Post on 28-Jan-2015
description
1
CS 56 • Networking Essentials
Chapter 3 Understanding Network Architecture
Sonny Huang
2
Chapter 3 Understanding Network Architecture
Outline
Access Methods
How Networks Send Data
Ethernet
3
Access Methods
Access Methods1. Function of Access MethodsA. Overview
1. Rules that define how a computer puts data on and takes it from the network cable is known as access method.2. Access methods regulate the flow of network traffic.
B. Traffic Control on the Cable1. Access the cable without running into other data.
4
Access Methods
2. Receiving computer must access cable with reasonable assurance that data has not been destroyed in a data collision during transmission.
3. Prevents simultaneous access to the cable.Collision occurs if two computers put data on the cable at the same time
5
Access Methods
Video Note c03dem01, c03dem02, and d03dem03 videos show how access methods help ensure orderly transmission of data on a network.
2. Major Access MethodsA. Three methods to prevent simultaneous use of the
cable:1. Carrier-sense multiple access methods (with collision detection and with collision avoidance)2. Token-passing method3. Demand-priority method
6
Access Methods
B. Carrier-Sense Multiple Access with Collision Detection (CSMA/CD) Access Method
1. Each computer on the network, including clients and servers, checks the cable for network traffic.a. Only when there is no traffic, can data be sent.b. After data is sent, no other computer can transmit until original data reaches its destination.c. A data collision occurs when two or more computers send data at the same time.d. When this happens, each computer stops data transmission and waits to resend when it “senses” the cable is free.
7
Access Methods
Computers can transmit data only if the cable is free.
8
Access Methods
2. Contention Method.a. Computers on the network contend, or compete, for an opportunity to send data.b. If there is data on the cable, no other computer may transmit until the data reaches its destination and cable is free.c. Current implementations of CSMA/CD are so fast that users are not even aware they are using a contention access method.
Video Note c03dem04 and c03dem05 videos show the CSMA/CD access method.
9
Access Methods
Video Note c03dem06 video show why CSMA/CD is considered a contention method.
3. CSMA/CD Considerations.
a. More computers means more network traffic.b. Can be a slow access method depending on the number of users and the applications being used.
10
Access Methods
C. Carrier-Sense Multiple Access with Collision Avoidance (CSMA/CA) Access Method1. Least popular of the three major access methods.2. Each computer signals its intent to transmit before it actually transmits data to avoid transmission collisions.3. Broadcasting the intent to transmit data increases the amount of traffic on the cable and slows down network performance.
Video Note c03dem07 video shows the CSMA/CA access method.
11
Access Methods
D. Token-Passing Access Method1. A special type of packet, called a token, circulates
around a cable ring from computer to computer.2. Computer wanting to send data must wait for the free
token.3. When computer obtains free token, it can transmit its
data.4. No other computer can transmit while this computer is
transmitting.5. No contention and no collisions take place.6. No time is spent waiting to resend data due to traffic.
12
Access Methods
Token Passing access method
13
Access Methods
Receiving computer reads the data and sets a flag that says the data has been received. It then puts the packet back on the cable to the sending computer. The sending computer creates a new free token after it reads the flag that says the receiving computer has processed the data.
Video Note c03dem08, c03dem09, c03dem10, c03dem11, and c03dem12 videos show the token-passing access method.
14
Access Methods
E. Demand Priority Access Method1. Overview
a. Designed for 100 Mbps Ethernet, known as 100VG-AnyLAN.b. Sanctioned and standardized by the Institute of Electrical and Electronic Engineers (IEEE) in its 802.12 specification.c. Repeaters and end nodes are the two components that make up all 100VG-AnyLANs.d. Repeaters are responsible for searching for requests to send.e. Repeaters are responsible for noting all addresses, links, and end nodes, and verifying that they are all functioning.
15
Start-bus network access method for100VG-AnyLAN is demand priority
Access Methods
f. End node can be a computer, bridge, router, or switch.
16
Access Methods
2. Demand Priority Contentiona. Computers contend (transmit at the same time).b. Hub determines priority (could be based on data type).c. Higher priority is transmitted first; if same, then they are alternated.
3. Demand Priority Considerationsa. Uses four pairs of wires, allowing computers to transmit and receive at the same time.b. Transmissions through the hub.c. No network-wide broadcasts, only computer-to-hub-to-computer.
17
Access Methods
See major points of each of the access methods in the access methods summary on page 115.
Video Note c03dem13 and c03dem14 videos show the demand-priority access method.
18
How Networks Send Data
How Networks Send Data1. The Function of Packets in Network CommunicationsA. Data sent in small, manageable packets, each wrapped
with the essential information needed to get it from its source to the correct destination.1. Large files flood and tie up the network.
Large continuous streams of data slow down the network
19
Breaking data into packets
How Networks Send Data
2. Impact of retransmitting large units of data multiply network traffic.3. Small packets provide timely error recovery, and the network only has to resend small amount of data.
Although the terms “packet” and “frame” are often used interchangeably, there are some differences based on the type of network. Packet is a unit of information transmitted as a whole from one device to another.
20
How Networks Send Data
B. Sender breaks data into packets and adds special control
information to each packet making it possible to:1. Send original, disassembled data in small chunks.2. Reassemble data in proper order at destination.3. Check data for errors after it has been reassembled.
2. Packet StructureA. Packets contain several types of data, including:
1. Information, such as messages or files.2. Computer control data and commands, such as service requests.
21
How Networks Send Data
3. Session control codes (such as error correction) that indicate the need for retransmission.
B. Packet Components1. Common packet components:
a. Source address that identifies sending computerb. Data intended for transmissionc. Destination address that identifies the recipientd. Instructions telling network components how to pass datae. Information to receiving computer about how to reassemble the dataf. Error-checking information to ensure that packet arrives intact
22
How Networks Send Data
2. Headera. Alert signal indicating
packet is being transmitted
b. Source addressc. Destination addressd. Clock information to
synchronize transmission
3. Dataa. Actual data being sent: 0.5 KB to 4 KB in sizeb. Larger files broken into smaller multiple packets
23
How Networks Send Data
4. Trailera. Exact content of the trailer varies depending on the communication method, or protocol.
A protocol is a set of rules or standards designed to enable computers to connect with one another and to exchange information with as little error as possible.
b. Contains error-check component called cyclical redundancy check (CRC), a number produced by mathematical computation at source.c If CRC number is calculated to be the same when it arrives, the data is assumed good; if not, it is retransmitted.
24
How Networks Send Data
C. Example: Packets in Printing1. Sending computer establishes a connection with the
print server.
Establishing a connection with a print server
25
Creating Packets
How Networks Send Data
2. Computer breaks the large print job into packets containing:a. Source and destination address.b. Data.c. Control
information.
26
How Networks Send Data
3. NIC in each computer examines the receiver’s address on all packets sent on its segment of the network.a. NIC compares the incoming packets destination address with its own address.b. NIC interrupts the computer only when it detects a packet addressed specifically to itself.
Examining the receiver’s address
27
How Networks Send Data
4. Packets enter the destination computer (print server) through the cable into the NIC.
Network interface card(NIC) accepts packets addressed to the printer server
28
How Networks Send Data
5. No computer resources are used until the NIC identifies a packet addressed to itself.
6. Network Operating System in the receiving computer reassembles the packets back into the original text file and moves the file into the computer’s memory.
Reassembled packets sent to the printer
29
Ethernet
Ethernet1. Origin of EthernetA. Late 1960’s, the University of Hawaii developed a
WAN called ALOHA using the CSMA/CD access method.
B. In 1972, a cabling and signaling scheme was invented at Xerox Palo Alto Research Center (PARC).
C. In 1975, the first Ethernet product was introduced; a system of 2.94 megabits per second (Mbps) to connect over 100 computers connected on one kilometer (.62 miles) of cable.
D. Xerox, Intel Corporation and Digital Equipment Corporation drew up standard for 10-Mbps Ethernet.
30
Ethernet
2. Ethernet SpecificationsA. International Organization for Standardization (ISO)
1. In 1978, ISO released a set of specifications for connecting dissimilar devices.2. This set of specifications is referred to as the Open System Interconnect (OSI) model.
The OSI model is described in detail in Chapter 5.
B. Ethernet specification performs the same functions as the OSI
physical and data-link layers of this model.1. These specifications affect how hardware links or passes information back and forth.2. Project 802 published by IEEE in 1980s.3. The IEEE 802.3 specification.
31
Ethernet
Ethernet became the basis for IEEE 802.3.
3. Ethernet FeaturesA. Most Popular Network Architecture
1. Ethernet media is passive, which means it requires no power source.2. Will not fail unless the media is physically cut or improperly terminated.
Simple Ethernet bus network terminated at both end
32
Ethernet
B. Ethernet Basics1. Topology: Linear bus or star bus2. Architecture: Baseband (digital: bidirectional)3. Access method: CSMA/CD4. Specifications: IEEE 802.35. Transfer speed: 10 or 100 Mbps6. Cable type: ThickNet, ThinNet, UTP
The most popular physical topology is a star, the manner a signal is transferred is still a bus.
33
Ethernet
C. The Ethernet Frame Format1. Ethernet breaks data into frames.
a. Package of information transmitted in a single unit (between 64 and 1518 bytes long).b. Ethernet uses 18 bytes for itself; therefore, data is between 46 and 1500 bytes long.
“Packet” and “frame” can be used interchangeably; in the context of Ethernet, the term “frame” is used.
34
Ethernet
2. Parts of the frame:a. Preamble, the start of the frameb. Destination and source addressc. Type of frame, usually IP or IPX (Novell’s Internetwork Packet Exchange)d. Error checking field (Cyclical redundancy check (CRC)
Specifics of frame layout vary from one type to another.
Sample Ethernet II frame
35
Ethernet
4. 10-Mbps IEEE StandardsA. 10BaseT Standard
See 10BaseT specifications summary in Table 3.4 on page 130.
1. 10-Mbps, baseband, over twisted-pair cable.2. Usually unshielded twisted-pair, but STP will work.
UTP is by far the most popular type of cable. STP is not recommended.
36
Ethernet
3. Uses hubs (multiport repeaters) from wiring closets.
4. Configured as star, but internally it is a bus.
A multiport Repeater can be used to extend an Ethernet LAN
37
A patch panel makes moving computer easy
Ethernet
5. Each computer has two pairs of wire: one pair receives data and one pair transmits.6. Cable type is Category 3, 4, or 5 UTP. Maximum length of 100 meters (328 feet); minimum of 2.5 meters (8 feet).7. 1024 computers maximum.
38
Ethernet
B. 10Base2 Standard
See 10Base2 specifications summary in Table 3.5 on page 132.
1. 10 Mbps, baseband, over thin coaxial cable.2. ThinNet with BNC connectors and terminators.3. BNC barrels to extend sections.
Be careful to limit the number of BNC barrels to maintain signal strength.
39
Ethernet
4. Maximum length of 185 meters (607 feet); minimum of 0.5 meters (20 inches).5. ThinNet can support maximum of 30 nodes per populated segment.6. Repeaters used to extend to total of 925 meters (3035 feet).
Repeaters count in the total number of 30 nodes per populated segment.
7. Relatively inexpensive, easy to install, easy to configure.
40
The thinnet 5-4-3 rules: 5 segments, 4 repeaters, and 3 populated segments
Ethernet
8. Uses 5-4-3 rule. The 5-4-3 rule can combine as many as five cable
segments connected by four repeaters; but only three segments can have stations attached. Thus, two segments are untapped and are often referred to as “inter-repeater links.”
41
Ethernet
C. 10Base5 Standard
See 10Base5 specifications summary in Table 3.6 on page 135.
1. 10 Mbps, baseband, over 500-meter (five 100-meter) segments.2. Called ThickNet or standard Ethernet.3. ThickNet uses bus topology and can support up to one hundred nodes per populated segment.4. ThickNet segment can be 500 meters (1640 feet) long for a total network length of 2500 meters (8200 feet).5. Transceivers are devices that transmit and receive, and also provide communications between the computer and the main LAN cable.
Thicknet cable composition
42
Ethernet
6. Transceiver cable uses a DIX or AUI 15-port to connect the transceiver to the NIC.
“AUI,” an acronym for attachment unit interface, is a 15-pin (DB-15) connector commonly used to connect a NIC to an Ethernet cable; AUIs and DIXs are discussed in Chapter 2.
43
Ethernet
7. N-series connectors, including N-series barrel connectors, and N-series terminators.
Thicknet backbone with attached transceiver and cable
44
Ethernet
8. Uses 5-4-3 rule.
The thicknet 5-4-3 rules: 5 backbone segments , 4 repeaters, and 3 populated segments
45
Ethernet
9. Maximum length of 50 meters (164 feet); minimum of 2.5 meters (about 8 feet).
D. Combining ThickNet and ThinNet Cable1. ThickNet cable is good for backbones.2. ThinNet cable is used for branch segments.3. Branch segments connect the computer to the backbone.
46
Ethernet
E. 10BaseFL Standard1. Ethernet using fiber-optic cable to connect computers and repeaters.2. Accommodates long runs between buildings.3. Maximum segment length of 2000 meters (about 6800 feet).
Speed is the same as standard Ethernet, but longer distances are possible due to fiber-optic cable.
47
Ethernet
5. 100-Mbps IEEE StandardsA. Able to support high-bandwidth applications:
1. Computer-aided design (CAD)2. Computer-aided manufacturing (CAM)3. Video4. Imaging and document storage
B. 100BaseVG AnyLAN Standard1. 100BaseVG AnyLAN Ethernet is IEEE 802.12 standard.
a. Combines Ethernet and Token Ring technology
b. Originally developed by Hewlett Packard
48
Ethernet
2. Specifications:a. Minimum data rate of 100 Mbpsb. Cascading star topology (parent/child) over Category 3, 4, and 5 twisted-pair and fiber-optic cablec. Demand priority access methodd. Ability to filter individually addressed frames at the hube. Support Ethernet frames and Token Ring
packets
100BaseVG AnyLAN is also referred to as 100BaseVG, VG, and AnyLAN.
49
Ethernet
3. Topology.a. Star topology in which all computers are
attached to a hub.b. Add child hubs to the central hub to expand
the network.
Parent hub controls transmission to computers attached by their child hubs.
Parents hubs with five attached child hubs
50
Ethernet
4. Limitations of 100BaseVG:a. Requires own hubs and cardsb. Two longest cables from hub to
computer less than or equal to 250 meters (810 feet)
c. Requires more wiring closets than 10BaseT
51
Ethernet
C. 100BaseX Ethernet Standard
See 100BaseX Media Specification listed in Table 3.7 on page 138.
1. Fast Ethernet2. Uses UTP Category 5 wire3. Star bus topology like 10BaseT4. Incorporates three media specifications:
a. 100BaseT4 (4-pair Category 3, 4, 5 UTP)b. 100BaseTX (2-pair Category 5 UTP or STP)c. 100BaseFX (2-strand fiber optic)
52
Ethernet
6. Performance ConsiderationsA. Ethernet architecture can use multiple communication
protocols:1. NetWare2. UNIX3. Windows4. Macintosh
|B. Segmentation1. Divides crowded segment into two less crowded segments using bridge or router.2. Reduces traffic and improves access time.
53
Ethernet
Router actually breaks the single segment into two; the bridge just creates two collision domains. Bridges and routers are discussed in Chapter 7.
Using a bridge to segment a network and reduce network traffic.
54
Ethernet
C. NOSs on Ethernet1. Microsoft Windows 95, Windows 98, and Windows 20002. Microsoft Windows NT Workstation and Windows NT Server3. Microsoft Windows 2000 Professional and Windows 2000 Server4. Microsoft LAN Manager5. Microsoft Windows for Workgroups6. Novell NetWare7. IBM LAN Server8. AppleShare9. UNIX
55
Ethernet
Review the table summarizing the specifications for Ethernet architecture on page 140.
56
Token Ring
Token Ring1. OverviewA. General Information
1. Token Ring: 802.5 specifications2. Introduced by IBM in 1984 for all its computer environments, from PCs to mainframes (SNA)
IBM Token Ring is one implementation of IEEE 802.5, which is covered in more detail in Chapter 5.
3. Adopted by ANSI/IEEE in 1985
57
Token Ring
American National Standards Institute (ANSI) is an organization that was formed in 1918 for the development and adoption of trade and communication standards in the United States. ANSI is the American equivalent of the International Organization for Standardization (ISO).
B. Token Ring Features1. Architecture
a. Logical ring is within the hub.b. Computers are wired in a star from the hub (star-wired ring topology).c. Token still passes in a ring from computer to computer.
58
Token Ring
Logical electrical mode of transmission. Hubs may be powered or not. All stations on a line must be same speed.
Logical ring, in which the physical ring is in the hub.
59
Token Ring
2. Token Ring Basicsa. Star-wired ring topologyb. Token passing access methodc. IBM Type 1, 2, 3 cabling (STP or UTP)d. Transfer rates of 4 Mbps or 16 Mbpse. Baseband transmissionf. IEEE 802.5 specifications
60
Token Ring
See Table 3.9, Components of a Token Ring Frame, on page 143.
C. Frame Formats
1. Start delimiter indicates the start of the frame.2. Access Control and Frame Control.
Instructor Note See Figure 5.8, Media Access Control, described in detail on page 223.
3. Source and Destination Addresses.4. Information or Data.
61
Token Ring
5. CRC Error Checking.6. End delimiter indicates the end of frame.7. Frame status tells whether the frame was recognized, copied, or whether the destination address was available.
2. How Token Ring Network Works
Token is a predetermined formation of bits that permits a computer to put data on the cable.
62
Token Ring
A. Overview1. First computer on the network generates a token.2. Token travels around the ring polling each computer.3. Computer signals that it wants to transmit data and takes control of the token.4. Computer cannot transmit unless it has possession of the token.5. No computer can transmit data while the token is in use.
63
Token Ring
Token passing is considered deterministic, which means that a computer cannot force its way on the network as in CSMA/CD.
|30|6. Each computer acts as a unidirectional repeater, regenerates the token, and passes it along.
7. After taking control of the token, it sends a data frame out on the network.8. Frame proceeds around the ring until it reaches the destination computer.9. Destination computer copies the frame into its receive buffer.10. Destination computer marks the status field as received.11. Frame continues around the ring until sending computer receives it.12. Sending computer removes the frame and generates a new token.13. Only one token at a time can be active on the network.14. Token travels in only one direction.
Instructor Note Does the token flow clockwise or counterclockwise? The answer is that it doesn’t matter. The direction taken depends on hardware connections. Logically, one can make the token travel in any direction or order. The designers of hubs determine the order in which each port is addressed, and the administrator determines the order in which computers are connected to the hub. The IEEE 802.5 standard says clockwise; the IBM publication SC30-3374, section 3, says counterclockwise.
B. Monitoring the System1. First computer online is assigned to monitor network activity by the Token Ring system and is responsible for:a. Frames being delivered and received correctly.b. Frames that have circulated the ring more than once.c. Ensuring only one token is on the network.2. Process of monitoring is called beaconing.a. Active monitor sends out a beacon announcement every seven seconds.b. Beacon passed from computer to computer throughout the ring.c. If station does not receive an expected announcement from its upstream neighbor it sends a message including:(1) Stations address(2) Address of neighbor that did not announce(3) Type of beacond. Ring attempts to diagnose and repair without disrupting the entire network.
Instructor Note When troubleshooting, be aware the network is trying to bypass the problem. This can sometimes mask the real failure.
e. If unable to reconfigure automatically, manual intervention is required.3. Recognizing a Computer.a. When a new computer comes online the initialization includes:(1) Checking for duplicate addresses.(2) Notifying others of its existence.3. Hardware Components
|31|A. The Hub1. Overviewa. Token ring hub is known by several names:(1) MAU (Multistation Access Unit)(2) MSAU (MultiStation Access Unit)(3) SMAU (Smart Multistation Access Unit)b. The hub’s internal ring automatically converts to an external ring at each connection point when a computer is connected.2. Capacitya. IBM MSAU has 10 connection ports creating a logical ring of eight stations (two other ports extend the ring).
Instructor Note Some MSAU units are passive. For example, the IBM 8228.
b. Up to 33 hubs on a ring supporting 72 computers that use UTP or 260 that use STP on a MSAU-based network.
Instructor Note Other vendors offer hubs with more capacity; capacity depends on the vendor and the hub model.
|32|c. Hubs must be connected in a ring; ring-out on one hub must connect to ring-in on the adjacent hub.
d. Patch cables connect many MSAUs on top of each other while still forming a continuous ring.3. Built-in Fault Tolerancea. NIC failure should bring down ring.b. MSAU will sense a failed NIC and disconnect from it.c. Faulty computer or connection will not affect rest of network.B. Cabling1. Overviewa. IBM Type 1, 2, 3; usually Type 3 UTP cabling.
|33|b. IBM states maximum cabling distance MSAU to computer is 46 meters (150 feet), although some vendors say 152 meters (500 feet).c. Maximum distance from one MSAU to another is 152 meters (500 feet).d. Single Token Ring can accommodate only 260 computers with STP cable and 72 computers with UTP cable.2. Patch Cablesa. Extend connection between computer and MSAU or between two MSAUs.b. IBM Type 6, up to a maximum of 46 meters (150 feet).3. Connectorsa. Media interface connectors (MICs), also known as universal data connectors, connect Types 1 and 2 cable.b. Neither male nor female so they can be connected together.c. RJ-45 for 8-pin connection for Type 3 cable.d. RJ-11 for 4-pin connection for Type 3 cable.e. Media filter required to make connection between Token Ring interface card and standard RJ-11/RJ-45 telephone jacks.4. Media Filtersa. Required with Type 3 telephone twisted-pair cable.b. Convert cable connectors and reduce line noise.5. Patch Panelsa. Organize cable that runs between a MSAU and a telephone punchdown block.b. Punchdown block is a kind of hardware that provides terminal connections for bare network cable ends.6. Repeatersa. Increase Token Ring cable distances.b. Actively regenerates and retimes the Token Ring signal.7. NICsa. Either 4 Mbps or 16 Mbps.b. If the network is a 4 Mbps network, the 16 Mbps cards can be used, but will revert to 4 Mbps mode (16 Mbps networks will not accept slower 4 Mbps cards because they cannot speed up).
Instructor Note Different speed cards can exist on different logical segments.
8. Fiber-Optic Cablea. Well suited for unidirectional, high-speed fiber-optic cable.b. Although more expensive, it greatly increases the range (up to 10 times that of copper).4. Future of Token Ring Networks
Instructor Note See Chapter 6 for discussion of protocols, routers, and bridges.
A. Overview1. Token ring is losing market share to Ethernet.2. Large companies are selecting Token Ring to support mission-critical applications.B. Users of token ring face the following challenges:1. Complexity, manageability, cost and space requirements2. Bridge and segment congestion3. Upgrading to high-speed technologies
Instructor Note See Token Ring specifications in Table 3.10 on page 151.
Clockwise flow of the token around the logical ring
64
Token Ring
6. Each computer acts as a unidirectional repeater, regenerates the token, and passes it along.7. After taking control of the token, it sends a data frame out on the network.8. Frame proceeds around the ring until it reaches the destination computer.9. Destination computer copies the frame into its receive buffer.10. Destination computer marks the status field as received.11. Frame continues around the ring until sending computer receives it.12. Sending computer removes the frame and generates a new token.
65
Token Ring
13. Only one token at a time can be active on the network.14. Token travels in only one direction.
Does the token flow clockwise or counterclockwise? The answer is that it doesn’t matter. The direction taken depends on hardware connections. Logically, one can make the token travel in any direction or order. The designers of hubs determine the order in which each port is addressed, and the administrator determines the order in which computers are connected to the hub. The IEEE 802.5 standard says clockwise; the IBM publication SC30-3374, section 3, says counterclockwise.
66
Token Ring
B. Monitoring the System1. First computer online is assigned to monitor network
activity by the Token Ring system and is responsible for:a. Frames being delivered and received correctly.b. Frames that have circulated the ring more than once.c. Ensuring only one token is on the network.
2. Process of monitoring is called beaconing.a. Active monitor sends out a beacon announcement every seven seconds.b. Beacon passed from computer to computer throughout the ring.
67
Token Ring
c. If station does not receive an expected announcement from its upstream neighbor it sends a message including:
(1) Stations address(2) Address of neighbor that did not
announce(3) Type of beacon
d. Ring attempts to diagnose and repair without disrupting the entire network.
When troubleshooting, be aware the network is trying to bypass the problem. This can sometimes mask the real failure.
68
Token Ring
e. If unable to reconfigure automatically, manual intervention is required.
3. Recognizing a Computer.a. When a new computer comes online the initialization includes:
(1) Checking for duplicate addresses.(2) Notifying others of its existence.
69
Hub showing the internal ring and clockwise token path
Token Ring
3. Hardware ComponentsA. The Hub1. Overview
a. Token ring hub is known by several names:(1) MAU (Multistation Access Unit)(2) MSAU (MultiStation Access Unit)(3) SMAU (Smart Multistation Access Unit)
b. The hub’s internal ring automatically converts to an external ring at each connection point when a computer is connected.
70
Token Ring
2. Capacitya. IBM MSAU has 10 connection ports creating a logical ring of eight stations (two other ports extend the ring).
Some MSAU units are passive. For example, the IBM 8228.
b. Up to 33 hubs on a ring supporting 72 computers that use UTP or 260 that use STP on a MSAU-based network.
Other vendors offer hubs with more capacity; capacity depends on the vendor and the hub model.
71
Token Ring
c. Hubs must be connected in a ring; ring-out on one hub must connect to ring-in on the adjacent hub.
d. Patch cables connect many MSAUs on top of each other while still forming a continuous ring.
72
Token Ring
3. Built-in Fault Tolerancea. NIC failure should bring down ring.b. MSAU will sense a failed NIC and disconnect from it.c. Faulty computer or connection will not affect rest of network.
73
Token Ring
B. Cabling1. Overview
a. IBM Type 1, 2, 3; usually Type 3 UTP cabling.b. IBM states maximum cabling distance MSAU to computer is 46 meters (150 feet), although some vendors say 152 meters (500 feet).c. Maximum distance from one MSAU to another is 152 meters (500 feet).d. Single Token Ring can accommodate only 260 computers with STP cable and 72 computers with UTP cable.
74
Token Ring
2. Patch Cablesa. Extend connection between computer and MSAU or between two MSAUs.b. IBM Type 6, up to a maximum of 46 meters (150 feet).
3. Connectorsa. Media interface connectors (MICs), also known as universal data connectors, connect Types 1 and 2 cable.b. Neither male nor female so they can be connected together.c. RJ-45 for 8-pin connection for Type 3 cable.d. RJ-11 for 4-pin connection for Type 3 cable.
75
Token Ring
e. Media filter required to make connection between Token Ring interface card and standard RJ-11/RJ-45 telephone jacks.
4. Media Filtersa. Required with Type 3 telephone twisted-pair cable.b. Convert cable connectors and reduce line noise.
5. Patch Panelsa. Organize cable that runs between a MSAU and a telephone punchdown block.b. Punchdown block is a kind of hardware that provides terminal connections for bare network cable ends.
76
Token Ring
6. Repeatersa. Increase Token Ring cable distances.b. Actively regenerates and retimes the Token Ring signal.
7. NICsa. Either 4 Mbps or 16 Mbps.b. If the network is a 4 Mbps network, the 16 Mbps cards can be used, but will revert to 4 Mbps mode (16 Mbps networks will not accept slower 4 Mbps cards because they cannot speed up).
Different speed cards can exist on different logical segments.
77
Token Ring
8. Fiber-Optic Cablea. Well suited for unidirectional, high-speed fiber-optic cable.b. Although more expensive, it greatly increases the range (up to 10 times that of copper).
78
Token Ring
4. Future of Token Ring Networks
See Chapter 6 for discussion of protocols, routers, and bridges.
A. Overview
1. Token ring is losing market share to Ethernet.2. Large companies are selecting Token Ring to support mission-critical applications.
B. Users of token ring face the following challenges:1. Complexity, manageability, cost and space requirements2. Bridge and segment congestion3. Upgrading to high-speed technologies
79
Token Ring
See Token Ring specifications in Table 3.10 on page 151.
80
AppleTalk Network
AppleTalk EnvironmentA. AppleTalk1. Overview
a. Apple Computer introduced AppleTalk in 1983 as proprietary network architecture for small groups.b. Built into Macintosh computers.c. AppleTalk Phase 2 is the current release of AppleTalk.
2. When a device attached to a LocalTalk network comes online, the following occurs in order:a. Device checks to see if it has stored an address from a previous networking session. If not, the new device assigns itself a random address from a range of allowable addresses.
81
AppleTalk Network
b. The device broadcasts the address to see if any other device is using it.c. If no other device is using the address, the device stores address for the next time it comes online.
AppleTalk network
82
AppleTalk Network
B. LocalTalk1. Overview
a. CSMA/CA access method in STP bus or tree topology.b. Cabling components include cables, connector modules, and cable extenders.c. Supports a maximum of 32 devices.d. Other vendors offer more robust solutions.
LocalTalk connector module with a LocalTalk cable
83
AppleTalk Network
Can also accept UTP or fiber-optic cable. 2. Not widely used by Ethernet or Token Ring because:
a. Maximum communication data rate is 230.4 Kbps.b. LocalTalk NICs for PC-compatible computers are obsolete.
C. AppleShare1. AppleShare is the file server on AppleTalk network.2. LocalTalk networks can be joined using zones.3. Each zone is a workgroup and is given a name.
84
AppleTalk Network
Workgroups are divided into zones to ease congestion.
Three zones joined together to form a larger network.
85
AppleTalk Network
D. EtherTalk1. EtherTalk allows AppleTalk to use Ethernet coaxial
cable (ThinNet or ThickNet).2. EtherTalk card allows a Macintosh computer to connect
and 802.3 Ethernet network.
E. TokenTalk1. TokenTalk connects Macintosh II to an 802.5 Token
Ring.2. Compatible with AppleTalk Phase 2.
F. AppleTalk considerations1. PCs and mainframes can use AppleTalk.2. Apple encourages third-party use of network.
86
AppleTalk Network
2. ArcNet EnvironmentA. Overview
1. Datapoint Corporation developed the Attached Resource Computer Network (ArcNet) in 1977.
ArcNet predates IEEE Project 802 standards, but loosely maps to the 802.4 standard.
Simple Start-wired ArcNet network
87
AppleTalk Network
2. Simple, inexpensive, flexible network architecture designed for workgroup-sized networks.3. Token-passing bus networks using broadband cable.4. Star bus or bus topology.
ArcNet is currently more expensive than Ethernet.
B. How ArcNet works:
1. Token-passing access method in a star bus.2. 2.5 Mbps, although ArcNet Plus is 20 Mbps.3. Token moves in numerical order of card address, regardless of where it is on the network.
88
AppleTalk Network
| 4. ArcNet packet includes source and destination addresses, and up to 508 bytes of data.C. Hardware1. Hubs can be active, passive, or smart.
Instructor Note Smart hubs have all the features of active hubs and usually add diagnostic features such as reconfiguration detection and operator control-of-port connections.
2. Uses 93-ohm RG-62 A/U coaxial cable.
Instructor Note Unlike active hubs, unused ports on passive hubs must be terminated; they can also use mixture of hubs and bus.
3. Maximum cable length between active hub and workstation is 610 meters (2000 feet).4. Maximum distance between drops on a linear coaxial bus is 305 meters (1000 feet).5. Using UTP, maximum cable distance is 244 meters (800 feet) between devices on both star and bus topologies.
Instructor Note Discuss ArcNet Specifications, Table 3.11 on page 159.
89
AppleTalk Network
| 4. ArcNet packet includes source and destination addresses, and up to 508 bytes of data.
Token movement based on numerical order
An ArcNet packet contains source and destination addresses
90
AppleTalk Network
C. Hardware1. Hubs can be active, passive, or smart.
Smart hubs have all the features of active hubs and usually add diagnostic features such as reconfiguration detection and operator control-of-port connections.
2. Uses 93-ohm RG-62 A/U coaxial cable.
Unlike active hubs, unused ports on passive hubs must be terminated; they can also use mixture of hubs and bus.
91
AppleTalk Network
3. Maximum cable length between active hub and workstation is 610 meters (2000 feet).4. Maximum distance between drops on a linear coaxial bus is 305 meters (1000 feet).5. Using UTP, maximum cable distance is 244 meters (800 feet) between devices on both star and bus topologies.
Discuss ArcNet Specifications, Table 3.11 on page 159.
92
AppleTalk Network
D. EtherTalk1. EtherTalk allows AppleTalk to use Ethernet coaxial cable (ThinNet or ThickNet).2. EtherTalk card allows a Macintosh computer to connect and 802.3 Ethernet network.E. TokenTalk1. TokenTalk connects Macintosh II to an 802.5 Token Ring.2. Compatible with AppleTalk Phase 2.F. AppleTalk considerations1. PCs and mainframes can use AppleTalk.2. Apple encourages third-party use of network.
|38|2. ArcNet EnvironmentA. Overview1. Datapoint Corporation developed the Attached Resource Computer Network (ArcNet) in 1977.
Instructor Note ArcNet predates IEEE Project 802 standards, but loosely maps to the 802.4 standard.
2. Simple, inexpensive, flexible network architecture designed for workgroup-sized networks.3. Token-passing bus networks using broadband cable.4. Star bus or bus topology.
Instructor Note ArcNet is currently more expensive than Ethernet.
B. How ArcNet works:1. Token-passing access method in a star bus.2. 2.5 Mbps, although ArcNet Plus is 20 Mbps.
|39|3. Token moves in numerical order of card address, regardless of where it is on the network.|40|4. ArcNet packet includes source and destination addresses, and up to 508 bytes of data.
C. Hardware1. Hubs can be active, passive, or smart.
Instructor Note Smart hubs have all the features of active hubs and usually add diagnostic features such as reconfiguration detection and operator control-of-port connections.
2. Uses 93-ohm RG-62 A/U coaxial cable.
Instructor Note Unlike active hubs, unused ports on passive hubs must be terminated; they can also use mixture of hubs and bus.
3. Maximum cable length between active hub and workstation is 610 meters (2000 feet).4. Maximum distance between drops on a linear coaxial bus is 305 meters (1000 feet).5. Using UTP, maximum cable distance is 244 meters (800 feet) between devices on both star and bus topologies.
Instructor Note Discuss ArcNet Specifications, Table 3.11 on page 159.