Transcript EETS8320 Fall 2006 - Lyle School of Engineering

Digital Telecommunications
Technology - EETS8320
Fall 2006
Lecture 1
Overview and Introduction
(Slides with notes.)
Page 1
© 1996-2006 R.Levine
Introduction: EETS8320
• Subject Area: Digital coding and multiplexing of
telecommunications transmissions (formerly in
course EETS8302)
• Digital telecommunications switching (formerly
in course EETS8304)
• Descriptive and semi-technical treatment
– About 70% of our students do not have an
engineering or science undergraduate degree,
although many work in the telecom industry.
– Each student can write a term paper at a technical
level appropriate to their own background and
knowledge
Page 2
© 1996-2006 R.Levine
Course Administrative Matters
• 13 weeks of class each 3 hours (consecutive), with
slides and notes
• Each student takes a multiple-choice midterm quiz
(1hour) and writes a term paper (approx 20 pages or
5000 words) on a pre-approved topic.The letter
grade on the term paper is substantially your final
grade.
• If your midterm quiz numerical grade is above
average* your final letter grade is increased by one
“step” on SMU’s grade scale: Example, B+  A• If your midterm is below average, no deduction is
made. Your paper grade is then your course grade.
*Notes explain details.
Page 3
© 1996-2006 R.Levine
Course Objectives
• One objective: to give students sufficient
understanding of the technology to make
intelligent decisions in the present and future
– This course is focused on science and technology, because
understanding technology is important.
– Adequate understanding of both technology and business are
very important in the telecommunications industry.
– Knowledge of business-economics alone is not sufficient!
– Knowledge of technology alone, with ignorance of economics is
also not sufficient!
• The Iridium system, ISDN and In-flight telephones are unsuccessful telecom products often cited as examples of
economic or technological bad judgement.
– Human interface (ease of use) is also a factor in some cases.
Page 4
© 1996-2006 R.Levine
Problem Products: Iridium
• Iridium, a world-wide direct satellite telecom
system of 1990s
• Technologically impressive, but…
– Priced higher than most potential customers would
pay:
• Handset $3000 (price later significantly reduced)
• Service $3/min or more (price later slightly reduced)
• Designers and implementers were aware of possible low
sales risk due to high prices.
– Unexpected low cost terrestrial competition in
populated areas harmed Iridum .
Page 5
© 1996-2006 R.Levine
Iridium: More Recent History
• High-budget Customer Base was never large enough:
– For example, oil exploration crews in Siberia? Very few of these!
– Native farmers in Kazakhstan? They could not afford Iridium.
– Callers from an ocean liner? Sounds promising:
• Existing Inmarsat satellite telephone calls are $10/minute!
• But e-mail to/from most ocean liners is free!!
• If you build it, will they come?… Apparently, No!
– Customer enrolment was only a tiny fraction of Iridium management’s
estimates. Several top Iridium executives resigned.
– Iridium Corp. filed for Chapter 11 bankruptcy protection in August, 1999.
This allows continued operation while a plan is made to hopefully
reorganize and eventually pay creditors (bondholders, etc.). Shareholders
are not protected in Chapter 11. Reorganized as Iridium Satellite LLC in
Dec. 2000
– US Government subscribers are almost the only present users, while
Iridium operates in reorganization.
– Development of several competitive LEO satellite systems (e.g.
Globalstar) stopped. Licenses were cancelled or returned to the FCC.
Page 6
© 1996-2006 R.Levine
Problem Products: ISDN
• Fully digital end-to-end telecommunication via 64
kbit/s channels derived from pre-existing digital
telecom channels
• Was viewed as the unquestioned future direction
of PSTN voice-data service in the 1980s.
– Bone of contention among major telephone
switch manufacturers.
• Ultimately in limited use but not in consumer
demand due to high cost
– Unanticipated availability of low-cost 53 kbit/s
V.90 modems in 1990s diverted many potential
customers
Page 7
© 1996-2006 R.Levine
50
Z 30
S1
15
Y
A
.
S2
W
100
Motor
Fuel.
30
D2
.25 0.50
1.00
X
Unit price ($)
.
50
5.00
B
Quantity minutes bought or sold
100 D1
Quantity bought or sold
Quantity bought or sold
A Page from Economic Theory
100
50
10.00
Unit price ($/gal)
In-flight
Phone
Service.
.
0.50
1.00
Unit price/min ($)
C
• Supply-demand curves graphically describe
how buyers and sellers find an equilibrium
price per unit and quantity sold. See notes.
Page 8
© 1996-2006 R.Levine
Problem Products: In-Flight Telephone Service
•
Very costly to install due to severe aircraft radio interference standards.
– Originally allowed only aircraft-originated calls. Aircraft destination
calls are supported in some systems in a somewhat inconvenient way..
•
None were ever profitable. Many competitive systems addressed improved
convenience, video games, etc., but not low price.
– Test market studies show that sales improve dramatically below 20
cents/min, but existing systems can't meet that price.
•
Verizon Airfone plans, after 21 years, to discontinue service by the end of
2006. In-flight phone service will end on over 1,000 aircraft operated by
United, Delta, Continental, US Airways, Air Canada and Cathay Pacific.
– Proposed systems using customers' existing cell phones inside the aircraft are
still in preliminary development.
– Use of VoIP via Internet links (priced at ~$10/h)) provided in some flights is
another alternative
•
Historical note: First (analog) Airphone system developed by Jack
Goeken, also famed as founder of MCI.
Page 9
© 1996-2006 R.Levine
Customer Perception of
“Fairness” is Important
• Some system proposals did not succeed due to
negative customer perception of “fairness”
– Two types of limited play video disks were test
marketed circa 1998 as “no return” methods for
video rentals. Both rejected by customers.
– System software for wireless air time charges
paid by land-line originator were developed,
due to industry pressure circa 2000, but 100%
of participants in US marketing tests would not
choose this billing method.
Page 10
© 1996-2006 R.Levine
Concerns about 3G Wireless
•
Some telecom industry observers fear that 3G, and other advanced wireless
data technologies, will suffer fates like those of Iridium and ISDN.
– First Generation (1G) wireless was analog cellular technology used
from1981 to mid 1990s. Very few still in use.
– Second Generation (2G) wireless utilizes digital speech coding, used
from early 1990s to the present. Technologies include GSM, CDMA, North
American TDMA, iDEN (NexTel) and others. Range of available bit rate
per user is about 6 kbit/s to 20 kbit/s
– Two and a Half (2.5G or 2-1/2G) designs are packet data systems,
achieving available bit rate per user up to about 60 kbit/s to 140 kbit/s.
Used for Internet access and packet voice (VoIP).
– Third Generation (3G) utilizes various types of CDMA (UMTS,
CDMA2000). Provides bit rates up to 2 Mbit/s. Major applications are
viewing high quality visual entertainment (HDTV images) or possibly
transferring massive files via Internet.
– Fourth Generation (4G) utilizes OFDM to achieve 16 Mbit/s or higher bit
rates. Applications similar to 3G but even “faster.”
Page 11
© 1996-2006 R.Levine
2.5 G
• These “3G sceptics” believe that the major growth in cellular
industry will come from lower cost voice service and less
glamorous data services like e-mail.
– They therefore designed a packet data technology upgrade
based on GSM, called GPRS (and a higher data rate version
named EDGE), called 2.5G, or “2 and a half G.”
– The cost of installing GPRS or EDGE in an existing GSM base
system is relatively small. (In contrast, all 3G systems require
costly new or additional base radio replacements.)
• GPRS upgrades are already in place in some countries in Europe,
N.America, Asia, Australia, etc.
• Use of GPRS (later EDGE) in USA (by AT&T and Cingular) to replace
IS-136 TDMA is under way today. Voice Stream (T-Mobile) has a GSM
starting point technology and thus a less costly upgrade to GPRS and
EDGE. Merger rumors between GSM-technology firms are continually
rife.
– GPRS provides up to 171 kbit/s per subscriber, EDGE up to 384
kbit/s.
Page 12
© 1996-2006 R.Levine
Other Aspects
• Previous slides did not analyze or compare:
– Radio bandwidth required for higher bit rates.
– Sensitivity of each technology to “noise” and
interference limits the number of simultaneous
“conversations” in each cell and thus the system
capacity
– Cost and complexity of each technology
– Power consumption in active and standby modes,
affecting battery “life.”
– An accumulation of negative aspects like these can
severely degrade the theoretical performance of
real systems
Page 13
© 1996-2006 R.Levine
Customer Preference Issues
• In some cases, potential customers won't buy
because they perceive cost or terms of sale
inherently unattractive. Examples:
• Today North American wireless subscribers view air
time costs over about US$ 0.20/min as excessive
• North American telephone users reject “caller pays”
for wireless destination calls (although this is
accepted in many other countries).
• In a non-telecom case, customers reject a “self
destructing” or “pay per view” video disc.
Page 14
© 1996-2006 R.Levine
Sometimes Non-Technical Problems Dominate
• “Morse code” telegraph was not practical for
most end users because of the special skill
required to send with a “key” and receive by
listening to “di-dahs”
– Electro-mechanical Teletypewriter machines only
require the ability to read and type (keyboard) but
they were costly, bulky and noisy.
– Telephone station sets were always relatively
small, quiet when idle, and require only the ability
to speak and hear understand the language of the
other person.
• Special teletypewriters are available for deaf or hard of
hearing telephone users.
Page 15
© 1996-2006 R.Levine
How Does Customer Perceive Acceptable Price vs. Performance?
• In some cases, the technical performance is
adequate, but end users perceive the price as
excessive and won’t buy the product.
• This non-technical aspect of product development is
supposed to be addressed by customer surveys,
focus groups, etc., but sometimes they predict
incorrectly.
• Telecom items perceived as overpriced:
– Iridium originally charged $3000 for a handset and $3 per
minute air time
– Scheduled airline in-flight telephones. Those systems still
charged at least $2 or more per minute,due to high operating
costs.
• Most end users (apparently) won’t pay over about
$0.10 to $0.20/minute for “air time”
Page 16
© 1996-2006 R.Levine
Best to Understand Technology Yourself
• Make well-founded decisions yourself
– Less dependence on the opinions of others
– Your instructor earns most of his income from being a
“technology expert” consultant, but would still rather have his
clients understand the technology themselves!
• Separate the wheat from the chaff when exaggerated
product claims are made
• Make realistic and profitable product and service
plans
– Do customers exist for this product or service?
– Are they willing to pay a compensatory price for the product
at projected costs?
– Why is the product competitively advantageous? What are
the competitive products or services?
Page 17
© 1996-2006 R.Levine
Now: Telecom Technology
• Having said enough for now about the reasons and
motivations for telecom products, we turn to the
technology of telecommunication.
• There are two ways to convey information:
– Send a physical object. Historically, the customary object is a
letter (e.g. on papyrus, parchment and later on paper).
– Send some “energy” in the form of an electromagnetic wave.
In ancient times, light was involved in viewing signals or
semaphore signals at a distance. Privacy and data rate
improvements had to wait for the discovery and a minimal
understanding of electricity.
Page 18
© 1996-2006 R.Levine
Historical Overview: Telegraph
• Invented by Samuel F.B.Morse (an artist, not a scientist)
greatly assisted by Alfred Vail* .
– Inter-city telegraph demonstrated by Morse in 1837.
– Several less practical European telegraph systems preceded
Morse
• For example, Morse (and others) thought that electrical
signals travelled “instantaneously” from telegraph key to
the sounder (receiver), since the complete theory of
electromagnetic waves was not formulated until 1860-90
by J.C. Maxwell, O. Heaviside, et al.
* Coincidentally, a relative of Theodore Vail, president of AT&T about 60
years later
Page 19
© 1996-2006 R.Levine
Telegraph Main Features
• Current flow around a circuit including a
battery, telegraph key (on-off switch), a single
wire (typically iron, later copper) with the
earth as a return path.
• Worked adequately up to about 30 miles,
depending on earth conductivity.
• About 1849 the repeater allowed longer links
by chaining 30 mi sections via an electromechanical “relay” (switching contacts
operated by an electromagnet).
Page 20
© 1996-2006 R.Levine
Telecom Overview: Telephone
The telephone was invented in 1876 by Alexander G. Bell
(a speech teacher, not a scientist). Born in Scotland, Bell
immigrated to Canada and then the USA.
• The telephone had the significant advantage that no
special skill (such as learning Morse code) was required
to use it!
– Ease or convenience of use is often a deciding factor
in the success of one technology over another.
• Bell’s microphone (called “transmitter”) produced
electric current proportional to instantaneous air
pressure. Earphone (“receiver”) reversed the
process, converting the electrical waveform back
into acoustic (sound) form.
Page 21
© 1996-2006 R.Levine
Some Business History
• Bell was financed by his wealthy industrialist father-inlaw, Gardiner G. Hubbard, a man with a history of
business and legal contention with the (then) large
Western Union Telegraph Company
• Bell’s original objective was to send several
independent telegraph signals over the same circuit
– Today we would describe his plan as frequency division
multiplexing (FDM) of amplitude modulated Morse code.
• He discovered by accident that his equipment could
transmit speech
– He added a new claim to his already filed patent covering this
– When the telephone became commercially important, major
patent litigation followed, ultimately decided by the US
Supreme Court
Page 22
© 1996-2006 R.Levine
Business Conflicts
• Bell and Hubbard offered the patent to Western Union
(WU) at first for $100,000
– This was an immense sum* in 1876, when a large house cost
less than $1000.
– WU turned them down, due to dislike of Hubbard
• A famous negative evaluation letter (probably not authentic) is
available on this course web site.
• The letter also is a prime example of
– “Not Invented Here” (NIH) attitude, ignoring good outside ideas
– Lack of proper appreciation of the advantages of the invention
– Inability to accurately foresee that improvements are possible to
overcome the initial disadvantages of the invention
*The Internet web site http://eh.net/ehresources/howmuch/dollarq.php
that contains a history of US dollar inflation, indicates that $100,000
in 1876 had the purchasing power of $1,705,922.48 in the year
2005.
Page 23
© 1996-2006 R.Levine
Early Competitive Moves
• WU, after recognizing the fast growth of the telephone,
quickly decided to get back into competition
• They hired the best available inventor, Thomas A.
Edison, to invent a significantly improved microphone
circa 1878
– Edison studied the telephone, found its most important
weakness, and came up with a solution based on Bell’s “liquid
transmitter.” Bell’s liquid transmitter was a variable resistance
microphone used in his first working voice transmission, but it
was impractical because it used an acid-water solution as the
variable resistance material. Edison substituted a sealed
capsule of powdered carbon as the pressure sensitive variable
resistance element. This “carbon microphone” invention was
also later improved by German-American Emil Berliner as well.
Page 24
© 1996-2006 R.Levine
Business Strategies
• The improvement in audio loudness (permitting longer
telephone wires and thus more wire coverage area per
central office) gave the carbon microphone a strong
economic competitive advantage.
• But WU could not operate a telephone system without
infringing the basic Bell patent, either.
– Negotiations were stalled, until the Bell company suggested
something which would be illegal under present anti-trust law…
• WU agreed in 1879 to stay out of the telephone
business for 20 years in return for 1% of the income
from the telephone.
– The telephone industry grew so fast that Bell was soon able to
buy most of WU shares. WU became a subsidiary of Bell from
circa 1900 until divested in a famous 1914 anti-trust case.
Page 25
© 1996-2006 R.Levine
Early 20th Century
• American Telephone & Telegraph (the renamed Bell Telephone
company), was headed for many years by Theodore Vail,
coincidentally a nephew of Alfred Vail, Morse’s collaborator.
• Vail vigorously bought out other telephone operating companies in
most major cities, leaving only rural areas to the independents
(formed after the Bell patents expired). This acquisition stopped in
1914*.
• AT&T purchased Western Electric Co. (electric equipment
manufacturer originally so named to save the cost of repainting the
entire sign in a former Western Union repair shop), “vertically
integrating” manufacturing and telephone operations
• AT&T established its Long Lines division, providing long distance
connection between all North American and foreign cities.
*In 1914 AT&T’s negotiator made the “Kingsbury commitment” to not buy out any more
independent telephone companies, thus settling a federal antitrust lawsuit.
Page 26
© 1996-2006 R.Levine
Some Technological Transmission Advances
• Single wire with earth return was replaced in 1890s by a subscriber “loop” of
current carrying copper wire.
– An innovation by J.J.Carty, who became head of AT&T R&D and ultimately
established Bell Telephone Laboratories.
Some AT&T accomplishments during the first half of the 20th century:
• DeForest’s “Audion” triode vacuum tube amplifier was improved and
adapted for analog voice frequency amplification, leading to coast to coast
long distance telephone connections.
• Gilbert S. Vernam invented the Vernam Cipher cryptography method for
teletypewriters during WW 1
• The quality and noise of analog telephone connections were improved in
1920s by H.S.Black’s invention of “negative feedback” at Bell Laboratories.
• Frequency Division Multiplexing (FDM) using single side band (SSB)
modulation was developed by John R.Carson at Bell Laboratories. Basis of
Analog telephone multiplexing.
• Microwave co-ax cable was developed by Lloyd Espenscheid of Bell Labs.
Used today for T-3 and other signals.
Page 27
© 1996-2006 R.Levine
More Business Developments
• The Anti-trust Division of the US Justice Department investigated AT&T in
1914, 1937, 1948, 1965, 1972. Each investigation led to consensual
settlements which further restricted the scope of AT&Ts business.
– 1914: “Kingsbury Commitment” stopped acquisition of independent telephone
companies and divested WU from AT&T
– 1937: AT&T divested non-telephone businesses (appliances, motion picture sound
systems, etc.) and offshore manufacturing.
• ITT (originally International Telephone and Telegraph Corp.) was founded by brothers
Sosthenes and Hernand Behn. They were sugar brokers in Porto Rico who first bought the
Puerto Rico telephone company. Then they founded Cia. Telefonica Espana in 1923. In
1937, using J.P. Morgan funds, they bought all off-shore Western Electric factories. They
later founded other telephone companies in Latin America. ITT sold its telephone
manufacturing businesses in 1990s to Alcatel, and now owns hotels, insurance companies
and some non-telephone manufacturing firms.
– 1948: AT&T agreed to license all patents to competitors
– 1969: AT&T agreed to allow connection of customer-owned equipment (result of
FCC and court CarterPhone decision) rather than renting. AT&T had previously
always rented equipment to the subscriber, a method learned from the United Shoe
Machinery company in the early Boston years.
– 1984: AT&T divested local telcos (RBOCs) but retained long distance and
manufacturing. (Manufacturing later separated under the Lucent name.)
Page 28
© 1996-2006 R.Levine
Other Business Events
• AT&T, until 1984 divestiture, received 1% of gross income
of all RBOCs
– Also was part owner of Bell Canada and Northern Electric, its
manufacturing subsidiary, until 1970s. This became Nortel
Networks, no longer owned by AT&T.
– Extensive cross-licensing of patents with other major
telephone equipment manufacturers in other countries as
well.
• Example: Crossbar telephone switch was developed under
cross-licensing agreement with L.M.Ericsson of Sweden
• AT&T acquired NCR (formerly National Cash Register) in
1989, then spun it off as part of the 1996 separation into
three businesses. Lucent (with Bell Laboratories) is only a
manufacturer and recently merged with Alcatel. AT&T is
today an operating company in long distance. Its
cellular/PCS activity is a separate corporation, now merged
with Cingular Wireless.
– Both AT&T and Lucent have started several spin-offs also
Page 29
© 1996-2006 R.Levine
Some Major Telecom Vendors
with Dallas-Ft.Worth Presence
• Alcatel (France) acquired most ITT manufacturing operations and
Rockwell (Collins) telecom products, and Digital Switch Corp. (DSC),
and integrated them with its existing products in the 1980-90s.
• Ericsson (Sweden), another long term telecom manufacturer
worldwide, has operations here.
• Fujitsu, NEC (Nippon Electric Co.) are two separate independent
Japanese telecom manufacturers with Dallas area operations
• Motorola, primarily in Fort Worth (cellular and paging equipment)
• Nokia (Finland), strong in cellular/PCS handsets but also makes
cellular infrastructure and landline telecom switchgear
• Nortel Networks (formerly Northern Telecom) is a descendant of
Northern Electric of Canada.
• Siemens (Germany) a long term telecom and general electrical
equipment maker, now reducing its presence in telecom.
This area is sometimes called “Telecom Corridor™” or “Switch Alley”
Page 30
© 1996-2006 R.Levine
Some Telephone Operating Companies
• Originally 7 (now 4) Regional RBOCs, with
consolidation of SWBell-PacTel-SNET and NYNEXGTE-Bell Atlantic (now Verizon), etc.
• GTE, arising from mid-century consolidation of many
independent local telcos, merged with Bell Atlantic and
Primeco wireless to form Verizon (rhymes with horizon)
in 2000
• Scattered remaining independents in some smaller
cities (e.g. Rochester NY, etc.)
• Numerous Inter-Exchange Carriers (IXCs) the largest 3
being AT&T, MCI and Sprint.
• The government operates Post, Telephone and
Telegraph (PTT) administrations in many other
countries; but “privatization” is spreading rapidly
Page 31
© 1996-2006 R.Levine
Digital Telecom Revolution
• The T-1* digital multiplexing system, introduced by Bell Labs in
1961, ultimately led to an almost complete conversion of the North
American public switched telephone network (PSTN) to digital
transmission and (later) digital switching
• T-1 was a rare and uniquely successful product because it is:
– Immediately equal or lower in cost than the prior analog FDM
multiplexer. Cost improved more later with product evolution as well.
– Carefully designed to be backwards compatible with all switching and
prior art transmission equipment at connection interfaces
– Better signal quality than FDM multiplex
– More capacity (24 voice channels on the same wires that previously
carried only 12 channels)
*Also written T1. Since T-1 is a trade name, DS-1 is an approximately equivalent term used in
standards documents, etc.
Page 32
© 1996-2006 R.Levine
Is “Digital” Always Better?
• The “error” introduced by conversion from analog to digital
representation can be controlled and limited in advance by the
designer of the A/D converter. Called “quantizing” error.
• Digital representation of information does not suffer from cumulative
“noise” errors. Transmission over a longer distance only causes time
delay, not distortion.
– This is the result of a system design in which the two binary digital
voltage levels (typically 0 and 5 volts) differ by much more than the
typical “noise” voltage level (typically 0.001 volts).
• But… digital representation typically uses more (radio) bandwidth
than analog representations.
– This problem can be reduced by use of data compression coding in
some cases.
– When the channel is extremely “noisy,” (e.g., a cellular radio link) error
protection coding must be used, and this requires part of the total
channel bit rate to be devoted to bits for this purpose. Cellular radio
systems typically devote half the physical bit rate capacity to error
protection.
Page 33
© 1996-2006 R.Levine
T-1 Benefited From Prior Technology
• PSTN voice signals were historically already low-pass audiofrequency filtered to attenuate audio power above approx. 3.5 kHz
audio frequency
– Necessary for FDM multiplexing and well-verified to support intelligible
conversation
– Permits accurate digital waveform coding at 8000 samples/second
• T-1 design was an early application for transistors
– Repeaters are installed at 6000 ft. intervals in outdoor or difficultaccess locations and must operate reliably and consume little power
– Vacuum tube devices would not be practical
• T-1 uses PCM (pulse code modulation) waveform coding with
logarithmic companding
– 8-bit binary coding of each waveform sample, with non-uniform
voltage steps, produces uniform signal to noise ratio over a wide
range of audio loudness
8 bit/sample • 8000 sample/sec = 64,000 bit/second = 64 kb/s
Page 34
© 1996-2006 R.Levine
Incorporation of Call-Processing Signals
• Two methods for signaling are in general North American use
1. “Robbed bit” signaling uses the least significant bit of the PCM in every
6th frame to convey supervision (channel busy/idle) status. Five of
every six consecutive waveform samples are not affected.
Systems for 12 and 24 multi-frame synchronizing patterns are used to
ensure that the signaling equipment uses the proper bit
• Robbed bit signaling leaves 56 kb/s (7 bits of every sample) for the
subscriber, even if not multi-frame synchronized
2. Common channel signaling uses a reserved digital channel (either 64 or
1536 kb/s in North America) to convey messages in packet data form
between switching systems regarding the call processing on numerous
other channels
Common channel signaling system Number 7 is today’s world-wide
standard, with some national variants There are many abbreviations:
(SS7, CCS7, etc.) In some cases, different abbreviations imply different
national variants of Common Channel Number 7.
Page 35
© 1996-2006 R.Levine
Further Digital Multiplexing
•
Higher level digital multiplexing systems were developed with better economy for high
traffic corridors:
– T-1 (DS-1): so called North American (and Japan) Primary Rate digital
multiplexing. 24 channels at 1.544 Mb/s
– T-1C: a double capacity system (48 channels) now rarely used. Not mentioned in
international standards. 3.152 Mb/s
– T-2 (DS-2): a quadruple capacity system (96 channels). Called M12 or
Secondary Rate. Combines 4 DS-1 tributaries. Seldom installed today. 6.312
Mb/s
– T-3 (DS-3): Combines 7 DS-2 tributaries. M13 multiplexers produce this Tertiary
level rate from 28 T-1 tributaries. 44.736 Mb/s. Uses co-axial cable or
microwaves.
– Different and mostly incompatible “T-4” higher level digital multiplexers using coaxial cable or microwaves were developed in different countries (US, Canada,
Japan) but were relatively little used since only a few routes have enough traffic
to make this economically feasible.
– European digital multiplexers of similar characteristics are widely used in other
countries.
Page 36
© 1996-2006 R.Levine
Higher Level Multiplexer Trends
• DS-1, DS-2, DS-3 multiplexers are designed to
accommodate small time-varying inaccuracy in the
bit rate of the incoming tributaries (plesiochronous
multiplexing)
• An undesirably large portion of the total bit rate (bit
“overhead”) is needed to handle this, and the
necessary process for de-multiplexing a single DS-1
or single 64 kb/s voice channel (DS-0) is complicated
and costly
• These difficulties, and the increased use of fiber
optic transmission and more accurate digital bit
stream synchronization, has led to development of
new and fundamentally improved multiplexing
designs
Page 37
© 1996-2006 R.Levine
EM Wave Transmission Media
• Radio transmission. Non-guided via open
space
– Inferior channel characteristics due to fading,
interference, etc.
– But portability makes cellular service valuable, and
absence of intermediate equipment between
microwave towers gives lower cost.
• Guided electromagnetic waves:
– Via twisted pair wires, co-axial cable. Typically
using “repeaters” to compensate for signal loss
– Via optical fiber, in the infra-red optical frequency
range. Electro-optical or all-optical signal amplifiers
are used to compensate for losses.
Page 38
© 1996-2006 R.Levine
SONET and SDH
• A multiplexing format normally used on optical fiber, but
lowest bit rate members of the family can be transmitted
via co-axial cable or microwave radio.
– SONET (Synchronous Optical Network) in North America
– SDH (Synchronous Digital Hierarchy) elsewhere
• In a refreshing departure from previous international
incompatibility, these standards are virtually identical.
SONET includes a lowest bit rate version at 51.84 Mb/s
which is not used in SDH, but higher rates such as
155.52 Mb/s etc. are common to both standards and are
compatible when similarly configured.
Page 39
© 1996-2006 R.Levine
Digital Transmission and Switching
• The rapid growth of digital multiplexing transmission
systems (almost 100% of the North American network
today) led to a parallel development of digital local and
long distance switches. These switches are more
compact, use less power, and are more reliable than
their electro-mechanical predecessors, and mostly
contain automatic self-test equipment to permit
efficient use of fewer repair personnel
• Digital Switching is now included in the present course
EETS8320.
Page 40
© 1996-2006 R.Levine
Digital Switch Basics
• End office digital switches typically support traditional
analog telephone sets, and in some cases ISDN or
proprietary digital telephone sets. The analog voice
waveform voltage is periodically measured (“sampled”)
and each voltage is converted via analog/digital
converters and digitally coded into a bit stream.
• From trunk connections, separate channels of digital
information are separated from the bit streams.
• Digital channel data is stored temporarily (typically for
125 microseconds) in a local memory in the switch.
• Desired outgoing channel bit streams are multiplexed
together to connect to other switches.
• Microprocessor internal to the switch controls routing of
connections.
© 1996-2006 R.Levine
Page 41
Switch Types
• End-office switch: both trunks and telephone sets
– Formerly called Class 5
• Trunk-trunk switch (no telephone sets)
– Used to complete long distance connections between end
switches
– Used (with radio base stations) for cellular radio systems
– Formerly designated as Class1 to Class 4 based on details
of application in the network.
• Private Branch Exchange (PBX) switches, used
primarily for business users to establish both external
and internal calls
• “Intercom” switch. Connects telephone sets or
“hands-free” stations for internal calls only. Rare
today.
Page 42
© 1996-2006 R.Levine
Switch Features
• All digital switches are microprocessor
controlled and have many features. Some
examples:
– Call waiting (signal during a conversation that
another caller is attempting to reach you, and
ability to answer that caller)
– Incoming call forwarding to another number when
desired
– 3-way conference calling via a conference bridge
• PBXs in particular have a large repertoire of
sophisticated features.
Page 43
© 1996-2006 R.Levine
Digital Speech Coding
• A technical “race” has continued for the last quarter
century between speech coding technology and
transmission technology
– Lower bit rate speech coders are typically more complex
devices, but they allow carrying more conversations in a
transmission medium with a fixed total bit rate
– Innovations such as fiber optic transmission and integrated
circuits have reduced the cost of high transmission bit rates
• The public telephone industry almost changed over to
32 kb/s ADPCM* speech coding in the early 1980s, but
the lower cost of fiber stopped this plan
• Radio systems such as cellular and PCS appear to be
the main present use for lower bit rate speech coders.
* Adaptive Differential PCM
Page 44
© 1996-2006 R.Levine
Other Speech Coding
• Digital speech coding methods generally fall into one of
two categories:
1. Waveform coding. Examples include:
– PCM (Mu-law and A-law pulse code modulation: used in DS-1 and
E-1)
– ADPCM (adaptive differential PCM - typically 32 kb/s)
– Delta Modulation (DM) and CVSD (continuously variable-slope DM)
2. Audio Power Spectrum Coding. Examples include:
–
–
–
–
Sub-band coding
RELP (regular - or residual - pulse excited linear predictive coding)
CELP (code excited linear predictive)
VSELP, ACELP (vector sum ELP, Algebraic code ELP)
Page 45
© 1996-2006 R.Levine
Some Speech Coder Bit-rates
Typical Applications
Type
Bit-rate
Applications
Linear Binary PCM
640 kb/s
Compact Disk Music
Mu-law, A- law PCM
64 kb/s
Telephone
ADPCM
32 kb/s to 16 kb/s
DM, CVSD
32 kb/s
Telephone; digital
PCS*
Military digital
communication
Sub-band
32-16 kb/s or lower
xELP family
16- 4 kb/s
Some land mobil e and
satellites
digital cellular and
PCS, voice over
Internet
Lower bit rate coders are generally less satisfactory
than higher bit rates.
*PCS= Personal Communications Service, a cellular radio system usually with
digital speech coding
Page 46
© 1996-2006 R.Levine
Non-voice “Bearer” Services
• Due to their near-ubiquitous presence, readily
available investment capital, and the franchise held
by many telephone operating companies to install
wire, cable or fiber, many other services are also
under development and use in the telephone system
and related systems
• Images: telefax, video, other images
• Data: Internet access, data bases, and related
information
• Digital coding of any originally analog information
(such as video) is seen as the optimum method for
combined transmission
– but verify that the entire system is really advantageous!!
Page 47
© 1996-2006 R.Levine
Telephone Data Modems*
• Digital data can be transmitted via telephone voice channels
using an audio frequency carrier signal which is modulated to
convey binary information by changing its:
– Amplitude (instantaneous voltage or power level). This method is
used alone only for Morse Code
– Phase (relative time delay of oscillatory waveform peaks and
valleys vis-à-vis a standard “clock” signal)
– Frequency (the quantity of cycles per second; the musical pitch)
• Recent modem (modulator-demodulator) designs mostly use
QAM (quadrature amplitude modulation) a combination of
amplitude and phase modulation
• V.90 or V.92: In one direction, various voltage amplitude levels
are each used to represent a specific 7-bit binary data value.
• ADSL: A special type of multi-carrier QAM modem is used via
telephone subscriber wires to carry high bit-rate digital Internet
signals in a frequency band above the usual voice frequencies.
* Modem is an invented word made of the first syllables taken from the two words
Modulator and DEModulator.
© 1996-2006 R.Levine
Page 48
Modem Properties
• Data modems today also include automatic equalizers to
compensate for individual voice channel characteristics that would
otherwise cause undesired waveform changes.
• Data rates of up to 9.6, 14.4, 28.8 and 33.6 kb/s are feasible using
classic adaptive QAM modem technology
• Higher bit rates up to 56 kb/s* use direct PCM encoding at one
end
– Fully digital connection at transmitting end. Analog connection at
receiving end. Signal voltage can be measured with sufficient
accuracy at receiving end to infer the PCM code value used.
– Full 64 kb/s throughput requires a specifically installed digital line
such as ISDN or DDS.
* V.90 and V.92 modems today are legally limited to 53 kb/s. The highest voltage
levels of PCM are prohibited to avoid “crosstalk” with other wire pairs in the same
cables.
Page 49
© 1996-2006 R.Levine
Fully Digital Telephone Services
• ISDN (integrated services digital network) and
proprietary digital services (DDS, etc.)
• Special digital signals used on the subscriber loop
• Permits end-to-end 64 or 56 kb/s digital service
• For voice, analog-digital conversion is performed in
the ISDN telephone set rather than in the central
office switch
• Unfortunately ISDN is very costly, but has had a
recent small surge in utilization due to Internet
access applications. Some critics view ISDN as an
early example of the “Iridium syndrome”
• Emergence of 56 kb/s V.90/92 modems has
severely reduced the use of ISDN
Page 50
© 1996-2006 R.Levine
Packet Data Systems
• In several types of data networks, data is transmitted in “packet”
format
– A small block of consecutive data bits from each particular source has a
“header” pre-pended. The header contains, among other things, a code
number indicating the destination. This is used to control routing.
– Typically an error-detecting code is appended to the end of the packet.
• In some systems, all packets are the same “size” (length) – in others
each packet is of different size, typically based on source data rate.
• Packets from different sources are transmitted via the same
channel, one after another
• Most systems use a special “flag” bit pattern, 01111110, as a
separator between packets.
– The internal packet bit stream is pre-modified (bit stuffing) to exclude
any false occurrences of the”flag pattern.
– At the receiving end, the bit stuffing process is undone
Page 51
© 1996-2006 R.Levine
Why Packets?
• Many types of digital information sources are “bursty”
in time
– Brief “bursts”of high bit rate data are separated by some time
intervals during which no data bits are generated
– Data coding methods which remove redundant information
from “raw” speech or video typically produce bursty data
• A number of different packet transmissions can be
multiplexed on a shared channel in a high bit rate
medium (co-ax, fiber, etc.) more efficiently than using
a separate channel for each source, provided that all
data sources do not continually produce data bursts
simultaneously
Page 52
© 1996-2006 R.Levine
ATM (Asynchronous Transfer Mode)
• ATM “Payload” data is transmitted in fixed size
packets (called here “cells”) of 48 bytes (384 bits)
with a 5 byte identification header (53 bytes total)
• ATM signals can be transmitted e.g. via the
“payload” of SONET/SDH at 50 Mb/s or more
gross bit rate
• Due to its small packet size, ATM has little signal
delay, and is theoretically superior to other packet
formats for digitally coded voice.
• ATM is an interesting alternative to LAN/WAN
technologies such as Ethernet, although
presently far more costly
Page 53
© 1996-2006 R.Levine
Telefax (Facsimile,FAX)
• Groups 1 and 2 FAX are obsolescent.
• Group 3 FAX is in worldwide use. A page image is typically
transmitted in less than a minute at 9.6 kb/s binary data
rate via internal modem over a voice grade PSTN channel.
• Group 3 FAX uses binary data compression coding of
black/white pixel (or pel) picture elements (“dots”)
– Line difference coding takes advantage of vertical “lines” in the
image
– Run-length coding takes advantage of large contiguous areas of
white or black, and of long run zero line differences produced by
line difference coding.
– Huffman coding takes advantage of repeated appearance of certain
binary bit patterns in the FAX bit stream
Page 54
© 1996-2006 R.Levine
Other Data Compression Methods
• Lossless data compression methods exploit redundant
data bit patterns when present, and accurately
regenerate original data when decoded
– Plain language text has well-known frequently occurring
characters (E T A O I N etc.) and infrequently occurring
characters (J Z Q etc.), a fact that is exploited by Morse code
and Huffman coding
– LZW (Lempel-Ziv-Welch) coding dynamically adjusts
transmission codes to use short binary patterns for frequent
symbols and longer binary patterns for infrequent symbols. LZW
is one type of “algebraic” coding.
– Huffman coding is a non-dynamic formal lossless data
compression method similar to LZW
Page 55
© 1996-2006 R.Levine
Lossy Coding
• The word “lossy” implies data compression with imperfect
reconstruction of the original information
– Used when human perception can be “fooled,” for:
• Video and still pictures with continuous brightness range (gray or
color)
• Audio spectrum (speech) coding. Does not reproduce the exact
sound waveform
• Lossy image coding typically approximates the spatial
brightness pattern using a family of orthogonal functions
– Discrete Cosine Transform is popular for images, video
• Frame difference and motion extrapolation methods are
used with video as well
• Video, which requires over 40 Mb/s with simple waveform
coding, can be lossy-encoded at 64 to 128 kb/s (via Px64
code) with acceptable (not high) quality and coding delays
Page 56
© 1996-2006 R.Levine
Error Protection Coding
• Use of additional bits with the “payload” data can be
used to
– Correct a limited quantity of bit errors
– Detect (but not correct) larger quantity of bit errors
• Error detection codes are often used in conjunction with an
Automatic request to Retransmit (ARQ) strategy to retransmit
pieces of the data (typically packets) unless they are soon
acknowledged as received OK, via a message sent back to
transmitter from the receiving end equipment.
• Widely used with error-prone radio channels and delay-tolerant
signals such as for wireless call processing messages
• Also used in T-1 extended super frame version, and
SONET/SDH multiplexing systems, to continuously
monitor transmission accuracy.
Page 57
© 1996-2006 R.Levine
Encryption
• Important when the transmission is physically open to
interception by unauthorized persons
– Particularly for wireless, radio, microwave, etc.
• Encryption methods can also be used to authenticate messages
– Only a person who knows the correct “secret key” can properly
encrypt or decrypt a message
• The most widely used physical level encryption method is the
Vernam cipher
– A “secret” encryption cipher bit stream is “added” to the bits before
transmission, then the same cipher bit stream is “subtracted” out at
the receiver*
– The practical complications in this process relate to generating,
distributing and synchronizing the cipher bit streams at both the
transmit and receive ends.
*Not normal arithmetic addition and subtraction. Rather the XOR, or ring sum or modulo-2
logical operation
Page 58
© 1996-2006 R.Levine
End of Lecture 1
• The rest of the sessions involve a more
detailed description of the technical
topics just listed.
Page 59
© 1996-2006 R.Levine