Chapter 9

1. Chapter 9 Local Area Networks

2. Introduction • Network- a set of two or more interconnected devices or computers • LAN may consist of PCs and/or MACs, mainframes, minicomputers, etc. • Local area is typically within an office, a building, or a group of buildings. The distance among the computers and devices is only one characteristic that makes a network a local area network. • Due to new technologies, LANs now often span much longer distances, up to many miles in some cases. chapter 9

3. Figure 9.1 a networked arrangement and a single multi-user system chapter 9

4. Introduction Continued • What features differentiate a LAN from a wide area, WAN? • Generally, WANs are geographically larger than LANs. • Often WANs are used to connect many more computers than LANs • LANs are owned and operated by a single company mostly, and their wiring, systems, and devices are on the company's premises • WAN is often provided by common carriers such as the telephone companies • However, many companies provide their own private WANs by leasing or purchasing transmission equipment. • LANs make use of media access techniques that are different than the physical interfaces to WAN. LANs may use different protocols from WANs • The interconnection of LANs over WANs is very common in corporations today • New networking and switching technologies are blurring the lines between the two types of networks. chapter 9

5. Motivation for Using a LAN • Two primary reasons: • to share resources such as printers or files • to improve communications between users in a workgroup, office, department, or company • Disk sharing: • provides access to commonly used programs reducing total disk space requirements • provides access to commonly needed data. • With only a shared "copy" of the information, each user is assured that it is always accurate and up-to-date. • allows centralized backup of files chapter 9

6. Figure 9.2: In addition to saving on hardware, LAN also results in software savings chapter 9

7. Motivation for Using a LAN Continued • LAN can also offer improved security. • Security features such as passwords are built into the servers • System attached to the LAN could be diskless workstation (no floppy disk drive). • No software or data resides on the system, and none can be copied to a floppy disk and taken away. • Also limits the introduction of viruses or loading unauthorized programs. chapter 9

8. Motivation for Using a LAN Continued • Improves communications • Electronic mail and workgroup applications • Documents, programs, and data files exchanged as attachments using electronic mail among LAN users. • Workgroup applications allow multiple users to cooperate in performing various tasks chapter 9

9. Components of Local Area Networks • Often consist of interconnected computers including personal computers, UNIX workstations, multi-user systems, and even large mainframe computers. • Computers containing shared resources are called servers. • The users and applications running on other systems access these servers over the LAN, and are called clients, hence the name client server computing chapter 9

10. Terminology • Standalone workstation - a workstation that is not connected to a network, but relies on its own hard disk for data storage and applications. • Client - a workstation connected to a network. A person whose workstation is part of a network may also be called a client, or that person may be known more informally as a user. • Servers - store shared data and programs on their hard disks. They can also perform management functions, such as determining which users have access to certain programs. • Client-server - a network that uses a server to enable clients to share data, data storage space, and devices. chapter 9

11. Figure 9-3: Client Server LAN Configuration chapter 9

12. Components of Local Area Networks Continued • Print server- this computer accepts print requests from other systems on the LAN, temporarily writes the data to be printed onto its disk storage, then sends the data to the printer. • The print server manages the print streams from several systems concurrently • A directly attached printer acts as a print server. Other systems on the LAN can direct print streams to the printer. chapter 9

13. Components of Local Area Networks Continued • A LAN adapter is a hardware board (e.g., a PC adapter card) that can be inserted into an expansion slot in the PC. • LAN adapters are commonly called network interface adapters (NIAs) or network interface cards (NICs). • The adapter provides the interface to the PC or device on one side and the network on the other side. • The adapter provides a socket into which the LAN medium, such as a cable, can plug. • The adapter has to translate the signals used within the PC or printer into signals used on the LAN cable. chapter 9

14. Network Interface Cards (NICs) Types of NICs • Industry Standard Architecture (ISA) • MicroChannel Architecture (MCA) • Extended Industry Standard Architecture (EISA) • Peripheral Component Interconnect (PCI) FIGURE: Four primary bus architectures

15. Network Interface Cards (NICs) FIGURE : Three kinds of bus connections on the same board

16. Network Interface Cards (NICs) • NICs may connect to interfaces other than a PC’s bus • For laptop computers, Personal Computer Memory Card International Association slots may be used to connect NICs • PCMCIA • Also called PC card • Developed in the early 1990s to provide a standard interface for connecting any type of device to a portable computer

17. Network Interface Cards (NICs) FIGURE: Typical PCMCIA NIC FIGURE : Parallel port NIC

18. Network Interface Cards (NICs) FIGURE : Wireless NIC and transceiver FIGURE : Ethernet NICs for printers

19. Basics of Local Area Networks • IEEE 802-series of standards is focused on LAN interconnection and operation • The IEEE LAN standards are specific set of standards that conform to the lower layers of the OSI reference Model • The most popular types of LAN physical media are coaxial cable, twisted-pair wire, and fiber optic cable chapter 9

20. Coax in LANs • The original type of coaxial cable used in Ethernet LANs is known as thick coax, approximately 1/2 inch thick and is relatively inflexible and difficult to install. • Main advantage of coaxial cable - less susceptible to interference than twisted-pair wire, and supports relatively high rates of data transmission over greater distances. • The main disadvantage - more expensive that other media, and in the case of thick coax, its size and inflexibility makes it more difficult to install. • Typical data rates over coaxial cable LANs is 10 Mbps with distances ranging from 100's to 1000's of feet. • "thin" coax came into widespread use because it is less expensive than thick coax, and easier to work with. • It is approximately 1/4 inch (0.63 cm) thick and very similar to the coax used for cable TV. • But, this coax supports shorter distances than the thick coax. chapter 9

21. Twisted-pair Cables in LANs • Unshielded twisted-pair (UTP) is the standard telephone cable • Main advantages of UTP wiring - inexpensive, flexible, easy to install and available • In the past, only relatively low speed transmissions were possible over UTP but newer techniques are now in use that will support UTP speed in the 100 megabits per second range. • Disadvantages: • more susceptible to electrical interference. Not a huge problem in offices, but can be a problem in factories where electrical machinery is in use. • Another problem is that signals lose their strength as they are transmitted over UTP, high attenuation. chapter 9

22. Twisted-pair Cables in LANs Continued shielded twisted-pair, STP: • shield provides protection from noise, thus eliminating somewhat the problems associated with susceptibility to noise that plagues UTP • shield also helps keep signals from emanating out of the wire, important in certain environments. • A negative of STP though, is that the shield increases the cost of the wire so it is typically more expensive than UTP chapter 9

23. Fiber Optic cabling in LANs • Advantages: • high transmission rates, up to 100 Mbps and therefore greater bandwidth • immune to electrical interference because they use light rather than electricity • very thin and flexible. Thus, a fiber optic cable is easy to install, and ideal for bundling many fibers together to create a cable that carries very large amount of traffic. • Less attenuation than copper wiring (longer length of a fiber optic cable before repeaters are required). • Very secure. Any tap into the cable interrupts the flow of light and is easily detected. • Disadvantages: cables and adapters more expensive • Commonly used for high speed backbones chapter 9

24. Transmission Techniques • Baseband: • Signals placed directly onto the transmission medium (not modulated by a carrier) • A stream of such digital pulses represents the information being transmitted. • The signal takes up the entire bandwidth of the transmission medium. • Signals from multiple sources can be transmitted via the technique known as Time Division Multiplexing or TDM as seen in Chapter 2. • Relatively inexpensive. No special equipment is required to modify the signals. • One problem -signals lose strength as they are carried over longer distances and must be regenerated. chapter 9

25. Figure 9.5: Baseband Transmission chapter 9

26. Transmission Techniques Continued • Broadband: • signals modulated to different frequency ranges typically provided using radio frequency modems. • Frequency Division Multiplexing (FDM) is used where each signal occupies a different frequency range. • The different frequency ranges called logical channels. • Therefore, multiple channels are available using broadband transmission. (technique used in standard cable TV). • Each logical FDM channel could also be shared by multiple applications by using TDM within the channel. chapter 9

27. Transmission Techniques Continued • Broadband provides: • greater bandwidth than baseband and is, therefore, able to carry more information and support greater distances than baseband. • ability to carry voice, data, and video at the same time. • Disadvantages: • Typically more expensive than baseband transmission. • more difficult to configure and costly to modify. • Best suited to large installations chapter 9

28. Figure 9.6: Broadband Transmission chapter 9

29. Media Access Methods • All devices attached to a LAN share the transmission medium. • What happens if multiple devices attempt to send data onto the LAN at the same time? • A media access control (MAC) method determines how multiple devices share the transmission medium. chapter 9

30. Media Access Methods Continued • Carrier Sense Multiple Access with Collision Detection (CSMA/CD) • Devices must sense the presence of a carrier signal (presence of a carrier signal indicates that information is currently being transmitted) • If a carrier signal is not detected, it is an indication that the LAN is free and the device can attempt to send. • If another device had sent information at the same time, though, the listening devices would detect a collision. This is where the "collision detection" part of the name comes from. • When such a collision is detected, the devices stop transmitting and wait for some period of time before attempting to transmit again. chapter 9

31. Media Access Methods CSMA/CD Continued • CSMA/CD is the access method used in Ethernet and 802.3 LANs and became an international standard when IEEE 802.3 was approved. • Traffic increases may result in much more frequent collisions degrading network performance. • Rule of thumb: CSMA/CD LANs (Ethernet) work fine with constant traffic load of 30 percent of capacity and traffic bursts of about 60 percent of capacity; • otherwise divide the Ethernet into additional segments to reduce the number of stations that are sharing a given segment's bandwidth. chapter 9

32. CSMA/CD continue • CSMA/CD is a probabilistic access control method (the opportunity to transmit is not guaranteed). • The possibility that a device might not gain access to the network at a critical time is unacceptable. • This is one reason why IBM invented the Token-Ring network. • Token passing is the media access technique used in token ring LANs: • A token ring operates as a logical ring. • The transmitter of each device is connected to the receiver of the next device in the ring enabling devices to pass messages around the ring. chapter 9

33. Figure 9.8: Token Passing chapter 9

34. Token Ring continued • Token: special type of data frame that circulates around the ring. • A device can transmit only when it is in possession of the token • After a data frame is transmitted, the device releases the token to the network so that other devices can transmit. • Apparently simple method but: • How is the loss of a token detected and how is a new token created? • What steps should a station take if it stops receiving data? • What if a station that was to receive a frame goes off the network? • How is the network to identify frames that have circulated the network too many times? chapter 9

35. Token Ring continued • Ring error monitor (REM) can regenerate lost tokens and remove bad frames from the network. • Additionally, each device is capable of signaling certain problems by transmitting a beacon signal. • Error detection and diagnostic tools available on a token ring are quite extensive. By contrast, no such tools are built into Ethernet (CSMA/CD) networks. • The mechanism that controls a token ring network is much more involved and expensive than that required for Ethernet. • Token ring is deterministic (every device is guaranteed a chance to transmit each time the token circulates the ring ) while Ethernet is probabilistic. chapter 9

36. Token Ring vs. CSMA/CD • Token passing may suffer less performance degradation than CSMA/CD in very large LANs. • Because contention for the transmission medium is more orderly than with CSMA/CD, eliminating collisions, timeouts, and subsequent retries. • Token passing allows stations to transmit whenever a free token is available, but they may have to wait a while to get a token. • A potential problem is that stations that get a token can "hog" the LAN. • Implementations attempt to minimize this by placing a limit on the amount of time a single station can send before passing on the token. chapter 9

37. LAN Topologies • Topology of the LAN: actual physical layout of a LAN or the arrangement in which devices are interconnected. • Most widely used topologies: • bus • star • ring chapter 9

38. Bus Topology • All stations (systems and devices) directly connected to the same transmission medium, usually a physical cable. • simple, and as a result, often inexpensive. • very common on LANs • type of topology originally specified for Ethernet LANs. • Information on bus-based LANs is broadcast to all connected stations. • Transmissions go in both directions along the bus or cable. • All stations, thus, receive all transmissions. • Each station has a unique address assigned to it. • Station address included in the data frames that carry information on the bus. chapter 9

39. Figure 9.9: LAN BUS topology chapter 9

40. Star Topology • Each station connected to a central piece of equipment commonly called a hub. • All communications go through the hub which amplifies and retransmits signals providing connectivity among stations • Individual stations are not directly connected to one another but indirectly through the central hub. • One problem with a star is that if the central hub fails, the network is inoperable. • Redundancy features are often built into hubs making them very reliable. Also, hubs can be configured so that a bypass is possible in the event that a component fails. chapter 9

41. Figure 9.10: LAN STAR topology chapter 9

42. Ring Topology • Stations directly connected to other stations form what looks like a ring. • Unlike a star, adjacent stations on the ring are directly cabled to one another. • No central hub. Information flows in one direction around a ring. Each station receives all signals from the adjacent station, regenerates and retransmits frames to the next station on the ring. • Eliminates problem of depending on a central switch but dependent on each individual station on the ring. • If a station fails, or the link between stations fails, the ring can become inoperable • There are solutions to this problem such as redundant links and ways to bypass failed stations. chapter 9

43. Ring Topology Continued • As with other LAN topologies, each station on a ring has a unique address. • A station look at the destination address in a frame to determine if a frame is intended for it. • If so, the frame is pulled off the ring. • If not, the frame is retransmitted to the next station. • Token passing schemes are often used on ring-base LANs. • For example, IBM's token-Ring LAN uses a ring topology on which a token passing media access control method is used. chapter 9

44. Figure 9.11: LAN RING topology chapter 9

45. Ethernet • Jointly by Xerox, DEC, and Intel • One of the most popular types of LAN in use today. • Original Ethernet specification called for coaxial cable as the transmission medium. • Today, Ethernet LANs make use of other types of cabling such as twisted-pair wire as well. • Ethernet uses a bus topology • The main advantage of Ethernet is its relatively low cost. • Inexpensive Ethernet adapters are available for most PCs and Ethernet interfaces are supported in a wide range of computer equipment. • Multivendor support of Ethernet makes it a popular choice in many cases chapter 9

46. Figure 9.12: Ethernet uses LAN BUS topology chapter 9

47. Ethernet Continued • Ethernet LAN uses a CSMA/CD contention protocol (defined as part of IEEE 802.3 standard). • Ethernet and IEEE 802.3 specifications are similar, but not identical. • One difference: Ethernet frame headers include a type field that indicates the higher layer protocol in use. • For example, the type field could indicate that either TCP/IP or XNS (Xerox Network System) protocols were used across the Ethernet LAN. • Instead of a type field, the header of an 802.3 frame includes a length field indicating the length of the data contained in the information portion of the frame. chapter 9

48. IEEE Networking Specifications TABLE IEEE 802 standards

49. Figure 9.13: Ethernet frame compared to IEEE 802.3 frame chapter 9

50. Ethernet Continued • Originally designed for transmission rates of 10 Mbps • Work going on for Ethernet-like LANs at higher speeds (fast Ethernet at 100 Mbps). • Ethernet LANs with large number of users and heavy traffic demands may result in performance problems • Primarily due to the characteristics of the CSMA/CD (collisions, forcing stations to wait before retrying their transmission). • An upper limit of 10 Mbps also becomes a problem: • when transmitting information requiring high bandwidth such as video images. • It also makes Ethernet less attractive as a backbone LAN to interconnect other LANs. • But, multiple Ethernet LANs can be easily interconnected by devices called bridges forming large logical LANs. chapter 9