| Network Hardware |
 |
Network hardware enables networking devices—nodes—to cable together into a network. That’s pretty straightforward and even redundant, at least in words. The CompTIA Network+ exam expects you to know the network hardware used in Ethernet and TCP/IP networks in some detail, even the hardware no longer in use. This chapter looks at network interfaces and the magic boxes that connect devices, both wired and wirelessly.
 |
Network Interfaces |
A network interface enables a computing device to send and receive data across a network. All modern computing devices have a network interface, either wired to connect to an Ethernet network or wireless to connect to a wireless access point. Traditionally called a network interface card (NIC) because the interface came on a separate expansion card, today most devices have built-in NICs (though most techs still call them NICs). Figure 4.1 shows a network interface on a motherboard.
FIGURE 4.1 Gigabit Ethernet network interface
NICs have several jobs. They provide a connection to the network media, of course, either through cables or wireless radio frequency waves. NICs also have a physical address to enable communication with other devices on the network. Finally, NICs take data from the operating system and encapsulate that data into a frame suitable for traversing the Physical-layer network.
MAC Address
Inside every NIC, burned onto some type of ROM chip, is special firmware containing a unique identifier with a 48-bit value called the media access control address, or MAC address.
No two NICs ever share the same MAC address—ever. Any company that makes NICs must contact the Institute of Electrical and Electronics Engineers (IEEE) and request a block of MAC addresses, which the company then burns into the ROMs on its NICs. Many NIC makers also print the MAC address on the surface of each NIC, as shown in Figure 4.2. Note that the NIC shown here displays the MAC address in hexadecimal notation. Count the number of hex characters—because each hex character represents 4 bits, it takes 12 hex characters to represent 48 bits.
The MAC address in Figure 4.2 is 004005–607D49, although in print, we represent the MAC address as 00–40–05–60–7D–49. The first six digits, in this example 00–40–05, represent the number of the NIC manufacturer. Once the IEEE issues those six hex digits to a manufacturer—often referred to as the organizationally unique identifier (OUI)—no other manufacturer may use them. The last six digits, in this example 60–7D–49, are the manufacturer’s unique serial number for that NIC; this portion of the MAC is often referred to as the device ID.
Would you like to see the MAC address for your NIC? If you have a Windows system, type ipconfig /all from a command prompt to display the MAC address (see Figure 4.3). Note that ipconfig refers to the MAC address as the physical address.
FIGURE 4.3 Output from ipconfig /all
Local Lingo
MAC-48 and EUI-48 IEEE forms MAC addresses from a numbering name space originally called MAC-48, which simply means that the MAC address will be 48 bits, with the first 24 bits defining the OUI, just as described here. The current term for this numbering name space is EUI-48. EUI stands for Extended Unique Identifier. (IEEE apparently went with the new term because they could trademark it.)
Exam Tip
Most techs just call them MAC addresses, as you should, but you might see MAC-48 or EUI-48 on the CompTIA Network+ exam.
Travel Assistance
EUI-64 Although we’re not expected to exhaust the supply of 48-bit MAC addresses until the next century, IEEE has released a 64-bit successor called, appropriately, EUI-64. The numbering scheme is the same, but adds a couple more octets to bring it up to 64 bits. Currently, the only common uses for EUI-64 are FireWire and IPv6. (See Chapter 5 for information on the latter.)
Organizing the Data: Ethernet Frames
All network technologies break data transmitted between computers into smaller pieces called frames, as you’ll recall from Chapter 1. Functioning at the Data Link layer (layer 2), NICs encapsulate data from the higher levels of the OSI model into frames to enable communication at the Physical layer.
Using frames addresses two networking issues. First, frames prevent any single machine from monopolizing the shared bus cable. Second, they make the process of retransmitting lost data more efficient.
Looking at the first scenario, if a sending computer sends a file as a single huge frame, the frame will monopolize the cable and prevent other machines from using it until the entire file gets to the receiving system. Using relatively small frames enables computers to share the cable easily—each computer listens on the segment, sending a few frames of data whenever it detects that no other computer is transmitting.
Second, in the real world, bad things can happen to good data. When errors occur during transmission, the sending system must retransmit the frames that failed to get to the receiving system in good shape. If a word processing document were transmitted as a single massive frame, the sending system would have to retransmit the entire frame—in this case, the entire document. Breaking the file up into smaller frames enables the sending computer to retransmit only the damaged frames.
Because of these benefits—shared access and more efficient retransmission—all networking technologies use frames, and Ethernet is no exception to that rule. A basic Ethernet frame contains six pieces of information: the preamble, the MAC address of the frame’s recipient, the MAC address of the sending system, the type of data, the data itself, and a frame check sequence. Figure 4.4 shows these components.
FIGURE 4.4 Ethernet frame
Preamble
All Ethernet frames begin with a preamble, a 64-bit series of alternating ones and zeroes that ends with 11. The preamble gives a receiving NIC time to realize a frame is coming and to know exactly where the frame starts.
Destination and Source MAC Addresses
The second and third portions of the frame identify the destination and sending NICs, respectively. In early Ethernet networks, every node on the network received every frame because they were on a shared bus cable. Nodes would use the destination MAC address to determine whether to process each frame. In today’s switched Ethernet, the switch keeps a record of which node connects to which switch port and forwards data to the precise destination node. The only common traffic that goes to all nodes is broadcast traffic sent to the destination MAC of FF-FF-FF-FF-FF-FF, the broadcast address.
Travel Advisory
Switches Sending Multiple Frames Switches send frames to all nodes in two nonbroadcast circumstances: unknown unicasts and multicasts. If the switch doesn’t recognize the MAC address, it sends the frame to every node. Additionally, some switches can’t filter multicasts and instead let the NICs handle the filtering.
Type
An Ethernet frame may carry one of several types of data. The Type field helps the receiving computer interpret the frame contents at a very basic level. This way, the receiving computer can tell if the frame contains, for example, IPv4 data or IPv6 data. (See Chapter 5 for more details on IPv4 and IPv6.)
The Type field does not tell you if the frame carries higher-level data, such as an e-mail message or webpage. You have to dig deeper into the data section of the frame to find that information.
Data
The data part of the frame contains whatever the frame carries, what’s called the payload. This can be an IP packet or some other type of packet used on the network. Regardless of the type of data in the payload, an Ethernet frame must have a payload of at least 46 bytes. If the payload is smaller than 46 bytes, the sending NIC will automatically add extra bytes—a pad—to bring the payload up to the minimum.
Frame Check Sequence
The frame check sequence (FCS)—Ethernet’s term for the cyclic redundancy check—enables Ethernet nodes to recognize when bad things happen to good data. Machines on a network must be able to detect when data has been damaged in transit.
To detect errors, the computers on an Ethernet network attach a special code to each frame. When creating an Ethernet frame, the sending machine runs the data through a special mathematical formula and attaches the result, the frame check sequence, to the frame. The receiving machine opens the frame, performs the same calculation, and compares its answer with the one included with the frame. If the answers do not match, the receiving machine asks the sending machine to retransmit that frame.
 |
Repeaters, Hubs, and Bridges |
Early 10-Mbps Ethernet networks ran through various evolutionary progressions, as you’ll recall from Chapter 3, and clever network designers came up with many ways to link any and all of the varieties together. They came up with devices to extend the bus topology networks from 500 m or 185 m to double or more in length. They invented devices to accommodate 10BaseT networks and sit in the center of the star bus topology.
A repeater takes in an electronic signal, corrects any voltage drops, and then retransmits the cleaned signals. That electronic signal with Ethernet is the frame. A repeater increases the maximum possible distance between machines by linking together two or more segments.
A hub enables a specific number of devices to connect to the network (one per port), and often, but not always, includes a separate port to enable hubs to be linked together to extend the network. Some hubs also include connectors for different media types—BNC, RJ-45, or fiber—but they cannot interface between networks that are based on different technologies.
Hubs are multiport repeaters, sending out every frame that one node sends to every other node connected to the hub. As a result, hubs enable multiple devices to be in the same network, but they’re also all in the same collision domain and the rules of CSMA/CD apply.
Exam Tip
Repeaters and hubs are considered layer-1 (Physical-layer) devices because they work with the electrical signal.
A bridge acts like a repeater or hub to connect two Ethernet segments, but it goes one step beyond—filtering and forwarding traffic between those segments based on the destination MAC addresses. To filter traffic means to stop it from crossing from one network to the next; to forward traffic means to pass traffic originating on one side of the bridge to the other.
Exam Tip
Because bridges work with MAC addresses, they operate at layer 2, the Data Link layer, of the OSI networking model. They function in the Link/Network Interface layer of the TCP/IP model.
A newly installed Ethernet bridge initially behaves exactly like a repeater, passing frames from one segment to another. Unlike a repeater, however, a bridge monitors and records the network traffic, eventually reaching a point where it can begin to filter and forward. A bridge usually requires only a few seconds to gather enough information to start filtering and forwarding.
Although bridges offer a good solution for connecting two segments and reducing bandwidth usage, these days you’ll mainly find bridges used in wireless, rather than wired, networks. (I cover those kinds of bridges at the end of this chapter.)
 |
Switched Ethernet |
While plain-vanilla 101 BaseT Ethernet performed well enough for first-generation networks (which did little more than basic file and print sharing), by the early 1990s networks used more-demanding applications, such as Lotus Notes, SAP business management software, and Microsoft Exchange, which quickly saturated a 10BaseT network. Fortunately, those crazy kids over at the IEEE kept expanding the standard, giving the network tech in the trenches a new tool that provided additional bandwidth—the switch.
The Trouble with Hubs
A classic 10BaseT network with a hub can only have one message on the wire at any time. When two computers send at the same time, the hub dutifully repeats both signals. The nodes recognize the collision and, following the rules of CSMA/CD, attempt to resend. Add in enough computers, and the number of collisions increases, lowering the effective transmission speed for the whole network. A busy network becomes a slow network because all the computers share the same collision domain.
Switches to the Rescue
An Ethernet switch looks like a hub because all nodes plug into it (Figure 4.5). But switches don’t function like hubs inside. Switches come with extra smarts that enable them to take advantage of MAC addresses, effectively creating point-to-point connections between two conversing computers. This gives every conversation between two computers the full bandwidth of the network.
FIGURE 4.5 Hub (top) and switch (bottom) comparison
To see a switch in action, check out Figure 4.6. When you first turn on a switch, it acts exactly as though it were a hub, passing all incoming frames right back out to all the other ports. As it forwards all frames, however, the switch copies the source MAC addresses and quickly creates an electronic table of the MAC addresses of each connected computer. The table is called a source address table (SAT).
FIGURE 4.6 A switch tracking MAC addresses
Exam Tip
One classic difference between a hub and a switch is in the repeating of packets during normal use. Although it’s true that switches initially forward all frames, they filter by MAC address in regular use. Hubs never learn and always forward all frames.
As soon as this table is created, the switch begins to do something amazing. When a computer sends a frame into the switch destined for another computer on the same switch, the switch acts like a telephone operator, creating an on-the-fly connection between the two devices. While these two devices communicate, it’s as though they are the only two computers on the network. Figure 4.7 shows this in action. Because the switch handles each conversation individually, each conversation runs at 10 Mbps.
FIGURE 4.7 A switch making two separate connections
Each port on a switch is in its own collision domain, plus the switch can buffer incoming frames. That means that two nodes connected to the switch can send data at the same time and the switch will handle it without any collision.
With half-duplex switches, collisions can occur and the rules of CSMA/CD apply. These collisions can only happen between the switch and a node, not between two nodes. If the switch forwards a frame to a node from any other node at the same time that the node tries to send a frame to any node through the same switchport, that’s a collision.
Network developers eventually figured out how to make switches and NICs run in full-duplex mode so they could send and receive data at the same time. With full-duplex Ethernet, CSMA/CD is disabled and no collisions can occur. Each node will always get the full bandwidth of the network.
Local Lingo
Layer 2 Switch Because a switch filters traffic on MAC addresses (and MAC addresses run at layer 2 of the OSI seven-layer model), they are sometimes called layer 2 switches.
Unicast messages—from one node specifically to another node—usually go only to the intended recipient when you use a switch. (The exception is when the switch does not have a MAC address in its SAT, as mentioned earlier.) The switch will send all broadcast messages to all the ports. You’ll commonly hear a switched network called a broadcast domain to contrast it to a hub-based network with its collision domain.
Spanning Tree Protocol
Because you can connect switches together in any fashion, you can create redundant connections in a network. These are called bridging loops (Figure 4.8).
FIGURE 4.8 A bridging loop
Exam Tip
The CompTIA Network+ exam refers to bridging loops as switching loops. The terms mean the same thing, but bridging loop is more common. Be prepared for either term on the exam.
In the early days of switches, making a bridging loop in a network setup would bring the network crashing down. A frame could get caught in the loop, so to speak, and not reach its destination.
The Ethernet standards body adopted the Spanning Tree Protocol (STP) to eliminate the problem of accidental bridging loops. Switches with STP enabled can detect loops, communicate with other switches, and set the looped port’s state to blocking.
STP-enabled switches use a frame called a Bridge Protocol Data Unit (BPDU) to communicate with each other to determine things like the distances between them and to keep track of changes on the network.
Travel Assistance
Switches today all have STP enabled, and network designers create bridging loops in their networks to provide fault tolerance. Ports set as blocking still listen to the traffic on the network. If a link fails, the blocking port can become a forwarding port, thus enabling traffic to flow properly.
 |
Connecting Ethernet Segments |
You can combine multiple switches to create much bigger broadcast domains than any one switch can support. You can do this in two ways without using a bridge: via an uplink port or a crossover cable.
Uplink Ports
Uplink ports enable you to connect two switches using a straight-through cable. They’re always clearly marked on the switch, as shown in Figure 4.9. To connect two switches, insert one end of a cable to the uplink on one switch and the other end of the cable to any one of the regular ports on the other switch. To connect more than two switches, you must daisy-chain your switches by using one uplink port and one regular port. Figure 4.10 shows properly daisy-chained switches.
FIGURE 4.9 Typical uplink port
FIGURE 4.10 Daisy-chained switches
Working with uplink ports is sometimes tricky, so you need to take your time. Switch makers give their uplink ports many different names, such as crossover, MDI-X, and OUT. There are also tricks to using uplink ports. Refer to Figure 4.9. See the line connecting the uplink port and the port labeled 2X? You may use only one of those two ports, not both at the same time. Additionally, some switches place a button on one of the ports; you press this button to make it either a regular port or an uplink port (Figure 4.11). Pressing the button electronically reverses the wires inside the switch.
FIGURE 4.11 Pressing the Normal/Uplink button toggles the port between regular and uplink
Exam Tip
Two terms you might see on hubs and switches and, consequently, on the exams: MDI and MDIX. A media-dependent interface (MDI) is a regular port on a hub or switch. A media-dependent interface crossover (MDIX) is an uplink port.
Crossover Cables
Switches can also connect to each other via special twisted-pair cables called crossover cables. A standard cable cannot be used to connect two switches without using an uplink port because both switches will attempt to send data on the second pair of wires (3 and 6) and will listen for data on the first pair (1 and 2). A crossover cable reverses the sending and receiving pairs on one end of the cable. One end of the cable is wired according to the T568A standard, whereas the other end is wired according to the T568B standard (Figure 4.12). With the sending and receiving pairs reversed, the switches can hear each other; hence, the need for two standards for connecting RJ-45 jacks to UTP cables.
FIGURE 4.12 A crossover cable reverses the sending and receiving pairs.
A crossover cable connects to a regular port on each switch. Keep in mind that you can still daisy-chain even when you use crossover cables.
In a pinch, you can use a crossover cable to connect two computers together using 10BaseT NICs with no switch between them at all. This is handy for quickie connections, such as for a nice little home network or when you absolutely, positively must chase down a friend in a computer game!
Be careful about confusing crossover cables with uplink ports. First, never connect two switches by their uplink ports with a straight-through cable. This creates two crossings, and you always need an odd amount of crossings for this to work. Take a straight-through cable; connect one end to the uplink port on one switch and the other end to any regular port on the other switch. Second, if you use a crossover cable, just plug each end into any handy regular port on each switch.
Travel Advisory
Crossing Crossovers If you mess up your crossover connections, you won’t cause any damage, but the connection will not work. Think about it. If you take a straight-through cable (that is, not a crossover cable) and try to connect two PCs directly, it won’t work. Both PCs will try to use the same send and receive wires. When you plug the two PCs into a switch, the switch electronically crosses the data wires, so one NIC sends and the other can receive. If you plug a second switch to the first switch using regular ports, you essentially cross the cross and create a straight connection again between the two PCs! That won’t work. Luckily, nothing gets hurt—except your reputation if one of your colleagues notes your mistake!
 |
Routers |
Routers connect networks together. They filter and forward traffic relying on layer-3 addresses to get the traffic to the proper network. Because of the more sophisticated addressing schemes used by routers, they can support multiple routes between networks. The networks connected by the routers can be different types of networks (for example, Ethernet and DOCSIS).
Local Lingo
DOCSIS Cable companies use the network standard called DOCSIS to enable communication via coax and cable modems. You’ll learn more about this in Chapter 8.
Router technology holds together the Internet, providing multiple pathways between the hundreds of networks that make up the entire system (see Figure 4.13). Most routers, however, see much more humble service doing the same kinds of things as bridges, linking networks or sites to form a wide area network (WAN).
FIGURE 4.13 A network with multiple routers
Exam Tip
Routers operate at layer 3 of the OSI model, also known as the Network layer.
Routers rely on layer-3 addressing, also called network addressing. A network address tells the router two pieces of information: the specific machine to which a packet should be delivered, and the network on which that machine resides. The IP addresses used on the Internet are examples of network addresses (see Chapter 5 for a discussion of IP addressing).
Given the destination network address, a properly configured router can determine the best route to the destination machine. In Figure 4.13, router 1 needs to deliver a packet from machine A to machine D. Assuming that all the links between the routers operate at the same speed, the most efficient route for the packet should be router 1 to 5 to 4 to 6. That route requires four hops. A hop is the process of passing through a router en route to the final destination. Alternative routes exist, but the router ignores them because they require more hops. In the event of a break in the link between routers 1 and 5, router 1 should automatically calculate the next-best route (if one exists) and redirect traffic to the alternative route: router 1 to 2 to 3 to 4 to 6 (see Figure 4.14).
FIGURE 4.14 Properly configured routers choose the best route.
Travel Assistance
You’ll find more discussion on routers and routing in Chapter 6. I can’t do them justice here before giving you details about addressing in a TCP/IP network, which you’ll get in Chapter 5.
 |
Wireless Access Points |
A wireless access point (WAP) enables nodes with Wi-Fi NICs to connect into a LAN using radio waves. WAPs support one or more 802.11 standards and nodes that use those standards. Many WAPs support multiple standards, thus enabling a variety of Wi-Fi devices to connect and network. All 802.11n WAPs, for example, can support devices running 802.11n, 802.11g, and 802.11b; some support 802.11a devices as well.
WAPs provide a bridge between wired and wireless networks. Wired nodes using CSMA/CD can communicate seamlessly with wireless nodes using CSMA/CA, and vice versa.
WAPs designed for small office/home office (SOHO) environments generally come as part of a much larger bundle of hardware, usually called a “router,” though routing is but one of the functions. The Linksys box in Figure 4.15, for example, is a combination WAP, router, and four-port switch, plus it offers excellent security features through its firmware.
FIGURE 4.15 Linksys WAP, router, and switch, all-in-one box
Objective 4.01: Network Interfaces A network interface card gets you on the network, but it must be the right fit for your computer system or whatever you’re plugging it into. You must also make sure that it has the right connectors for the network media and that it supports the right network standard and speeds—for example,100/1000 Mbps Ethernet. One very important task performed by a NIC is to give your computer system a unique address in the network; this is known as the media access control (MAC) address. This address is unique on every NIC produced, but the top 24 bits of the MAC address can be used to identify the NIC (or NIC component) manufacturer. Objective 4:02: Repeaters, Hubs, and Bridges Repeaters provide a simple way to extend a bus-based network over the basic 185-meter (Thinnet) or 500-meter (Thicknet), 30/100-device (Thin/Thick) segment maximums. Hubs behave like multiport repeaters to connect machines to a star bus network. Both devices have no specific understanding of the data they are passing between ports and therefore do not provide any form of traffic management.
Bridges can link together segments of a network, just like a repeater, but they also build up an internal MAC address table, identifying the computer systems on either side of their interfaces so they can determine whether a data packet actually needs to pass across the bridge; this helps with traffic management and thus improves overall network performance. Bridges pass broadcast network traffic because they break up the collision domain, but not the broadcast domain.
Objective 4.03: Switched Ethernet Switches are layer-2 devices that act like a bridge, with each port on the switch creating a network segment. This means that each port on the switch is its own collision domain. The switch filters traffic by sending the data only to the port that has the destination MAC address connected to it. This increases overall network performance and adds to the security of the network.
Objective 4.04: Connecting Ethernet Segments You can combine multiple switches to create bigger broadcast domains in two ways without using a bridge: via an uplink port or a crossover cable. To connect two switches, insert one end of a straight-through cable to the uplink and the other end of the cable to any one of the regular ports on the other switch. The CompTIA Network+ name for an uplink port is MDIX. To connect using a crossover cable, plug the ends of the cable into regular ports on each switch.
Objective 4.05: Routers Routers can do all the things that bridges can do, but they also support multiple paths between networks and can link together dissimilar network types, such as Ethernet and fiber. Routers are often used for wide area networking, providing interconnectivity between networked sites. Routers do not pass broadcast traffic; instead, they manage the identification of machine locations themselves, using network addresses that contain both a machine’s address and the network on which it is located.
Objective 4.06: Wireless Access Points Wireless access points are popular network devices that allow a wireless client to connect to a wired network. Today’s WAPs offer a number of capabilities such as routing and switching.
REVIEW QUESTIONS
1. What is the minimum size of an Ethernet frame?
A. 8 bytes
B. 64 bytes
C. 128 bytes
D. 256 bytes
2. Which part of the frame contains information used for checking the validity of the frame?
A. FCS
B. Pad
C. Preamble
D. Type
3. Which device enables you to connect two networks, but discards broadcast traffic?
A. Bridge
B. Hub
C. Switch
D. Router
4. At which layer of the OSI model do switches work?
A. Layer 1
B. Layer 2
C. Layer 3
D. Layer 4
5. How are the connectors wired on a crossover cable?
A. One end is T568A; the other end is T568B.
B. Both ends are T568A.
C. Both ends are T568B.
D. One end is an RJ-45; the other end is an RG-6.
6. What is the purpose of a preamble in an Ethernet frame?
A. It gives the receiving NIC time to realize a frame is coming and to know when the frame starts.
B. It provides the receiving NIC with the sending NIC’s MAC address so communication can continue.
C. It provides error-checking to ensure data integrity.
D. It contains a description of the data that is to follow so the receiving NIC knows how to reassemble it.
7. Which device enables you to connect two networks that use different technologies?
A. Hub
B. Repeater
C. Router
D. Switch
8. Upon looking at the front of a switch, you notice something labeled MDI-X. What is this for?
A. It is a special receptacle for the power cable.
B. It is a regular port used to connect computers.
C. It is an uplink port used to connect the switch to another switch.
D. It is the brand name of the switch.
9. What feature of switches prevents the problem of bridging loops?
A. STP
B. TCP/IP
C. IEEE 802.3
D. UTP
10. What feature of switches keeps track of which MAC address goes to each port?
A. FCS
B. SAT
C. STP
D. UTP
REVIEW ANSWERS
1. B The minimum size of an Ethernet frame is 64 bytes.
2. A The frame check sequence FCS contains error-checking information.
3. D Routers connect networks, but discard broadcast traffic.
4. B Switches work at layer 2, the Data Link layer, because they manage traffic using MAC addresses.
5. A A properly wired crossover cable is T568A on one end and T568B on the other. All T568A or all T568B defines a straight-through cable. RJ-45 is for UTP; RG-6 is for coaxial. You wouldn’t find them on the same cable.
6. A The preamble gives the receiving NIC time to realize a frame is coming and to know when the frame starts.
7. C A router enables networks to connect even if the networks use different technologies.
8. C Uplink ports are often labeled MDI-X.
9. A Spanning Tree Protocol (STP) prevents bridging loops. STP also stands for shielded twisted-pair; knowing the context in which the acronym is being used is important.
10. B A switch keeps track of which MAC address goes to each port with its source address table (SAT).
