Guide to Networking Essentials Fifth Edition

1
Guide to Networking EssentialsFifth Edition

• Chapter 7
• Network Architectures

2
Objectives

• Compare and contrast media access methods used in
  network architectures
• Describe the operation of Ethernet
• Differentiate between Ethernet standards and
  speeds
• Explain the four Ethernet frame types and how
  they are used

3
Objectives (continued)

• Describe the token ring architecture and its
  components
• Describe the AppleTalk network architecture
• Explain the function of Fiber Distributed Data
  Interface
• Describe other LAN and WAN architectures and
  their role in todays networks

4
Putting Data on the Cable Access Methods

• Given that network architectures communicate in a
  number of different ways, some factors in network
  communications must be considered
• How computers put data on the cable
• How they ensure that the data reaches its
  destination undamaged

5
Function of Access Methods

• The way in which computers attached to a network
  share the cable must be defined
• A collision results from two or more devices
  sending a signal along the same channel at the
  same time
• Splitting data in small chunks helps prevent
  collisions
• Channel access methods specify when computers can
  access the cable or data channel
• Ensure that data reaches destination by
  preventing computers from sending messages that
  might collide
• Every computer on a network must use the same
  access method

6
Major Access Methods

• Channel access is handled at the MAC sublayer of
  the Data Link layer in the OSI model
• Five major types of channel access
• Contention
• Switching
• Token passing
• Demand priority
• Polling

7
Contention

• In early networks based on contention, computers
  sent data whenever they had data to send
• As networks grow, outgoing messages collide more
  frequently, must be sent again, and then collide
  again
• To organize contention-based networks, two
  carrier access methods were created
• CSMA/CD
• CSMA/CA

8
Carrier Sense Multiple Access with Collision
Detection (CSMA/CD)
9
Carrier Sense Multiple Access with Collision
Avoidance (CSMA/CA)

• When the computer senses that no other computer
  is using the network, it signals its intent to
  transmit
• Other computers with data to send must wait when
  they receive the intent-to-transmit signal and
  send their intent-to-transmit only when channel
  is free
• The overhead created by intent-to-transmit
  packets reduces network speed significantly
• Used in wireless LANs with an access point
• Wireless NIC tells access point its intents to
  transmit
• Access point hears transmissions from all
  devices, so it can determine whether its okay to
  transmit

10
Switching

• Switching nodes are interconnected through a a
  switch, which controls access to the media
• Contention occurs only when multiple senders ask
  to reach the same receiver simultaneously or when
  the simultaneous transmission requests exceed the
  switchs capability to handle multiple
  connections
• Advantages fairer, centralized management
  (enables QoS), switch can have connection ports
  that operate at different speeds
• Disadvantage higher cost

11
Token Passing
12
Demand Priority

• Demand priority channel access method used
  solely by the 100VG-AnyLAN 100 Mbps Ethernet
  standard (IEEE 802.12)
• 100VG-AnyLAN runs on a star bus topology
• Intelligent hubs control access to the network
• Hub searches all connections in a round-robin
  fashion
• When an end node has data to send, it transmits a
  demand signal to the hub
• The hub then sends an acknowledgement that the
  computer can start transmitting its data
• The major disadvantage of demand priority is price

13
Polling
14
Choosing an Access Method
15
Choosing an Access Method (continued)
16
Choosing an Access Method (continued)
17
The Ethernet Architecture

• 1960s and 1970s many organizations worked on
  methods to connect computers and share data
• E.g., the ALOHA network at the University of
  Hawaii
• 1972 Robert Metcalf and David Boggs, from
  Xeroxs PARC, developed an early version of
  Ethernet
• 1975 PARC released first commercial version (3
  Mbps, up to 100 computers, max. 1 km of total
  cable)
• DIX developed standard based on Xeroxs Ethernet
  (10 Mbps)
• 1990 IEEE defined the 802.3 specification
• Defines how Ethernet networks operate at layers
  1-2

18
Overview of Ethernet

• Ethernet is the most popular network architecture
  
• Advantages easy to install, scalable, broad
  media support, and low cost
• Supported transmission speeds 10 Mbps to 10 Gbps
• Uses the NICs MAC address to address frames
• Ethernet variations are compatible with one
  another
• Basic operation and frame formatting is the same
• Cabling, speed of transmission, and method by
  which bits are encoded on the medium differ

19
Ethernet Operation

• Ethernet is a best-effort delivery system
• It works at the Data Link layer of the OSI model
• Relies on the upper-layer protocols to ensure
  reliable delivery of data
• Understanding the following concepts is
  important
• How Ethernet accesses network media
• Collisions and collision domains
• How Ethernet handles errors
• Half-duplex and full-duplex communications

20
Accessing Network Media

• Ethernet uses CSMA/CD in a shared-media
  environment (a logical bus)
• Ethernet device listens for a signal or carrier
  (carrier sense) on the medium first
• If no signal is present, no other device is using
  the medium, so a frame can be sent
• Ethernet devices have circuitry that detects
  collisions and automatically resends the frame
  that was involved in the collision

21
Collisions and Collision Domains
22
Ethernet Error Handling

• Collisions are the only type of error for which
  Ethernet automatically attempts to resend the
  data
• Errors can occur when data is altered in medium
• Usually caused by noise or faulty media
  connections
• When the destination computer receives a frame,
  the CRC is recalculated and compared against the
  CRC value in the FCS
• If values match, the data is assumed to be okay
• If values dont match, the data was corrupted
• Destination computer discards the frame
• No notice is given to the sender

23
Half-Duplex Versus Full-Duplex Communications

• When half-duplex communication is used with
  Ethernet, CSMA/CD must also be used
• When using a switched topology, a computer can
  send and receive data simultaneously (full-duplex
  communication)
• The collision detection circuitry is turned off
  because collisions arent possible
• Results in a considerable performance advantage

24
Ethernet Standards

• Each Ethernet variation is associated with an
  IEEE standard
• The following sections discuss many of the
  standards, some of which are obsolete or had
  limited use
• Keep in mind that Ethernet over UTP cabling has
  been the dominant technology since the early
  1990s, and will likely to continue to be for the
  foreseeable future

25
100 Mbps IEEE Standards

• The most widely accepted Ethernet standard today
  is 100BaseT, which is also called fast Ethernet
• The current IEEE standard for 100BaseT is 802.3u
• Subcategories
• 100BaseTX Two-pair Category 5 or higher UTP
• 100BaseT4 Four-pair Category 3 or higher UTP
• 100BaseFX Two-strand fiber-optic cable
• Because of its widespread use, the cable and
  equipment in fast Ethernet are inexpensive
• Architecture of choice for all but heavily used
  servers and multimedia applications

26
100BaseTX

• 100BaseTX is the standard thats usually in mind
  when discussing 100 Mbps Ethernet
• Requires two of the four pairs bundled in a
  Category 5 twisted-pair cable
• Although three cable types are available for
  100BaseT, 100BaseTX is the most widely accepted
• Generally called fast Ethernet

27
100BaseT4

• 100BaseT4 Ethernet uses all four pairs of wires
  bundled in a UTP cable
• Advantage capability to run over Category 3
  cable
• One of the biggest expenses of building a network
  is cable installation, so many organizations with
  Category 3 cabling chose to get the higher speed
  with the existing cable plant by using 100BaseT4
  instead of 100BaseTX

28
100BaseFX

• 100BaseFX uses two strands of fiber-optic cable
• Advantages
• Impervious to electrical noise and electronic
  eavesdropping
• Can span much greater distances between devices
• Disadvantage far more expensive than
  twisted-pair
• Rarely used as a complete 100BaseTX replacement
• Used as backbone cabling between hubs or switches
  and to connect wiring closets between floors or
  buildings
• Connect client or server computers to the network
  when immunity to noise and eavesdropping is
  required

29
100BaseT Design Considerations
30
100BaseT Design Considerations (continued)
31
10 Mbps IEEE Standards

• Four major implementations of 10 Mbps Ethernet
• 10Base5 Ethernet using thicknet coaxial cable
• 10Base2 Ethernet using thinnet coaxial cable
• 10BaseT Ethernet over UTP cable
• 10BaseF Ethernet over fiber-optic cable
• Of these 10 Mbps standards, only 10BaseT and
  10BaseF are seen today
• 10Base2 and 10Base5 are essentially obsolete

32
10BaseT
33
10BaseF
34
Gigabit Ethernet IEEE 802.3ab and 802.3z
Standards

• Gigabit Ethernet implementations
• 802.3z-1998 covers 1000BaseX specifications,
  including the L (long wavelength
  laser/fiber-optic), S (short wavelength
  laser/fiber-optic), and C (copper jumper cables)
• 802.3ab-1999 covers 1000BaseT specifications,
  which require four pairs of 100 ohm Category 5 or
  higher cable

35
1000BaseT
36
1000BaseLX
37
1000BaseSX
38
1000BaseCX
39
10 Gigabit Ethernet 10 Gbps IEEE 802.3ae Standard

• Defined to run only on fiber-optic cabling, both
  SMF and MMF, on a maximum distance of 40 km
• Provides bandwidth that can transform how WAN
  speeds are thought of
• Runs in full-duplex mode only
• CSMA/CD is not necessary
• Primary use as network backbone
• It also has its place in storage area networks
  (SANs)
• Will be the interface for enterprise-level servers

40
10 Gigabit Ethernet 10 Gbps IEEE 802.3ae
Standard (continued)

• Standards
• 10GBASE-SR Runs over short lengths (between 26
  and 82 meters) over MMF
• For high-speed servers, SANs, etc.
• 10GBASE-LR Runs up to 10 km on SMF
• For campus backbones and MANs
• 10GBASE-ER Runs up to 40 km over SMF
• Primary applications are for MANs
• 10GBASE-SW Uses MMF for distances up to 300 m
• 10GBASE-LW Uses SMF for distances up to 10 km
• 10GBASE-EW Uses SMF for distances up to 40 km

41
Whats Next for Ethernet?

• Implementations of 40 Gbps Ethernet are underway
• Ethernet could increase tenfold every 4-6 years
• 100 Gbps Ethernet available by 2006 to 2008,
  terabit Ethernet by 2011, and 10 terabit Ethernet
  by 2015
• In October 2005, Lucent Technologies demonstrated
  for the first time the transmission of Ethernet
  over fiber-optic cable at 100 Gbps
• It will be able to transfer data across the city
  faster than todays CPUs can transfer data to
  memory
• This level of speed has major implications for
  the entertainment industry and many other areas

42
Ethernet Frame Types

• Ethernet supports four non-compatible frame types
• Ethernet 802.3 used by IPX/SPX on Novell NetWare
  2.x and 3.x networks
• Ethernet 802.2 used by IPX/SPX on Novell NetWare
  3.12 and 4.x networks
• Supported by default in Microsoft NWLink
• Ethernet SNAP used in EtherTalk and mainframes
• Ethernet II is used by TCP/IP
• All Ethernet frame types support a packet size
  between 64 and 1518 bytes, and can be used by all
  network architectures mentioned previously

43
Ethernet 802.3
44
Ethernet 802.2

• Ethernet 802.2 frames comply completely with the
  Ethernet 802.3 standard
• The IEEE 802.2 group didnt address Ethernet,
  only the LLC sublayer of the OSI models layer 2
• Since Novell had already decided to use the term
  Ethernet 802.3 to describe Ethernet raw, its
  generally accepted that Ethernet 802.2 means a
  fully 802.3- and 802.2-compliant Ethernet frame
• Ethernet 802.2 frames contain similar fields to
  802.3, with three additional LLC fields

45
Ethernet SNAP

• Ethernet SubNetwork Address Protocol (SNAP) is
  generally used on the AppleTalk Phase 2
• It contains enhancements to the 802.2 frame,
  including a protocol type field, which indicates
  the network protocol used in the frames data
  section

46
Ethernet II
47
Wireless Ethernet IEEE 802.11b, a, and g

• AP serves as the center of a star topology
  network
• Stations cant send and receive at the same time
• CSMA/CA is used instead of CSMA/CD
• 802.11b/a/g use handshaking before transmission
• Station sends AP an RTS and it responds with CTS
• Standards define a maximum transmission rate, but
  speeds might be dropped to increase reliability
• No fixed segment length
• Maximum of 300 feet without obstructions
• Can be extended with large, high-quality antennas

48
The Token Ring Architecture
49
Token Ring Function

• A token passes around the ring
• If an in use token is received from NAUN, and
  the computer has data to send, it attaches its
  data to the token and sends it to its NADN
• If received token is in use, NIC verifies if it
  is the destination station
• If not, the computer re-creates the token and the
  data exactly and sends them to its NADN
• If it is, data is sent to the upper-layer
  protocols
• Two bits in data packet are toggled and token is
  sent to NADN when original sender receives it,
  it frees the token and then passes it along

50
Beaconing
51
Hardware Components

• A hub can be a multistation access unit (MSAU) or
  smart multistation access unit (SMAU)
• IBMs token ring implementation is the most
  popular adaptation of the IEEE 802.5 standard
• Minor variations but very similar to IEEE specs
• IBM equipment is most often used
• 8228 MSAU has 10 connection ports, eight of which
  can be used for connecting computers
• The RO port on one hub connects to RI port on the
  next hub, and so on, to form a ring among the
  hubs
• IBM allows connecting 33 hubs

52
Cabling in a Token Ring Environment
53
The Token Ring Architecture (summary)
54
The AppleTalk Environment

• Designed for use in Macintosh networks (1983)
• Can be run over several physical architectures
  commonly run over Ethernet (EtherTalk)
• Easy to implement
• Dynamic scheme used to determine devices address
• AppleTalk Phase 1 supported only 32 computers per
  network, and only with LocalTalk cabling
• With hubs/repeaters, increased the number to 254
• AppleTalk Phase 2, EtherTalk, and TokenTalk
  (1989) allow more than 16 million computers

55
LocalTalk

• LocalTalk uses STP in a bus topology to allow
  users to share peripherals and data in a small
  home or office environment
• CSMA/CA channel access method
• Avoids more collisions, but cumbersome
• Maximum transmission speed of 230.4 Kbps
• Thus, this architecture was used primarily in
  small, Macintosh-only environments

56
EtherTalk and TokenTalk

• EtherTalk is the AppleTalk protocol running over
  a 10 Mbps IEEE 802.3 Ethernet network
• TokenTalk is the AppleTalk protocol running over
  a 4 or 16 Mbps IEEE 802.5 token ring network
• Both implementations require using a different
  NIC
• Since 1996, Apple Computer has offered systems
  with built-in Ethernet NICs or with options to
  add Ethernet or token ring to its systems at a
  low cost
• Mac OS X with an Ethernet interface can freely
  participate in a Windows-based network

57
The Fiber Distributed Data Interface (FDDI)
Architecture
58
The Fiber Distributed Data Interface (FDDI)
Architecture (continued)
59
Networking Alternatives

• Many other network architectures are available
• Some are good for specialized applications, and
  others are emerging as new standards
• Topics
• Broadband technologies (cable modem and DSL)
• Broadcast technologies
• ATM
• ATM and SONET Signaling Rates
• High Performance Parallel Interface (HIPPI)

60
Broadband Technologies

• Baseband systems use a digital encoding scheme at
  a single fixed frequency
• Broadband systems use analog techniques to encode
  information across a continuous range of values
• Signals move across the medium in the form of
  continuous electromagnetic or optical waves
• Data flows one way only, so two channels are
  necessary for computers to send and receive data
• E.g., cable TV

61
Cable Modem Technology
62
Digital Subscriber Line (DSL)

• Competes with cable modem for Internet access
• Broadband technology that uses existing phone
  lines to carry voice and data simultaneously
• Most prominent variation for home Internet access
  is Asymmetric DSL (ADSL)
• Splits phone line in two ranges Frequencies
  below 4 KHz are used for voice transmission, and
  frequencies above 4 KHz are used to transmit data
• Typical connection speeds for downloading data
  range from 256 Kbps to 8 Mbps upload speeds are
  in the range of 16 Kbps to 640 Kbps

63
Broadcast Technologies

• By definition one-way transmissions
• This changed in Internet access by satellite
  television systems
• Work on the principle that most traffic a user
  generates is to receive files, text, and graphics
• The average users computer sends very little
  traffic
• User connects to service provider through a modem
• Service provider sends data by satellite to the
  users home at speeds up to 400 Kbps
• E.g., service offered by DirectTV, through its
  DirectPC add-on products

64
Asynchronous Transfer Mode (ATM)

• High-speed network technology for LANs and WANs
• Connection-oriented switches
• Dedicated circuits are set up before
  communicating
• Data travels in fixed-size 53-byte cells (5
  byte-header)
• Enables ATM to work at extremely high speeds
• Quick switching
• Predictable traffic flow
• Enables ATM to guarantee QoS
• Used quite heavily for the backbone and
  infrastructure in large communications companies
• LAN emulation (LANE) required for LAN applications

65
ATM and SONET Signaling Rates
66
High Performance Parallel Interface (HIPPI)

• HIPPI (late 1980s) high-speed interface
  developed for supercomputers and high-end
  workstations
• Serial HIPPI is a fiber-optic version that uses
  point-to-point optical links for bandwidth up to
  800 Mbps
• In early 1990s, it was used as a network backbone
  and for interconnecting supercomputers
• With the advent of Gigabit Ethernet, interest in
  HIPPI as a LAN backbone decreased
• HIPPI-6400 (1998) up to 6.4 Gbps transfer rates
• Known as Gigabyte System Network (GSN)
• HIPPI and GSN are now exotic networking products
  and arent often found in typical corporate
  networks

67
Summary

• Cable access methods determine how a network
  architecture gains access to the network medium
• A network architecture defines how data is
  placed, transmitted, and at what speed, and how
  problems in the network are handled
• DIX introduced Ethernet, which later became the
  IEEE 802.3 standard, transmitting data at 10 Mbps
• Standards for 10Mbps, 100Mbps, 1000Mbps
  (Gigabit), and 10G indicate the supported network
  mediums
• 10 Gigabit Ethernet runs only over fiber-optic
  cable and only in full-duplex mode

68
Summary (continued)

• Token ring networks are reliable, fast, and
  efficient
• Capable of transmitting at 4 Mbps or 16 Mbps
• Macintosh computers use AppleTalk to communicate
• FDDI is an extremely reliable, fast network
  architecture that uses dual counter-rotating
  rings
• Cable modem technology delivers high-speed
  Internet access to homes and businesses
• ATM, a high-speed network technology designed
  both for LANs and WANs, uses connection-oriented
  switches