision Domain per Port

A less-obvious but powerful benefit also accompanies Ethernet switches. Because a switch port is connecting to a single device, there is no chance of having a collision. With no chance of collision, collision detection is no longer needed. With collision detection disabled, network devices can operate in full-duplex mode rather than half-duplex mode. In full-duplex mode a device can simultaneously send and receive at the same time.

When multiple devices are connected to the same shared Ethernet segment such as a Layer 1 hub, CSMA/CD must be enabled. As a result, the network must operate in half-duplex mode, which means that only a single networked device can transmit or receive at any one time. In half-duplex mode, a networked device cannot simultaneously transmit and receive, which is an inefficient use of a network’s bandwidth.

Distance and Speed Limitations

To understand the bandwidth available on networks, we need to define a few terms. You should already know that a bit refers to one of two possible values. These values are represented using binary math, which uses only the numbers 0 and 1. On a cable such as twisted-pair cable, a bit could be represented by the absence or presence of voltage. Fiber-optic cables, however, might represent a bit with the absence or presence of light.

The bandwidth of a network is measured in terms of how many bits the network can transmit during a 1-second period of time. For example, if a network has the capacity to transmit 10,000,000 (that is, 10 million) bits in a 1-second period of time, the bandwidth capacity is said to be 10 Mbps, where Mbps refers to megabits (that is, millions of bits) per second. Table 4-1 defines common bandwidths supported on various types of Ethernet networks.

Table 4-1 Ethernet Bandwidth Capacities

The type of cabling used in your Ethernet network influences the bandwidth capacity and the distance limitation of your network. For example, fiber-optic cabling often has a higher bandwidth capacity and a longer distance limitation than twisted-pair cabling.

Recall from Chapter 3 the contrast of single-mode fiber (SMF) to multimode fiber (MMF). Because of the issue of multimode delay distortion, SMF usually has a longer distance limitation than MMF.

When you want to uplink one Ethernet switch to another, you might need different connectors (for example, MMF, SMF, or UTP) for different installations. Fortunately, some Ethernet switches have one or more empty slots in which you can insert a gigabit interface converter (GBIC). GBICs are interfaces that have a bandwidth capacity of 1 Gbps and are available with MMF, SMF, or UDP connectors. This allows you to have flexibility in the uplink technology you use in an Ethernet switch.