Title: Chapter 2
1Chapter 2 Topics in Data Communications
2Data Communications ConceptsIntroduction
- Essential definitions for Data Communications
- Data, Signaling, Transmission Systems
- Analog Digital
- Data are entities that convey meaning, while
signaling is the transfer of encoded data thru a
transmission system - Analog versus digital signaling
- Digital signaling usually less expensive than
analog but care must be taken to properly
engineer system (e.g. - attenuation) - Combinations of analog digital data and signals
- Analog data -gt Analog signals
- Digital data -gt Analog signals (Key equipment is
a modem) - Analog data -gt Digital signals (Key equipment is
a codec) - Digital data -gt Digital signals
3Data Communications ConceptsAnalog versus
Digital Transmission Systems
- Analog systems transmit analog signals without
regard for the content of the signal - Amplifiers are used to boost the energy of the
signal - Amplifiers also boost the strength of any noise
on the line, introducing the possibility that the
signal could be lost - Digital Transmission Systems are concerned with
the content of the signal - Repeaters used to regenerate the signal,
overcoming attenuation - Repeaters output a new copy cleansed of any
noise, so noise is not cumulative (however, bit
errors can still occur if the signal is not
regenerated before it degrades too much)
4Data Encoding TechniquesIntroduction
- Encoding is the process of mapping digital data
into the appropriate signal elements for
transmission - Encoding may be very complex or as simple as
using binary signal elements (0s and 1s) - Encoding schemes are chosen to assist the
receiver in its two key tasks - Determining when the signal element begins and
ends (so sampling is done at the proper time) - Determining the value of the signal element (Is
it a one? A zero?) - Attenuation, data rate, noise all play a role
at receiver - With analog data the encoding scheme also plays a
key role in system performance but the details
are a little different
5Data Encoding TechniquesAnalog encoding of
digital data
- The basis for analog encoding is a base signal
called the carrier signal - Digital data is encoded (and decoded at the other
end) by a device called a modem - Three basic schemes for analog encoding of
digital data - Amplitude Shift-keying (ASK)
- Frequency Shift-keying (FSK)
- Phase Shift-keying (PSK)
- These schemes can be combined for more
sophisticated digital transmission systems that
carry more data per signal element
6Data Encoding TechniquesAnalog encoding of
digital data
- Amplitude-shift Keying
- Data represented by different amplitude levels of
the carrier signal - Simplest scheme, but inefficient and prone to
noise - Most valuable use is in optical systems
- Frequency-shift Keying
- Data represented by different frequency values
near the carrier signal frequency - Less prone to errors but requires more complex
circuitry - Phase-shift Keying
- Data represented by different phase shifts to the
carrier frequency - More efficient and noise resistant than ASK or
FSK but requires more complex circuitry
7(No Transcript)
8(No Transcript)
9Data Encoding TechniquesDigital Encoding of
Digital Data
- The most common way to encode digital data is to
use a binary signaling scheme consisting of two
voltage levels - NRZ-L (Non-Return to Zero Level)
- Each voltage level defines the value of the
digital data - Used only in very short connections
- NRZ-I (Non-Return to Zero Inverted)
- A transition at the beginning of a signal unit
denotes a binary one - This type of signaling is known as differential
signaling it is usually easier to detect a
transition out of the background noise and the
signals are polarity insensitive - Clocking and DC current are usually problems
10Data Encoding TechniquesDigital Encoding of
Digital Data
- Manchester Encoding
- Example of a bi-phase coding up to two signaling
transitions per signal element (needs more
bandwidth to transmit a given data rate) - The mid-signal transition provides clocking as
well as the data value (a zero data element is a
high-to-low transition and a one is a low-to-high
transition) - Used in Ethernet LANs (IEEE 802.3)
- Differential Manchester Encoding
- Another bi-phase code
- The mid-signal transition provides clocking the
transition at the beginning of the signal element
represents data (a zero data element has no
transition at the beginning of a bit time while a
one does) - Used in Token Ring LANs (IEEE 802.5)
11(No Transcript)
12Data Encoding TechniquesDigital Encoding of
Analog Data
- Pulse Code Modulation (PCM) is an example used
in the phone system to transmit analog data
across digital networks - Sampling rate based on the Nyquist theorem
- Digitized into 8 bit samples based on a nonlinear
scale that provides good reproduction of the
human voice - Other digital-to-analog encoding schemes
- Adaptive Differential Pulse Code Modulation
(ADPCM) - used with voice transmission - Delta Modulation - used rarely but also for voice
transmission systems - Code Excited Linear Prediction (CELP) - used in
very low-bandwidth voice and multimedia
communication systems
13Multiplexing Introduction
- Allows a transmission system to carry multiple
independent signals simultaneously for higher
efficiency - Two general schemes are in use FDM and TDM
- Frequency Division Multiplexing (FDM)
- Takes advantage of the fact that the useful
bandwith of the transmission system exceeds the
required bandwidth of a given signal - Allows frequency spectrum to be divided
allocated to different signal sources - Most commonly used with analog signaling and
transmission - Time Division Multiplexing (TDM)
14(No Transcript)
15MultiplexingTechniques
- Allows a transmission system to carry multiple
independent signals simultaneously for higher
efficiency - Two general schemes are in use FDM and TDM
- Time Division Multiplexing (TDM)
- Takes advantage of the fact that the maximum bit
rate of the system exceeds the required bit rate
of the digital signal - Each source is allocated a time slot in the
multiplexer - Analog signals can be time division multiplexed,
but it is very uncommon - Two varieties of TDM statistical and fixed
time-slot - Both FDM and TDM can be used in a synchronous or
asynchronous manner
16(No Transcript)
17Transmission Media Introduction
- The transmission media is the physical signal
path between the transmitter and the receiver - Can be guided (cables, waveguides, etc.) or
unguided (open air) - Our key concerns for transmission systems are
data rate and distance - Influencing factors
- Bandwidth of the media
- Transmission impairments
- Interference
- Number of Receivers
18Transmission Media (2)Twisted Pair Cable
- Consists of a minimum of two copper wires twisted
together and enclosed within a protective sheath - Advantages inexpensive, easy to work with, may
already be installed where needed - Disadvantages limited in distance, data rate,
and bandwidth susceptible to interference - Comes in two general varieties shielded twisted
pair (STP) and unshielded twisted pair (UTP) - Shielding provides more noise immunity,
especially at lower data rates - STP costs more and is more difficult to work with
19Transmission Media (3)Twisted Pair Cable
- Category 3 and Category 5 UTP
- Rating standards devised by the Electronic
Industries Association (EIA) - The higher the category the better the cable Cat
3 designed to support 10Mbps Ethernet while Cat 5
will support 100Mbps Ethernet - The key difference between the two categories is
the number of twists per unit length of cable - Near-end Crosstalk (NEXT) is a key transmission
impairment to minimize in any twisted pair
cabling system - While these are regarded as the most commonly
found UTP installations, there are higher
performance UTP choices
20Transmission Media Twisted Pair Cable
- High-performance Twisted Pair
- Category 5e (or enhanced Category 5) supports
125-MHz bandwidth on all four pairs, allowing
Gigabit Ethernet to run over UTP up to 100 meters - Attenuation (db/100m)12.3 for 100 MHz,
12.3-10Log(Vin/Vout) gives Log(Vin/Vout)-1.23
or Vin/Vout is about 1/10 (tenth of signal
magnitude exits from the UTP. - Crosstalk32db-10Log(Vneighbor/Vsignal), about
1/1000 crosstalk. - Cat 5 UTP provides 100 Mbps over 100m.
- Category 6 supports over 200-MHz bandwidth on
all four pairs could potentially run high data
rate ATM connections - Category 7 will require special shielding and
will likely support up to 700-MHz bandwidth on
each pair
21(No Transcript)
22Transmission Media Coaxial Cable
- Provides a two conductor transmission system
where one conductor is situated inside the outer
hollow conductor with an insulating dielectric in
between - Because of its structural characteristics coaxial
cable is more resistant to noise than twisted
pair - Harder to work with and more expensive than
twisted pair - Coax systems can be grouped in three categories
based on the type of signaling used - Baseband digital signaling occupies the entire
spectrum of the cable - Broadband carrier-band analog signaling is
used, allowing multiple channels on the cable - Carrierband carrier-band analog signaling with
low-end components signal occupies entire
spectrum of cable - Coaxial cable provides 100 Mbps over 1Km.
23Transmission MediaFiber Optic Cabling
- A transmission system composed of a guided medium
that allows the propagation of optical rays - A range of fiber optic cabling exists for various
needs, from ultra-pure fused silica (expensive
but high data rate) to plastic (cheap with lower
data rate for short runs) - Advantages
- Huge bandwidth capacity
- Smaller size and lightweight
- Lower attenuation
- Electromagnetic isolation (high security
minimal interference) - 100 Gbps over 10 Km (multimode fiber)
- Common transmitters used are LEDs for (low-cost
low-speed systems) or Injection Laser Diodes
(long-haul high-speed systems)
24Transmission MediaFiber Optic Cabling
- Basic fiber types
- Step-index multimode cheapest to manufacture
but allows light to travel different paths down
the fiber, causing signal distortion lowering
the maximum data rate. - Graded-index multimode Higher grade of fiber
with a varying refractive index that limits
distortion of the signal. - Singlemode contains a core with a diameter
close to the wavelength to be transmitted allows
only a single transmission path down the fiber
which practically eliminates distortion - Three wavelength windows provide the best light
propagation 850, 1300, 1550 nm - Most multimode systems use the 850 nm window
- Long-haul transmission systems use the 1550 nm
window because loss is lower at higher wavelengths
25Transmission Media Unguided Media
- Microwave
- Occupies the frequency spectrum from 1GHz to
30GHz can provide either a highly directional or
omni-directional system - There are 3 main challenges to using microwave
for data transmission - Frequency Allocation and licensing
- Interference
- Security
- Infrared
- Uses light in the infrared spectrum for data
transmission - Must be used line-of-sight or in an environment
that allows infrared waves to be reflected - Less issues associated with microwave but only
for specialized uses
26Data Communication NetworksIntroduction
- For most WAN and MANs, transmission of data
usually involves a number of intermediate
switching nodes that move the data between source
and destination - The complete set of end nodes, data links,
intermediate switches is known as a
communications network - There is a spectrum of communication switching
techniques the two main variations are circuit
and packet switching
27(No Transcript)
28Data Communication NetworksCircuit Switching
- Communication between end nodes is via a
dedicated communications channel - Communications via circuit switching involves
three phases - Circuit establishment the path is established
before any data is transferred. The path is
digital or analog and may include internal links
operated using TDM or FDM. - Data transfer
- Circuit disconnect release of resource
dedicated to the connection - The fixed capacity of the channel is allocated
for the duration of the connection can be very
inefficient with bursty traffic (repeatedly ON
during T and OFF later) - Circuit switching is best suited for synchronous
data such as voice or real-time video
29Data Communication NetworksPacket Switching
- Packet switching breaks data up into a series of
packets, each appended with enough control
information to ensure the packet transits the
network successfully from source to destination - Developed to address problems certain data
sources have with circuit switching - Bursty data transmission
- Source and destination must operate at the same
data rate - Inefficient resource allocation
- Connection setup can be too slow for certain
applications - (set up Virtual Circuit along logical connection
path links, send packets without routing decision
over VC) - In addition to addressing the above problems,
packet switching also has other benefits for data
transmission - Under heavy load the network will accept packets
but delay increases - Priorities for transmission of the packets can be
set - Data rate conversion along links with short store
and forward
30Data Communication NetworksPacket Switching
- Two main varieties datagrams or virtual
circuits - Datagram Approach Each packet is routed
independently of all others, leading to the
following consequences - Packets dont take the same routes, may arrive
out of sequence - Routing based on neighboring info on traffic and
failure - Possible packet discard (overflow in queues) and
no control-flow - No circuit setup time, so data flow begins
without delay - Data can easily flow around problems in the
network - Virtual Circuit Approach a preplanned route
through the network is established before any
data is sent - Requires logical circuit setup and teardown but
routes along the connection are shared
(identical) with other packets - A routing decision does not have to be made for
every packet - May provide enhanced services such as error
flow control, and packet sequencing not available
in a datagram environment
31(No Transcript)
32(No Transcript)
33(No Transcript)
34- Problem
- M bit to be transmitted using UDP and Virtual
Circuit (VC). t(UDP) and t(VC) is routing
overhead for UDP and VC for each link. There are
K links on the path. The largest packet has only
N bits and data rate is R. - Under what condition UDP and VC have the same
transmission time. - Using UDP T(UDP) kM/N(1/Rt(UDP))
- Using VC T(UDP) k(t(VC) M/NR)
- The same transmission time when kMt(UDP)/N
kt(VC). - However, UDP produces shorter time when M/N is
small.
35Data Communication NetworksHybrids
- Multi-rate Circuit Switching (to improve circuit
switching) - Extends circuit switching to allow one or more
fundamental channels to be bundled together to
provide a range of data connection rates - Examples of multi-rate switching are ISDN (2x64
Kbps and 1x16Kbps channels) and inverse
multiplexing - Frame Relay (FR)
- Packet switching was operating under high error
rate and overhead added to enhance redundancy and
reduce errors which produced the FR scheme (from
64 Kbps to 2 Mbps by removing overhead). - WAN service based on a connection-oriented packet
data protocol - Frame Relay evolved from X.25 the new protocol
was streamlined by eliminating features necessary
on earlier, less reliable X.25 data
communications networks - Cell Relay (ATM)
- A further evolution of connection-oriented packet
data services - Unlike frame relay fixed length data units
(cells) are used which allow high-speed hardware
based switching - Connection oriented, fixed cell size, fast
switching devices.
36Comparison
- Packet Switching vs. Circuit Switching
- Is packet switching a clear winner?
- Great for bursty data
- Resource sharing
- No call setup
- Excessive congestion packet delay and loss
- Protocols needed for reliable data transfer,
congestion control - Q How to provide circuit-like behavior?
- Bandwidth guarantees needed for audio/video apps
- Still an unsolved problem
37Routing
- Routing in Packet-Switched Networks
- Goal move packets among routers from source to
destination - Well study several path selection algorithms
(chapter 5) - Datagram network
- Destination address determines next hop
- Routes may change during session
- Analogy postal service
- Virtual circuit network
- Each packet carries tag (virtual circuit ID),
tag determines next hop - Fixed path determined at call setup time, remains
fixed through call - Routers maintain per-call state
38Taxonomy
39Network Access
- End hosts are connected to edge routers through
access networks - Types of access networks
- Residential access
- Company access
- Mobile access
- Types of physical media technologies for access
networks - Fiber
- Coaxial cable
- Twisted-pair telephone wire
- Radio spectrum
40Access Network
- Access Network Residential Access
- Connects home end systems to the network edge
- Typically, through an ISP
- End hosts are PCs
- AKA last mile
- Means of residential access dialup, DSL, Cable,
etc. - Dial-up modem
- Uses POTS line ? twisted pair copper wire
- Calls ISPs number
- Max. data rate 56 Kbps
- Phone line is tied up when connected to ISP
- Digital Subscriber Line (DSL)
- Does not tie up the phone line
- Uses existing twisted-pair line
- Asymmetric upstream and downstream data rates
- Downstream 384 Kb/s1.5 Mb/s
- Upstream 128256 Kb/s
- Hybrid Fiber Coaxial (HFC) Cable
41LAN access
- LAN access
- Company/university local area network (LAN)
connects end system to edge router - Ethernet
- Shared or dedicated cable connects end system and
router - 10 Mbs, 100Mbps, Gigabit Ethernet
- Deployment institutions, home LANs soon
- LANs Link layer (chapter 5)
- Wireless Access Networks
- Shared wireless access network connects mobile
end system to router at a base station - Laptops, PDAs, etc.
- Wireless LANs
- Radio spectrum replaces wire
- Wireless LANs are based on IEEE 802.11 b standard
(11 Mbps) - Wider-area wireless access
- CDPD (Cellular Digital Packet Data) wireless
access to ISP router via cellular network - Third Generation (3G) wireless packet-switched
wide-area Internet access at 384 Kbps
42Example 1
- How long will it take to send a file of 640,000
bits from host A to host B over a
circuit-switched network. - Suppose all links in the network are TDM with
- 24 slots and
- have a bit rate of 1.536Mbps
- It takes 500 msec to establish an end-to-end
circuit before host A begins transmitting to B - How long will it take to send file?
43Example 1
- How long will it take to send a file of 640,000
bits from host A to host B over a
circuit-switched network. - Suppose all links in the network are TDM with
- 24 slots and
- have a bit rate of 1.536Mbps
- It takes 500 msec to establish an end-to-end
circuit before host A begins transmitting to B - How long will it take to send file?
- Transmission rate for each circuit 1.536 Mbps /
24 64 Kbps - Time to send 640 Kbits file (640000 bits)/(64
Kbits/sec) 10 seconds - Including circuit setup overhead, time to send
file is 10.5 seconds - This calculation is independent of the of
end-to-end links and does not include propagation
delays
44Example 2
- Packet Switching
- Two forwarding mechanisms
- No segmentation ? message switching
- With segmentation?pipelining
- Example 7.5 million bits message sent over 3
links, each of 1.5 Mbps - Time required without segmentation
(7.5/1.5)x315 sec - Now segment packet into 5000 chunks each of 1500
bits - Time for whole packet 5.002 sec
- Pipelining results in reduction of delays as all
links are being utilized simultaneously
45Delay in Packet-Switched networks
- Transmission delay
- Rlink bandwidth (bps)
- Lpacket length (bits)
- Time to send bits into link L/R
- Propagation delay
- d length of physical link
- s propagation speed in medium (2x108 m/sec)
- Propagation delay d/s
46Example 3
- Packet Switching Calculation of delay
- A packet of L bits
- Q links between source and destination hosts
- Each link has a data rate of R bits/sec
- Assume
- No queuing delays
- No end-to-end propagation delays
- No connection establishment is required
- How long it takes to send this L bit packet from
source to destination? - Time to traverse the first link from source host
L/R seconds - Q-1 more such links are traversed before reaching
destination - Thus, total delay QL/R seconds ? more delay for
larger packets
47Example LAN Hierarchy
- N stations to be arranged in a LAN hierarchy
- Each ST transfers on the average M bits per
second, of which qM are to STs that are local to
its segment and (1-q)M to STs in other segments - How to arrange the LAN Hierarchy so that traffic
in one LAN segment does not exceed the segment
capacity B, where B is 0.8xData.Rate of a LAN
segment, i.e. a reference to a congestion point.
48Solution
- Suppose there are K segments each consists of N/K
stations, where k rages from 1 (1 seg) to N (N
segs). - In a segment, the locally destined traffic by is
qMN/K - In a segment, the exported traffic by is
(1-q)MN/K - In a segment, the imported traffic by is
(1-q)MN/K, because - there are (K-1) segments each generates (1-q)MN/K
external traffic, - Total external traffic for K-1 segments is
(K-1)(1-q)MN/K destined to (k-1) segments. - Thus, a segment imports (1-q)MN/K data.
- Total traffic for a segment is Q qMN/K (local)
(1-q)MN/K (exported) (1-q)MN/K (imported)MN/K
(1-q)MN/K MN(2-q)/K - Lets derivate Q w.r.t. K to find its minimum
value dQ/dK-MN(2-q)/K2 which is minimum for K
maximum or KN. - We may incease the number of STs in a segment
until Q(K)0.8xData.Rate or MN(2-q)/K lt
0.8xData.Rate or - MN(2-q)/(0.8xData.Rate) lt K.
49- wireless transmission
- Wireless local area networks (e.g IEEE 802.11)
- 2.4 GHz (microwave band) 1-2 Mbps, 150 m
- Infrared (IR) link 1-10 Mbps, 10 m
- Earth based cellular data
- Basic individual connection, 13 Kbps, 3 km
- Higher rate PCS (e.g. EDGE), 384 Kbps, 3 km
- Satellite links (GHz frequencies)
- geosynchronous, 600-1000 Mbps, continent
- Low earth orbit (LEO) 13 kbps - 400 Mbps, 800
km - (e.g., Motorola Iridium, 66 satellite
constellation) - services available from carriers
- Integrated Services Digital Network (ISDN) 144
Kbps - Asymmetric Digital Subscriber Lines 1.5-8
Mbps/16-640 Kbps - Cable modems 0.5-2 Mbps
- T1 (old electronic telephony standard) 1.544 Mbps
- T2 6.312 Mbps
- T3 44.736 Mbps
50(No Transcript)