Multiplexing

1
Multiplexing

• Break one high-speed physical communication
  circuit into several lower-speed logical circuits
• Allows many different devices to use the circuit
  simultaneously while it seems that each pair of
  devices has the physical circuit all to itself
• Usually done in multiples of 4
• Two multiplexers are needed for a circuit one
  to combine the original circuits into a
  multiplexed circuit and one to separate them back
  out into separate circuits
• Four types of multiplexing
• Frequency division multiplexing (FDM)
• Time division multiplexing (TDM)
• Statistical time division multiplexing (STDM)
• Wavelength division multiplexing (WDM)

2
Frequency Division Multiplexing

• Divides the circuit horizontally by assigning
  different frequencies (channels) to each logical
  circuit
• All signals exist in the media at the same time
  on non-interfering frequencies
• Available bandwidth is divided into channels and
  guardbands which separate the channels
• Channels do not need to have identical capacities

3
Time Division Multiplexing

• Divides the circuit vertically by allowing
  different devices to transmit on the circuit at
  different times
• Devices take turns using the circuit to transmit
  data
• Time is allocated to each device in turn, even if
  the device is idle
• More efficient than Frequency Division
  Multiplexing, as there is no need for
  guardbands
• The full capacity of the line is divided evenly
  among the multiplexed circuits

4
Statistical Time Division Multiplexing

• Takes advantage of the fact that not all devices
  will be transmitting all the time
• Multiplexed circuit bandwidth is typically
  smaller than the combined bandwidths of the
  individual circuits
• Circuit capacity requirements are determined
  statistically by analyzing the usage of the
  circuits to be multiplexed
• Given four 64kbps circuits, TDM would require a
  256kbps multiplex circuit
• If we find that, statistically, only two of the
  four circuits are typically active
  simultaneously, we can provision a 128kbps
  multiplex circuit using STDM
• Provides more efficient use of bandwidth
• Can cause time delay if all circuits become
  active simultaneously
• Increased complexity and overhead, since each
  transmission must include an indication of the
  circuit it belongs to

5
Wavelength Division Multiplexing

• A version of FDM used in fibre-optic cabling
• Originally, fibre-optic transmission used a
  single color (light wavelength) for
  transmitting 622Mbps (622 million bits per
  second)
• By attaching devices that transmit and detect
  different light wavelengths, multiple circuits
  can be multiplexed over a single fibre-optic
  cable
• Dense Wave-Division Multiplexing (adding TDM to
  WDM) has increased the capacity of a single
  fibre-optic cable to 400 Billion bits per second
• DWDM Experimental results over 1.5Tbps
• 1,500,000,000,000 bits per second
• Approximately 200 Gigabytes/second

6
Inverse Multiplexing (IMUX)

• Opposite of multiplexing
• Combines two or more low-speed circuits making
  them appear as a single high-speed circuit
• Commonly used to provide T1 circuits in Wide Area
  Networks
• Combines 24 low-speed (64kbps) circuits to create
  a single 1.544 Mbps circuit
• IMUX equipment not standardized, so the same
  vendor should be used for both ends of the
  circuit
• BONDING (Bandwidth on Demand Interoperability
  Networking Group) standards have been adopted by
  some vendors for using 6 separate ISDN links over
  six telephone lines for room-to-room
  teleconferencing

7
Digital Subscriber Line (DSL)

• Provides high-speed data transmission over
  traditional telephone lines
• Limitations on traditional phone lines is based
  in the telephone and switching equipment in the
  end offices
• Actual cabling (local loop) is capable of much
  higher transmission capacity
• Conversion from POTS to DSL involves changing end
  equipment, not rewiring the local loop, making it
  very cost effective
• The customer premises equipment (CPE) includes a
  line splitter, which separates the traditional
  voice traffic from the data transmissions
• The line splitter directs data transmissions into
  the DSL modem (also called a DSL router) which is
  both a modem and a Frequency Division Multiplexer
• At the telephone company end office, the local
  loop enters into the Main Distribution Frame
  (MDF), which works like the line splitter on the
  customer premises
• The MDF sends voice traffic to the voice
  telephone network, and sends DSL traffic to the
  DSL Access Multiplexer (DSLAM)
• The DSLAM demultiplexes the data streams and
  converts them to ATM data which can be
  distributed to the Internet Service Provider
  (ISP)
• The ISPs Point of Presence (POP) can be
  co-located or located elsewhere

8
Asymmetric Digital Subscriber Line (ADSL)

• Uses Frequency Division Multiplexing to create
  three separate channels over the one local loop
  circuit
• One traditional voice circuit
• One relatively high-speed simplex circuit from
  the end office to the customer premises
• One slightly slower half-duplex circuit,
  primarily intended for upstream traffic from the
  customer premises to the end office
• It is called asymmetric because each of the two
  data circuits have different capacities
• Each of the two data channels can be further
  subdivided using Time Division Multiplexing
• The size of the two digital channels depends on
  the distance from the CPE to the end office
• Shorter distance to End Office results in less
  attenuation of the signal, allowing higher
  frequencies to be used and yielding a faster
  connection
• Offering higher speed ADSL limits the number of
  potential customers, while offering lower speed
  decreases product attractiveness

9
Cable modems

• Digital service offered by cable companies
• Data over Cable Service Interface Specification
  (DOCSIS) is the dominant standard
• Standard most often used by cable companies
  running hybrid fibre coax (HFC) networks
• Architecture is very similar to DSL
• Main difference is that DSL is point-to-point,
  while cable modems share multi-point circuits
• Each user must compete with other users for the
  available capacity
• All messages on the circuit go to all computers
  on the circuit
• Cable TV circuit enters the customer premises
  through a cable splitter
• TV signals are sent to the TV network
• Data signals are sent to the cable modem (both a
  modem and an FDM)
• Standard coaxial cable circuit may be shared by
  from 300 to 1000 customers
• Not all of these cable TV customers will
  subscribe to high-speed internet
• The coax runs to a fibre node with an
  optical-electrical converter
• Each fibre node may service up to half a dozen
  cable circiuts
• The fibre nodes are connected to a distribution
  hub (also called a headend) through two separate
  circuits
• Upstream circuit connects to a cable modem
  termination system (CMTS)
• Downstream circuit carries both ordinary cable TV
  signal and downstream traffic destined for the
  customer premises

10
Local Area Networks

• Introduction
• Why use a LAN? Dedicated servers vs. Peer-to-peer
  LANs
• LAN Components
• NICs, Cables, Hubs and Network Operating Systems
• Traditional Ethernet (IEEE 802.3)
• Topology, Media Access Control, Ethernet Types
• Switched Ethernet
• Topology, Media Access Control, Performance
  Benefits
• Wireless LANs (IEEE 802.11)
• Topology, Media Access Control, Wireless Ethernet
  Types
• Other Wireless Technologies
• Infrared Wireless, Bluetooth
• Improving LAN Performance
• Improving Server Performance, Improving Circuit
  Capacity, Reducing Network Demand

11
Why use a LAN?

• There are two main benefits to using a local area
  network information sharing and resource
  sharing.
• Examples of information sharing include file
  sharing, exchanging e-mail, and using the
  Internet.
• Examples of resource sharing include sharing
  hardware and software, such as sharing an
  expensive printer.
• Another important resource sharing technique is
  to purchase software on a per seat basis. For
  example, only purchasing a 10-seat license for a
  software program on a 20 client network instead
  of purchasing 20 copies of the same program.

12
Dedicated Server Networks

• A basic LAN dichotomy exists between dedicated
  server LANs and peer-to-peer LANs which dont
  have servers. Since 90 of all LANs have a
  dedicated server, this chapter mostly focuses on
  server-based LANs.
• A dedicated server is a computer that is
  permanently assigned a specific server task such
  as being a Web server, e-mail server, file server
  or printer server.
• Servers also run a special operating system
  called a server network operating system.
• When many servers are part of a network, it can
  be referred to as a server farm.

13
Peer-to-Peer Networks

• Peer-to-peer networks do not use dedicated
  servers.
• Any computer on a peer-to-peer network can act as
  both a client, accessing resources or information
  on other computers on the network, or as a
  server, allowing access to attached information
  or resources.
• Peer-to-peer networks tend to be small networks.
• The main advantage of peer-to-peer networking is
  lower cost since there is no dedicated server,
  generally the most expensive network component.
• The main disadvantage is that peer-to-peer
  networks are generally slower than dedicated
  server networks, since each computer is less
  powerful and may be in use as a client and a
  server at the same time.

14
Basic LAN Components

• The six basic LAN components are
• 1. Clients
• 2. Servers
• 3. Network Interface Cards
• 4. Network Cables
• 5. Hubs and Switches
• 6. Network Operating System

15
Basic LAN Components
16
Network Interface Cards

• Network interface cards, also called network
  cards and network adapters include a cable socket
  allowing computers to be connected to the
  network.
• NICs are part of both the physical and data link
  layer and include a unique data link layer
  address (sometimes called a MAC address), placed
  in them by their manufacturer.
• Before sending data onto the network, the network
  card also organizes data into frames and then
  sends them out on the network.
• Notebook computers often use NICs that are
  plugged into the PCMCIA port.

17
Network Cables

• Each computer is physically connected to the
  network using a cable.
• These cables are either untwisted wire pairs
  (UTP, the most common choice), shielded twisted
  pair (STP), coaxial cable, or optical fiber.
• Wireless LANs use radio frequencies or infrared
  light instead of cables.
• Sometimes two different types of cabling can be
  linked using a special connector. A BALUN
  (Balanced-Unbalanced) is one such device that
  connects UTP and Coaxial Cable.

18
Hubs

• Hubs act as junction boxes, linking cables from
  several computers on a network. Hubs are usually
  sold with 4, 8, 16 or 24 ports.
• Some hubs allow connection of more than one kind
  of cabling, such as UTP and coax.
• Hubs also repeat (reconstruct and strengthen)
  incoming signals. This is important since all
  signals become weaker with distance.
• The maximum LAN segment distance for a cable can
  therefore be extended using hubs.

19
Network Hub
20
Network Operating Systems

• The NOS is the software that runs the LAN. It
  comes in two types Server NOSs Client NOSs.
• Server NOSs enable server to execute and respond
  to the requests sent to them as web server, print
  servers, file servers, etc.
• Client NOS functions are typically included in
  most OS packages such as Windows 98 and Windows
  2000.

21
Network Profiles

• The network profile specifies what resources on
  each server are available to the network for use
  by other computers, including data files,
  printers, etc.
• Devices that are not included in the network
  profile can not be used over the network.
• User profiles describe what each user on a LAN
  has access to.
• Most LANs also use auditing software which keeps
  track of which user has accessed what network
  resource.

22
Ethernet (IEEE 802.3)

• Almost all LANs today use Ethernet
• Originally, Ethernet was jointly developed by a
  consortium of Digital Equipment Corp., Intel and
  Xerox and was standardized as IEEE 802.3.
• Ethernet LANs that use hubs are sometimes called
  shared Ethernet.

23
Shared Ethernet Topology

• Ethernets logical topology is a bus topology.
• This means all computers on the network receive
  messages from all other computers, whether the
  message is intended for those computers or not.
• When a frame is received by a computer, the first
  task is to read the frames destination address
  to see if the message is meant for it or not.
• Although, a decade ago most Ethernet LANs used a
  physical bus, almost all Ethernets today use a
  physical star topology, with the networks
  computers linked into hubs.
• It is also common to link use multiple hubs to
  form more complex physical topologies

24
Ethernet Topology
25
Multiple Hub Ethernet Design
26
Media Access Control

• Ethernets medium access control protocol, called
  CSMA/CD, is contention-based
• With a contention-based protocol, frames can be
  sent by two computers on the same network at the
  same time, in which case they will collide and
  become garbled.
• CSMA/CD, can thus be termed ordered chaos
  because it tolerates, rather than avoids,
  collisions caused by two computers transmitting
  at the same time.

27
CSMA/CD

• Stands for Carrier Sense Multiple Access w/
  Collision Detect
• Carrier Sense computers listen to the network to
  see if another computer is transmitting before
  sending anything themselves.
• Multiple Access all computers have access to the
  network medium.
• Collision Detect if they detect a collision
  (CD), they then wait a random amount of time and
  resend the frame (It has to be random in order to
  avoid another collision).

28
Ethernet Physical Media Standards

• Ethernet Media are formatted as follows
  Value1Base/BroadValue2
• Value 1 Data Rate for Medium 10 10Mbps
• Base or Broad
• Base Baseband Mode meaning only one (digital)
  channel
• Broad Broadband (analog) cable transmissions
  use more than one channel (e.g., cable TV)
• Value2 (relates to maximum distance possible in
  hundreds of meters or cable type T twisted pair,
  F fiber)

29
Types of Ethernet

• Seven types of shared Ethernet have been in use
• 10Base5 thick Ethernet, uses thick coax. This
  is the original Ethernet specification. Now
  uncommon.
• 10Base2 thin Ethernet, uses thin coax. Became
  popular in the early 1990s as a cheaper
  alternative to 10Base5. Now uncommon.
• 10BaseT twisted pair Ethernet, most common type
  of Ethernet. Uses Cat 3 and Cat 5 UTP. Common but
  rapidly losing ground to 100BaseT.
• 100BaseT also called Fast Ethernet, has
  replaced 10BaseT in sales volume. Uses Cat 5 UTP
  (Sometimes combined 10/100 Ethernet is found in
  which some segments run 10BaseT and some run
  100BaseT is also used by some organizations).
• 1000BaseT Gigabit Ethernet. Maximum cable
  length is only 100 meters.
• 10GbE 10 Gbps Ethernet. Uses fiber and is
  typically full duplex.
• 40GbE 40 Gbps Ethernet. Uses fiber and is
  typically full duplex.

30
1.       
Types of Ethernet
31
Switched Ethernet Topology

• Switched Ethernet uses switches instead of hubs.
  
• While a hub broadcasts frames to all ports, the
  switch reads the destination address of the frame
  and only sends it to the corresponding port.
• The effect is to turn the network into a group of
  point-to-point circuits and to change the
  logical topology of the network from a bus to a
  star.

32
Basic Switch Operation

• Switches make forwarding decisions based on
  forwarding tables (similar to routing tables).
• When a frame is received, the switch reads its
  data link layer destination address and sends
  the frame out the corresponding port in its
  forwarding table.
• Switches making switching decisions based on data
  link layer addresses are called layer-2 switches.
• When a switch is first turned on, its forwarding
  table is empty. It then learns which ports
  correspond to which computers by reading the
  source addresses of the incoming frames along
  with the port number that the frame arrived on.
• If the switchs forwarding table does not have
  the destination address of the frame, it
  broadcasts the frame to all ports.
• Thus, a switch starts by working like a hub and
  then works more and more as a switch as it fills
  its forwarding table.

33
Media Access Control

• Switched Ethernet still uses CSMA/CD media access
  control, but collisions are less likely as each
  network segment operates independently.
• The networks modified topology also allows
  multiple messages to be sent at one time.
• For example, computer A can send a message to
  computer B at the same time that computer C sends
  one to computer D.
• If two computers send frames to the same
  destination at the same time, the switch stores
  the second frame in memory until it has finished
  sending the first, then forwards the second.

34
802.3 Ethernet versus switched Ethernet
35
Performance Benefits

• Switched Ethernet can dramatically improve
  network performance.
• Shared Ethernet 10BaseT networks are only capable
  of using about 50 of capacity before collisions
  are a problem
• Switched Ethernet, however, runs at up to 95
  capacity on 10BaseT.
• Another performance improvement can be made by
  using a 10/100 switch that uses a 100BaseT
  connection for the server(s) and/or routers,
  i.e., the network segments experiencing the
  highest volume of LAN traffic.

36
Wireless Ethernet (IEEE 802.11)

• Wireless LANs dispense with cables and use radio
  or infrared frequencies to transmit signals
  through the air.
• WLANs are growing in popularity because they
  eliminate cabling and facilitate network access
  from a variety of locations and for mobile
  workers (as in a hospital).
• The most common wireless networking standard is
  IEEE 802.11, often called Wireless Ethernet or
  Wireless LAN.

37
Wireless LAN Topology

• WLAN topologies are the same as on Ethernet
  physical star, logical bus
• Wireless LAN devices use the same radio
  frequencies, so they must take turns using the
  network.
• Instead of hubs, WLANs use devices called access
  points (AP). Maximum transmission range is about
  100-500 feet. Usually a set of APs are installed
  making wireless access possible in several areas
  in a building or corporate campus.
• Each WLAN computer uses an NIC that transmits
  radio signals to the AP.
• Because of the ease of access, security is a
  potential problem, so IEEE 802.11 uses 40-bit
  data encryption to prevent eavesdropping.

38
A wireless Ethernet access point connected into
an Ethernet Switch.
39
WLAN Media Access Control

• Wireless LANs use CSMA/CA where CA collision
  avoidance (CA). With CA, a station waits until
  another station is finished transmitting plus an
  additional random period of time before sending
  anything.
• Two different WLAN MAC techniques are now in
  use the Physical Carrier Sense Method and the
  Virtual Carrier Sense Method.

40
Physical Carrier Sense Method

• In the physical carrier sense method, a node that
  wants to send first listens to make sure that the
  transmitting node has finished, then waits a
  period of time longer.
• Each frame is sent using the Stop and Wait ARQ,
  so by waiting, the listening node can detect that
  the sending node has finished and can then begin
  sending its transmission.
• With Wireless LANs, ACK/NAK signals are sent a
  short time after a frame is received, while
  stations wishing to send a frame wait a somewhat
  longer time, ensuring that no collision will
  occur.

41
Virtual Carrier Sense Method

• When a computer on a Wireless LAN is near the
  transmission limits of the AP at one end and
  another computer is near the transmission limits
  at the other end of the APs range, both
  computers may be able to transmit to the AP, but
  can not detect each others signals.
• This is known as the hidden node problem. When it
  occurs, the physical carrier sense method will
  not work.
• The virtual carrier sense method solves this
  problem by having a transmitting station first
  send a request to send (RTS) signal to the AP. If
  the AP responds with a clear to send (CTS)
  signal, the computer wishing to send a frame can
  then begin transmitting.

42
Types of Wireless Ethernet

• Two forms of the IEEE 802.11b standard currently
  exist, utilizing the 2.5 GHz band
• Direct Sequence Spread Spectrum (DSSS) uses the
  entire frequency band to transmit information.
  DSSS is capable of data rates of up to 11 Mbps
  with fallback rates of 5.5, 2 and 1 Mbps. Lower
  rates are used when interference or congestion
  occurs.
• Frequency Hopping Spread Spectrum (FHSS) divides
  the frequency band into a series of channels and
  then changes its frequency channel about every
  half a second, using a pseudorandom sequence.
  FHSS is more secure, but is only capable of data
  rates of 1 or 2 Mbps.
• IEEE 802.11a uses Orthogonal Frequency Division
  Multiplexing (OFDM), operates in the 5 GHz band
  with data rates of up to 54 Mbps.
• IEEE 802.11g uses OFDM in the 2.5 GHz band,
  operates at up to 54 Mbps, and is compatible with
  802.11b

43
Infrared Wireless LANs

• Infrared WLANs are less flexible than IEEE 802.11
  WLANs because, as with TV remote controls that
  are also infrared based, they require line of
  sight to work.
• Infrared Hubs and NICs are usually mounted in
  fixed positions to ensure they will hit their
  targets.
• The main advantage of infrared WLANs is reduced
  wiring.
• A new version, called diffuse infrared, operates
  without a direct line of sight by bouncing the
  infrared signal off of walls, but is only able to
  operate within a single room and at distances of
  only about 50-75 feet.

44
Infrared Wireless LAN
45
Bluetooth

• Bluetooth is a 1 Mbps wireless standard developed
  for piconets, small personal or home networks.
• It may soon be standardized as IEEE 802.15.
• Although Bluetooth uses the same 2.4 GHz band as
  Wireless LANs it is not compatible with the IEEE
  802.11 standard and so can not be used in
  locations that use the Wireless LANs.
• Bluetooths controlled MAC technique uses a
  master device that polls up to 8 slave devices.
• Examples of Bluetooth applications include
  linking a wireless mouse, a telephone headset, or
  a Palm handheld computer to a home network.

46
Improving LAN Performance

• As networks become more and more intensively
  used, LAN performance becomes a critical issue.
• The measure of LAN Performance is throughput,
  i.e., the total amount of user data transmitted
  in a given period of time.
• LAN performance can be improved by identifying
  and eliminating bottlenecks, that is, points in
  the network where congestion is occurring because
  the network or device cant handle all of the
  demand it is experiencing.

47
Identifying Network Bottlenecks

• Two common network bottlenecks are related to
  server access
• If server performance is poor when server
  utilization is high (gt60), then the bottleneck
  is the server.
• If server performance is poor during periods of
  low server utilization (lt40), then the
  bottleneck is not the server but the network
  circuit.

48
Improving Server Performance

• Two types of server performance improvements are
  possible
• Software improvements such as choosing a faster
  Network Operating System, fine tuning network and
  NOS parameters for optimal performance.
• Hardware improvements such as adding a second
  server, upgrading the servers CPU, increasing
  its memory space, adding more hard drives or
  adding a second NIC to the server.

49
Improving Server Performance RAID

• Improving disk drive performance is especially
  important, since disk reads are the slowest task
  the server needs to do.
• Replacing one large drive with many small ones
  can improve server performance.
• RAID or Redundant Array of Inexpensive Disks,
  builds on this idea. RAID system can be used to
  both improve performance and increase reliability
  by building redundancy into the hard drives, so
  that a hard drive failure does not result in any
  loss of data.

50
Improving Circuit Capacity

• Improving circuit capacity can be done simply by
  upgrading one or all segments of a network to a
  faster protocol (which also means upgrading the
  NICs), such as
• Upgrading the network from 10BaseT to 100BaseT,
  or
• Upgrading the network segment to the server from
  10BaseT to 100BaseT
• Another approach to improving circuit capacity is
  by increasing the number of network segments to
  the server. Most servers can handle several
  network segments simply by adding additional NIC
  cards, thereby increasing access to the server

51
Network Segmentation a. Before b. After
52
Reducing Network Demand

• Performance can also be improved by reducing
  network demand. This can be done by
• Moving more files, such as heavily used software
  packages to client computers.
• Disk caching, software on client machines can
  also reduce server demand.
• Moving user demands from peak times to off peak
  times, by telling network users when peak usage
  times occur and encouraging users to not use the
  network as heavily during these times can also
  help improve performance.
• Delaying some network intensive jobs to off-peak
  times, such as running heavy printing jobs at
  night, can also improve performance.

53
Improving LAN performance

• Increase Server Performance
• Software Fine-tune the NOS settings
• Hardware
• Add more servers and spread the network
  applications across the servers to balance the
  load
• Upgrade to a faster computer
• Increase the server's memory
• Increase the number and speed of the server's
  hard disk(s)
• Upgrade to a faster NIC
• Increase Circuit Capacity
• Upgrade to a faster circuit
• Segment the network
• Reduce Network Demand
• Move files from the server to the client
  computers
• Increase the use of disk caching on client
  computers
• Change user behavior