Completing Ch 2: TCP/IP and OSI

1
Completing Ch 2 TCP/IP and OSI
Lecture 3
2
Physical addresses DL Layer

• Physical address is also known as the link
  address
• Physical address can be different sizes (depend
  on the network)
• Unicast type physical addresses single Rx
• Multicast type physical address multiple Rxs
• Broadcast type physical address all Rxs can
  pickup message

3
Physical Address Example
Most local area networks use a 48-bit (6 bytes)
physical address written as 12 hexadecimal
digits, with every 2 bytes separated by a hyphen
as shown below
07-01-02-01-2C-4B
A 6-byte (12 hexadecimal digits) physical
address
4
Logical Addressing - Network Layer
IP Addresses can be either unicast, multicast or
broadcast types
Going from network A physical address 10 to
network P physical address 95. Cant use the
physical address because different networks The
network layer address contains the uniqueness we
need from source to sink. Network layer address
is A-P Unit at this layer - datagram
5
IP Address Example
An Internet address (in IPv4) is 32 bits in
length, normally written as four decimal numbers
(or 4 octal numbers), with each number
representing 1 byte. How many bits is a byte ? A
nibble ?? The numbers are separated by a dot.
Below is an example of such an address. Call dot
notation
132.24.75.9
Example of IPv6 Address (128 bits)
6
Addresses in TCP/IP
Application Specific Address
Converts to a port address
7
Port addresses Transport Layer
Addresses of sending and receiving processes (j
and k)
Add IP address
Overhead (H2, T2) added for what ?
8
Port Address Example
A port address is a 16-bit address represented by
one decimal number as shown below.
753 A 16-bit port address
9
Relation- ship between Layers, Addresses, and
Units in TCP/IP
Messages
Segments
Datagrams (Packets)
Frames
Bits
Signals
10
Ch3 Underlying Technologies (1 of 3)
Lecture 3
11
Internet Underlying Technologies

• As we mentioned before, the Internet is an
  interconnection of backbone networks JOINED
  together via routers, gateways and switches
• Before getting into the higher level protocols,
  lets cover more concerning the underlying
  technologies. Lets talk about LANS and WANS and
  etc
• Internet is comprised of LANs, Point-to-Point
  WANs and Switched WANs
• We will cover LANS Ethernet, Token Ring (not in
  book), Wireless and FDDI Ring (not in book)
• We will cover Pt-to-Pt WANs Telephony Modem,
  DSL, Cable/Modem, T-Lines and SONET
• We will cover Switched WANs X.25, Frame Relay
  and ATM

Token Ring (not in book) FDDI Ring (not in book)
12
In putting these technologies in perspective

• Most of the technologies we are about to cover
  could be covered as a full blown course in some
  cases, multiple courses
• We will look at each technology from a high level
  in gathering a general appreciation and
  understanding of the technology

13
LOCAL AREA NETWORKS (LANS)

• LAN a data communication system connecting
  multiple INDEPENDENT devices such computers,
  servers, printers, etc..
• Covers up to a certain geographical area
  typically within a building or campus
• Some Popular LANs Ethernet, Token Ring, Wireless
  type LANs, and ATM LANs
• Ethernet LAN
• Ethernet is the more popular LAN protocol
• Designed in 1973 by Xerox
• Started out with a 10 Mbps data rate (bus
  topology)
• Today, 100 Mbps and 1 gigabit per second exist
  (gig1000 Mbps)
• IEEE 802.3 standard describes the Ethernet
  protocol

14
CSMA/CD

• The 802.3 standard describes the CSMA/CD standard
  as the access method for the original Ethernet
• CSMA/CD stands for carrier sense multiple access
  with collision detection.
• The transport medium is shared only one station
  or node can use the medium at a time
• All stations can receive a sent frame however,
  only the destination station takes it in (the
  other stations drop the frame)
• How can we make sure no two stations are using
  the transport medium at the same time ? If this
  happened, the 2 frames could collide
• CSMA/CD solves this problem

• CSMA/CD Process
• Every station has equal access to the medium
• Station listens to or senses the medium before
  sending frame if no data, it can send if data
  exist, wait
• Suppose 2 stations sense at the same time and
  find no data on the medium, crash will happen
• In this case, all stations sense the collision
  and each Tx send a jam signal to delete the
  data it sent
• Then each station waits a randomly amount of time
  and try it again this prevents a second
  collision

Notice that Station Z receives a collision signal
1 time period earlier than Station A
15
CSMA/CD

• Only one signal can travel down the transport at
  any time
• Node has to look out and make sure path is clear
• Which can be detected faster large or small
  signal ? Why (sensing)
• Sensing too much whats the problem
• Which will clear the transport faster ? Large or
  small ? Why (waiting)
• If you wait too little whats the problem
• If you wait too long whats the problem
• Whats the optimum wait time ?
• If the larger signal is moving very fast which
  can be detected faster ? Why ? Is speed a factor
  (sensing, waiting)
• Does the length of the transport play a factor in
  how fast something clears the hall ? (waiting)

16
CSMA/CD

• 3 factors relate to CSMA/CD
• 1. Minimum frame length
• 2. Data transmission rate
• 3. Collision domain (maximum network distance)
• The amount of time a station needs to wait in
  making sure no data is on the line) is minimum
  frame length divided by the data transmission
  rate. Why ?? (SpeedDistance/Time) the larger
  the frame, the longer the time to wait however,
  sensing is shorter)
• Amount of time to send the smallest frame (ie. an
  8 bit frame at 2 bps will take 8/2 4 seconds to
  send therefore, need to wait ATLEAST 4 seconds)
• Data transmission rate data transfer rate how
  fast to send a certain amount of bits from one
  device to another
• For the original Ethernet min frame size520
  bits, transmission rate10 Mbps and the max
  network distance2500 meters

17
Increasing Speed of Ethernet

• Decrease collision domain
• Increase minimum frame length

Detect faster
Larger frame
Smaller frame
18
Ethernet layers

• Ethernets data link layer is sub-divided into
  MAC Layer and LLC Layer
• MAC Layer media access control layer controls
  the CSMA/CD access method. Also performs the
  framing work.
• LLC Layer logical link control layer performs
  the error and flow control routines

19
Ethernet frame

• The Ethernet frame consist of 7 fields preamble,
  SFD, DA, SA, length/type of PDU, 802.2 frame (the
  actual data) and CRC
• Frame doesnt provide acknowledgment or hand
  shaking fields unreliable medium

• Preamble alternating 1s and 0s t alert and
  synchronize the Rx
• SFD start field delimiter signals the
  beginning of the frame
• Destination address contains the address of the
  next node( intermediate or Rx)
• Source address contains the address of the
  sending node (Tx or intermediate)

• Length of protocol data unit if less than 1518,
  defines the length up-and-coming data field if
  greater than 1536, tells the protocol that uses
  the service
• Data padding data encapsulated from higher
  layers size ranges between 46 bytes to 1500
  bytes
• CRC cyclic redundancy check error detection
  info

20
Ethernet implementation
Each device on an Ethernet network has a NIC
(network interface card)
Contains the physical address ah ha, this is how
I can change locations and still get emails

• Ethernet addressing
• Unicast
• Multicast
• Broadcast

• Some implementations of Ethernet
• 10BASE5 (thick ethernet)
• 10BASE2 (thin ethernet)
• 10BASE-T (twisted pair)
• 10BASE-FL (fiber link)

Connects host to medium and perform CSMA/CD
British naval connector or bayonet nut connector
for coaxial cable
21
Ethernet implementation
22
Ethernet implementation
Unshielded twisted pair
23
Ethernet implementation
24
Fast Ethernet implementation
100 Mbps Ethernet 2-wire type (100BASE-TX or
100BASE-FX) 4-wire type (only 100BASE-T4) To
make faster, collision domain was decreased 10
fold (250 meters vs 2500 meters)
25
Fast Ethernet implementation
26
Fast Ethernet implementation
27
Gigabit Ethernet implementation
Need for data rates higher than 100 Mbps resulted
in a 1000 Mbps Ethernet called gigabit
Ethernet Again, we had the choice to either
decrease the collision domain or increase the
minimum frame size Because 25 meters for the 100
Mbps Ethernet was short enough, the minimum frame
size was increased to get the desired
speed Another option is to do away with the
CSMA/CD overhead by connecting every station to
the hub using 2 separate paths (this will do away
with collisions) called full-duplex Ethernet
28
Gigabit Ethernet implementation
29
Ten-Gigabit Ethernet implementation
30
Token Ring LAN

• Token Ring is a protocol defined by IEEE 802.5
• Use a token passing ACCESS method

• Token Passing Method
• During idle times (network not being used), a
  token circulates
• The token is passed station to station until a
  station needs to send data
• When the station sends its data, it holds the
  token
• The data (or frame) circulates and get
  re-generated by each station
• The Rx takes in and COPY the frame (based on
  destination address)
• The data then continues back to the original Tx
• Token is then release to circulate

31
Token Ring layers

• Uses the same layers as Ethernet (MAC and LLC)
• LLC Layer logical link control layer performs
  the error and flow control routines (same as
  Ethernet)
• MAC Layer media access control layer it
  implements the Token Passing Access Method
  (versus Ethernets CSMA/CD access method)

Token Ring
32
Token Ring Data frame

• Token Ring frame defines 3 types of frames data,
  token and abort
• Data Frame carries a protocol data unit (actual
  data) and is addressed to a specific Rx (not
  broadcasted)
• Token Frame is the placeholder frame (token)
  and uses only 3 of the 9 fields (SD, AC and ED)
• Abort Frame doesnt carry any info and is used
  to stop transmission

• SD start delimiter alert and synch the Rx
• AC Access control - 3 bits set priority, 1 bit
  tells what type of frame, 1 bit is a monitor bit
  tells which station is monitoring or sending at
  the time, and 3 reservation bits for station
  wishing for access
• FC Frame control 1 bit tells if PDU is
  control info or data, 7 bits is used by Token
  Ring (ie. tells how to use AC field info)

• DA Destination address
• SA Source Address
• Data
• CRC cyclic redundancy check error checking
• ED end delimiter signals end of data
• FS frame status intermediate stations can set
  it letting the Tx know they read it, Rx can set
  it letting the Tx know it was copied and can be
  discarded now

33
Token Ring Implementation

• Token Ring is a series of shielded twisted pair
  transport medium linking each station into a ring
• Because the token needs to pass through each
  station with in the ring, if a station is down,
  it could be a problem
• Therefore, for each station, a switch is used to
  by pass the down (or disabled) station
• These bypass switches are packaged together as a
  MAU multi-station access unit

NOTE As we covered last lecture, dont confuse
the Ring Token technology with the Ring topology.
With a ring topology approach, you would want to
traverse in either direction (this is the main
benefit of a ring topology) explain Ethernet in
ring topology.
34
FDDI Ring

• FDDI stands for Fiber Distributed Data
  Interconnect
• Data rate is the same as Fast Ethernet (100 Mbps)
• Light signals versus electrical signals are used
• Uses a token passing access method with
  self-healing
• What do we mean by self healing ? Ability to
  detect and fix problems. The hardware
  automatically recognizes and fix problems

35
FDDI Ring

• How does the self-healing works ?
• Two independent rings connecting all stations are
  used dual counter-rotating rings
• The second ring is used only if a failure occurs
• Functions like a Token Ring LAN until a failure
  (ie. fiber cut, node failure)
• In this case, the intermediate (non-Rx) nodes
  keep copies of the sent frame too

36
FDDI Ring

• When the station detects it cant communicate
  with the adjacent station, it uses the second
  ring to bypass the adjacent station
• Given a fiber cut or node failure, this station
  is bypassed and the ring is closed

37
FDDI Frame