Ethernet Operation 289 first hears an FLP burst or some other signaling scheme. In the silent listening state, a portable computer is capable of conserving enough battery power to be quite worth- while, although this mode is not supported by the standard. Ethernet Duplex Operation Two duplex modes exist: half and full. For shared media, the half-duplex mode is man- datory. All of the coaxial implementations are inherently half duplex in nature and cannot operate in full duplex. UTP and fiber implementations can be operated in half duplex, but that mode is an administrative imposition. All 10-Gbps implementations are specified for full duplex only. In half duplex, only one station can transmit at a time. For the coaxial implementations, a second station transmitting causes the signals to overlap and collide, becoming cor- rupted. Because UTP and fiber generally transmit on separate pairs, the signals have no opportunity to overlap and become corrupted. Ethernet has established arbitration rules for resolving conflicts that arise when more than one station attempts to transmit at the same time. Both stations in a point-to-point full-duplex link are permitted to transmit at any time, regardless of whether the other station is transmitting at the same time. Autonegotiation avoids most situations in which one station in a point-to-point link is transmitting under half-duplex rules and the other is operating under full-duplex rules. Only two methods exist of achieving a full-duplex connection at speeds below 10 Gigabit Ethernet: ■ By using autonegotiation ■ By administratively forcing the interface mode If one station in a point-to-point link is autonegotiating and the other is not, the auto- negotiating station is required to select half duplex. Thus, if one end of a link is forced, it is incumbent upon the network support staff to force the other end as well. Failure to force both ends results in an artificially elevated error level and poor performance on the link. Duplex mismatches are perhaps the most common problem found on switched networks. Priority Resolution In the anticipated event that link partners share more than one common technology capability, the following list is used to determine which technology should be chosen from the offered configurations. In other words, it is desirable for 10-/100-/1000-Mbps 1102.book Page 289 Tuesday, May 20, 2003 2:53 PM 290 Chapter 5: Ethernet Fundamentals versions of copper-based Ethernet to agree automatically on the best way for two inter- faces to link. The list is ranked by priority, with the most desirable link configuration at the top: ■ Full duplex ■ Half duplex ■ Full duplex ■ Half duplex ■ Hull duplex ■ Half duplex Fiber-optic Ethernet implementations are not included in this priority resolution list because the interface electronics and optics do not permit easy reconfiguration between implementations. It is assumed that the interface configuration is fixed. If the two interfaces are capable of autonegotiating, they are already using the same Ethernet implementation, although there remain a number of configuration choices, such as the duplex setting or which station will act as the master for clocking purposes. Collision Domains and Broadcast Domains This section presents topics that relate to collision and broadcast domains. Excessive broadcasts and collisions can affect the network performance. This section discusses how to reduce the impact of broadcasts and collisions on the performance of the net- work. The term segment is defined, and different types of segmentation are discussed. Directly Connected Networks To understand collision domains, it is first necessary to examine the issues of what a collision is and how it is caused. To help explain collisions, it is useful to review Layer 1 media and topologies. As illustrated in Figure 5-24, some of the various types of directly connected networks include the following: ■ Shared-media environment—A shared-media environment occurs when multiple hosts have access to the same medium. For example, if several PCs are attached to the same physical wire or optical fiber, or share the same airspace, they all share the same media environment. In this case, they are said to share the same collision domain. Traditional bus-based (coaxial cable) Ethernet and hub-based Ethernet (UTP cable) are shared-media environments. 1102.book Page 290 Tuesday, May 20, 2003 2:53 PM Collision Domains and Broadcast Domains 291 ■ Extended shared-media environment—This is a special type of shared-media environment in which networking devices can extend the environment so that it can accommodate multiple access or longer cable distances. Using Ethernet repeaters or multiple hubs can create an extended shared-media environment. This creates an extended collision domain. ■ Point-to-point network environment—This is widely used in dialup network connections and is the most familiar to the home user. It is a shared networking environment in which one device is connected to only one other device, such as connecting a computer to an Internet service provider by modem and a phone line. Because this is a point-to-point dedicated connection, there is no potential for collisions. No other devices share the link. Figure 5-24 Directly Connected Networks Indirectly Connected Networks Some networks are indirectly connected, meaning that some higher-layer networking devices or some geographical distance is between the two communicating hosts. Fig- ure 5-25 also shows the two types of indirectly connected networks, which are described as follows: ■ Circuit-switched—An indirectly connected network in which actual electrical circuits are maintained for the duration of the communication. Circuit switching sets up a physical, end-to-end connection between the endpoints. The bandwidth Shared Media (Multiple Access) Extended Media (Multiple Access with Layer 1 Networking Device) Point-to-Point 1102.book Page 291 Tuesday, May 20, 2003 2:53 PM 292 Chapter 5: Ethernet Fundamentals is dedicated to this point-to-point connection. The current telephone system is still, in part, circuit-switched, although the telephone systems in many countries now are concentrating less on circuit-switched technologies. Because this is not a shared environment, there are no collisions. ■ Packet-switched—Instead of dedicating a link as an exclusive circuit connection between two communicating hosts, the source sends messages in packets. Each packet contains enough information for it to be routed to the proper destination host. Packet-switched networks frequently share physical media, but because log- ical point-to-point connections are created, there are no collisions. Figure 5-25 Indirectly Connected Networks Collisions and Collision Domains It is important to identify the medium as a shared environment because this shared environment causes collisions. A similar situation can occur with an automobile on a highway. If there is only one car, there is nothing to collide with. However, if more than one automobile is trying to use the same section of road at the same time, as shown in Figure 5-26, a collision occurs. The same is true for networks: If more than one computer tries to transmit data on the same network segment at the same time, a collision occurs. A collision is a situation that can occur when 2 bits propagate at the same time on the same network. A small, slow network could work out a system that allows only two computers to send messages, with both agreeing to take turns. The problem is that many computers are connected to large networks, with each one wanting to communicate millions of bits every section. Circuit-Switched Switching of Circuits Packet-Switched Switching of Packets 1102.book Page 292 Tuesday, May 20, 2003 2:53 PM Collision Domains and Broadcast Domains 293 Figure 5-26 Collision Collision domains are the connected physical network segments where collisions can occur. Collisions cause the network to be inefficient. Every time there is a collision on a network, all transmission stops for a period of time. This time is variable, as determined by a backoff algorithm for each network device, which is necessary to allow broadcast transmission to resume. Except for a single isolated Ethernet LAN environment, the types of devices that inter- connect the media segments define collision domains. These devices have been classified as OSI Layers 1, 2, or 3 devices. Layer 1 devices do not break up collision domains; Layers 2 and 3 devices do break up collision domains, as described in Figure 5-27. Breaking up or increasing the number of collision domains with Layer 2 and 3 devices is known as segmentation. Figure 5-27 Collision Domain Segmentation STOP STOP 1102.book Page 293 Tuesday, May 20, 2003 2:53 PM 294 Chapter 5: Ethernet Fundamentals Layer 1 devices, such as repeaters and hubs, serve the primary function of extending the Ethernet cable segments. By extending the network, more hosts can be added. How- ever, every host that is added increases the amount of potential traffic on the network. Because Layer 1 devices pass on everything that is sent on the medium, the more traffic that is transmitted within a collision domain, the greater the chances of collisions are. The final result is diminished network performance, which is even more pronounced if all the computers on that network demand large amounts of bandwidth. Simply put, Layer 1 devices extend collision domains, as shown in Figure 5-28, but the length of a LAN also can be overextended and can cause other collision issues. Figure 5-28 A Repeater Increases the Collision Domain The four-repeater rule in Ethernet states that no more than four repeaters or repeating hubs can be between any two computers on the network, as shown in Figure 5-29. To ensure that a repeated 10BASET network will function properly, the round-trip delay calculation must be within certain limits; otherwise, all the workstations will not be capable of hearing all the collisions on the network. Figure 5-29 A Repeater Increases the Collision Domain Collision Domain Collision Domain 1102.book Page 294 Tuesday, May 20, 2003 2:53 PM Collision Domains and Broadcast Domains 295 Repeater latency, propagation delay, and NIC latency all contribute to the four-repeater rule. Exceeding the four-repeater rule can lead to violating the maximum delay limit. When this delay limit is exceeded, the numbers of late collisions dramatically increase. A late collision occurs when a collision happens after the first 64 bytes of the frame are transmitted. The chip sets in NICs are not required to retransmit automatically when a late collision occurs. These late-collision frames add delay that is referred to as con- sumption delay. As consumption delay and latency increase, network performance decreases. This Ethernet rule of thumb also is known as the 5-4-3-2-1 rule. This means that the following guidelines should not be exceeded: ■ Host sections consist of — 5 sections of network media — 4 repeaters or hubs — 3 sections of the network ■ 2 sections are link sections (no hosts) ■ 1 large collision domain Segmentation For a networking professional, one important skill is the ability to recognize collision domains. Connecting several computers to a single shared-access medium that has no other networking devices attached creates a collision domain. This situation limits the number of computers that can use the medium, also called a segment. As illustrated in Figure 5-30, Layer 1 devices extend but do not control collision domains. Layer 2 devices segment or divide collision domains. Controlling frame propagation using the MAC address assigned to every Ethernet device performs this function. Layer 2 devices, bridges and switches, keep track of the MAC addresses and which segment they are on. By doing this, these devices can control the flow of traffic at the Layer 2 level. This function makes networks more efficient by allowing data to be transmitted on different segments of the LAN at the same time, without the frames colliding. By using bridges and switches, the collision domain effectively is broken up into smaller parts, each being its own collision domain. These smaller collision domains will have fewer hosts and less traffic than the original domain, and thereby increase the amount of bandwidth available to each host in that domain. The lower the amount of traffic is in a collision domain, the greater the chance there is that when a host wants to transmit data, the media will be available. This works well as long as the traffic between segments is not too heavy. Otherwise, the Layer 2 device actually can slow communication and become a bottleneck itself. 1102.book Page 295 Tuesday, May 20, 2003 2:53 PM 296 Chapter 5: Ethernet Fundamentals Figure 5-30 Layer 1 Devices Extend the Collision Domains Layer 3 devices, like Layer 2 devices, do not forward collisions. Because of this, the use of Layer 3 devices in a network has the effect of breaking up collision domains into smaller domains. Layer 3 devices perform more functions than just breaking up a col- lision domain. These devices and their functions are covered in more depth in the sec- tion, “Broadcast Domains.” Figure 5-31 illustrates that Layer 2 and Layer 3 devices can break up the collision domain. 1102.book Page 296 Tuesday, May 20, 2003 2:53 PM Collision Domains and Broadcast Domains 297 Figure 5-31 Limiting the Collision Domain Layer 2 Broadcasts To communicate with all collision domains, protocols use broadcast and multicast frames at Layer 2 of the OSI model. When a node needs to communicate with all hosts on the network, it sends a frame with a destination MAC address 0xFFFFFFFFFFFF (a broadcast). This is an address to which the NIC of every host must recognize. Layer 2 devices must flood all broadcast and multicast traffic. The accumulation of broadcast and multicast traffic from each device in the network is referred to as broad- cast radiation. Figure 5-32 illustrates a bridge forwarding the broadcast to all hosts on the network. Because the NIC must interrupt the CPU to process each broadcast or multicast group that it belongs to, broadcast radiation affects the performance of hosts in the network. Most often, the host does not benefit from processing the broadcast because it is not the destination being sought. Either the host does not care about the service that is being advertised or it already knows about the service. High levels of broadcast radia- tion noticeably can degrade host performance, as shown in Figure 5-33. The three sources of broadcasts and multicasts in IP networks are workstations, routers, and multicast applications. Workstations broadcast an Address Resolution Protocol (ARP) request every time they need to locate a MAC address that is not in the ARP table. For example, the command telnet mumble.com translates into an IP address through a Domain Name System (DNS) search, and then an ARP request is broadcast to find the actual station. Generally, IP workstations cache 10 to 100 addresses about 2 hours. The ARP rate for a typical workstation might be about 50 addresses every 2 hours, or 0.007 ARPs per second. Thus, 2000 IP end stations produce about 14 ARPs per second. Collision Domain Collision Domain Or Or 1102.book Page 297 Tuesday, May 20, 2003 2:53 PM 298 Chapter 5: Ethernet Fundamentals Figure 5-32 Layer 2 Broadcast Figure 5-33 Effect of Broadcast Radiation on Hosts in IP Network chpt_05.fm Page 298 Tuesday, May 27, 2003 9:09 AM . endpoints. The bandwidth Shared Media (Multiple Access) Extended Media (Multiple Access with Layer 1 Networking Device) Point-to-Point 11 02. book Page 2 91 Tuesday, May 20 , 20 03 2: 53 PM 29 2 Chapter. Layer 2 and 3 devices is known as segmentation. Figure 5 -27 Collision Domain Segmentation STOP STOP 11 02. book Page 29 3 Tuesday, May 20 , 20 03 2: 53 PM 29 4 Chapter 5: Ethernet Fundamentals Layer 1. the offered configurations. In other words, it is desirable for 10 - /10 0- /10 00-Mbps 11 02. book Page 28 9 Tuesday, May 20 , 20 03 2: 53 PM 29 0 Chapter 5: Ethernet Fundamentals versions of copper-based