|
Items in cart [1] |
Questions? Call 888-624-8373 |
| Copyright © 2009. National Academy of Sciences. All rights reserved. Terms of Use and Privacy Statement |
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 447
Page 447
53
Technology in the Local Network
J.C. Redmond, C.D. Decker, and W.G.
Griffin
GTE Laboratories Inc.
The local telephone network is that part of the telephone
network that connects individual subscribers, homes, businesses,
and so on, to an end-office switching center and includes the end
office itself. In its early embodiment, the local network was
simply a telephone (or other instrument) at a subscriber's
residence or place of business connected to a pair of wires that
led to a switching office (Figure 1). At the switching office,
connections were made between local users, or the signal was sent
via a tandem or long-distance path for subsequent connection
through another part of the telephone network.
The early design goals placed on the local network were
relatively simple (at least from our perspective today): to provide
reliable transmission and switching of voice signals that could be
easily understood at the receiving end. There were other
considerations, such as ringing the phone, that were also
necessary, but, mainly, the subscribers just wanted to hear
intelligible voices.
In concept, the local network is not much different today than
it was in the past except that the termination at the subscriber's
premises is made at a standard interface that does not include the
customer's on-premises wiring or telephone (or other equipment). Of
course, things are much more complex now, because the demands for
added bandwidth, new services, and overall cost efficiency have
greatly changed the design goals used to plan and implement the
network.
The local telephone network is evolving rapidly from its
historical manifestation as a narrowband connection of physical
addresses to a more complex network of networks that includes
narrowband, broadband, and variable-band transmissions to physical
and logical addresses. The added capabilities and increased
efficiency of the telecommunications network have allowed the
introduction of new data services such as frame relay and switched
multimegabit data service; developed new, intelligent features
across a broad spectrum of users; and positioned the network for
significant growth in the future.
History of the Local Network
At the time the telephone was introduced, the telegraph was
regarded as far more important to commerce. The product of the
telegraph was a written record of the communicated message. This
tangible record provided a link to the familiar handwritten
discourse of the commerce of the day. Because it did not provide a
record of the message, the telephone was initially regarded as a
novelty.
However, as the need for communications increased, the telephone
soon surpassed the telegraph as the medium of choice. In fact,
having a telephone became so popular that the proliferation of the
supporting wires became objectionable. Engineers were forced to
find a more compact means of running the wires from point to point.
The engineers found that wrapping the copper pairs with paper
insulation and encasing the resulting bundles of pairs in lead
allowed a more compact transmission medium, called a cable.
It was also obvious that it was impractical for users to string
a pair of wires from their location to each person to be called.
The solution was to run all of the pairs of wires from the
customer's premises to a central location. There, ''operators"
could connect sets of wires together according to instructions from
the customers. This created a center where customers were switched
at the central point of the wires. This was the origination of the
old-timer's call, "Hello, Central!"
OCR for page 448
Page 448
Figure 1
Traditional access to customers.
The switching at Central was originally accomplished by human
operators who manually interconnected the customers' calls. The
transmission medium was copper wires either cabled or open. Control
was provided by the customers' verbal instructions and the
operators' manual actions. This division in function (i.e.,
transmission, switching, control, and terminal equipment) still
exists in today's telecommunications networks, albeit in radically
different forms.
Transmission Media and
Multiplexing
Transmission equipment for telephony has evolved from simple
open wire carrying a single conversation to optical fibers that can
carry many thousands of conversations.
Signals from a home or business are carried over a twisted pair
of copper wires, called the local loop, to a centrally located
local office. Hundreds and even thousands of wire pairs are carried
together in a single large cable, either buried underground in a
conduit or fastened above ground to telephone poles. At the central
office, each pair of wires is connected to the local switching
machine. The transmission quality of the early installations was
highly variable. Today, however, the plant that is being installed
has the capability to transmit at least basic rate ISDN (144 kbps).
This is true even for long loops (greater than 12,000 feet) that
require loop extension equipment or lower-resistance wire. (Future
plans will reduce the number of those long loops.)
In order to reduce costs, methods were developed to combine
(multiplex) a number of subscribers on a single transmission medium
from the central office, with the individual wire pairs split off
nearer to the subscribers (Figure 2). As advances in technology
progressed, multiplexing kept pace by increasing the number of
conversations carried over a single path. Only a few years ago,
multiplexing provided tens of conversation paths over a pair of
wires. Initially this was accomplished by shifting each telephone
signal to its own unique frequency band.
A major advance in multiplexing was accomplished when normal
voice signals were converted into a coded digital form. In this
form, the digital signals could be regenerated repeatedly without
loss of the voice or other information content.
With today's time division multiplexing, each telephone signal
is converted to a digital representation. That representation is
inserted into fixed time slots in a stream of bits carrying many
digitized telephone signals, with the overall stream operating at a
high bit-rate. (An uncompressed voice signal requires 64,000 bits
per second [bps] in digital form.)
The multiplexed signals can be transmitted over a variety of
transmission media. The most common multiplexing system, called T1,
operates over two pairs of copper wires carrying 24 telephone
signals, at an overall bit-rate of 1.544 million bits per second
(Mbps). First installed in 1962, the system is still widely used
today.
With optical fiber, a beam of light is transmitted through a
very thin, highly pure glass fiber. The light travels in parallel
rays along the axis of the fiber. Many telephone signals are
multiplexed together, and the light
OCR for page 449
Page 449
Figure 2
Using pair gain for access to customers.
source is simply tuned on and off to encode the ones and zeroes
of the digital signal. A single strand of optical fiber used in
today's telecommunication systems can carry 30,000 telephone
signals. Also, the repeaters in an optical fiber can be separated
much farther (tens of miles as opposed to 1,000 feet) than in an
electrical system (see Figure 2).
The greatly increased capacity of fiber links at relatively low
cost has led to the practicality of "bypass," in which high usage
subscribers bypass the local provider and feed directly into the
telecom network (Figure 3).
The history of transmission media and multiplexing shows an ever
increasing progression of the total number of conversations that
can be carried over a specific generation of the technology. The
more conversations carried, the lower the cost per call, or the
greater the bandwidth available per subscriber.
Switching Equipment
The telephone network is a switched network. The connection from
one telephone to another is created and maintained only for the
duration of each individual telephone call. In the early days,
switching was performed manually, by operators who used cords to
connect one telephone line to another. The automation of switching
was first accomplished by allowing customers to directly control
electromechanical relays and switches through a "dial" attached to
the telephone instrument. Electromechanical switching equipment
reduced the need for human operators. However, the equipment's
capacity and capability for supporting new features was
limited.
Most of today's switching machines switch signals that are in
digital format. Digital switching interfaces well with the
time-division-multiplex technology of today's transmission
systems.
As technology has advanced, it has blurred some of our old
categorizations in the local networks. We now have remote units in
the feeder network of various sizes and capabilities that combine
the transmission, multiplexing, and switching roles previously
accomplished by discrete systems. Initially, these changes were
done to reduce costs, but now this added complexity has given the
networks much greater flexibility to grow and expand in
capability.
Signaling
The telephone system uses a myriad of control signals, some of
which are obvious to the customer and others of which are unknown
to him or her. The early signals between the customer and Central
were a crank on the magneto to summon the other end. Instructions
on whom to call were exchanged by verbal commands.
Today's customers are very familiar with the telephone's ring,
dial tone, dialing keypad tones, and busy tone. However, control
signals that are sent between switching offices, over circuits
separated from the voice
OCR for page 450
Page 450
Figure 3
Local access and transport area.
channel, do not directly involve the customer's attention. These
separated control signals used within this separate network are
called "common channel signaling system number seven," or SS7.
Although the initial motivation in the introduction of common
channel signaling was an improvement in call set-up, this change
has supported the movement of network intelligence out of
proprietary switching systems and into a set of distributed
processors, databases, and resource platforms connected through
well-defined industry standards.
Thus, we have seen the implementation of the intelligent network
in which intelligence is added to the network to implement new
features and services such as personal number usage, virtual PBXs,
voice response
OCR for page 451
Page 451
services, and others (see Figure 4). Further, this intelligence
in the network has the potential to evolve to be the service
infrastructure for future broadband and multimedia applications
such as interactive video games, remote learning, and others.
Telephones and Other Station
Apparatus
The telephone itself is a rather simple appliance. A microphone
(the transmitter) and an earphone (the receiver) are contained in
the handset. The modern keypad dialer sends unique combinations of
two single-frequency tones to the central office to indicate the
particular digits dialed.
Newer instruments, especially personal computers, are now common
to the network. Their capabilities include simultaneous voice and
picture communications and computers with telephone capabilities,
and they will include features yet to be invented.
Figure 4
The advanced intelligent network adds significant capability to the
network by using intelligence
at a service control point or intelligent peripheral to provide new
features and services.
Future Evolution of the Local
Network
There is currently a revolution under way in the local telephone
network that is being brought about by changes in the technical,
competitive, and regulatory arenas. From the technical perspective,
there has been a great advance in the ability of networks generally
to handle and process information. More precisely, there has been a
digital revolution brought about by the ever increasing power and
ever decreasing cost of microelectronics.
One of the real drivers in this revolution is electronics
technology, specifically progress in semiconductor fabrication
capabilities. Today's 0.5-µm feature size for complex
production integrated circuits is expected to
OCR for page 452
Page 452
shrink to 0.25 mm by 1999 and
0.18 mm by 2002. This trend will
allow microprocessor speeds to increase from today's approximately
200 MHz clock rates to 500 MHz over the same time frame (see Figure
5). In addition, DRAM (dynamic random access memory) chip capacity
will also increase from today's 16 Mb chips to an estimated 1024 Mb
capacity in 2002. Putting these advances in electronics to work in
development of RISC (reduced instruction set compiler) processors,
the heart of desktop workstations or set-top boxes, for example,
will mean that these processors can be expected to perform
calculations such as those needed, for example, for video
compression ten times faster in the years after 2000 than they do
today.
While the text files transferred today between users are
normally about 100 kb, it is not uncommon for files with graphics
information to routinely exceed 1 Mb. It is now becoming common for
users to attempt to send 1 Mb files over the existing telephone
modem lines, with the result that the commonly used techniques are
seen to be quite inadequate. Thus, at least a factor-of-ten
improvement in available data rate is required.
The data services that users will soon demand certainly exceed
the capability of the existing telecommunications network. As
traffic begins to include full-screen, high-resolution color
images, files will become of the order of 1 Gb in size, dictating a
further increase in capacity of three orders of magnitude. This
sort of increase in capability will require some fundamental
changes. Growth will be required not only in the pipelines that
provide the data but also in file server technology, network
management infrastructure, and user software to enable rich new
services.
Though we are dealing here explicitly with the local telephone
network, the impact of this revolution also affects local networks
of all kinds, including telephone networks, CATV networks, private
data networks, cellular radio networks, and so on—a profound
technological convergence.
Figure 5
RISC (reduced instruction set compiler) processor performance
trends. Data for this graph
were taken from manufacturers' specifications as well as from an
article by Ohr 1.
As an example of the data growth envisioned for the network, the
increase in DS-1 and DS-3 access lines is enlightening. Figure 6
shows this growth to the year 2003.
Key elements in this robust capability for digital processing
are that it is independent of content (such as voice or video) and
permits the distribution of the processing to the periphery of the
network, thus allowing computing power to migrate to the user.
These developments have enabled the conception of a whole range of
interactive services that meld and blend the areas of
communications, computers, and video. In fact, it is this
commonality of technology that has led to other implications in the
areas of competition and regulation.
OCR for page 453
Page 453
It is the desire to offer broadband video and interactive
services that has created the incentive for the local exchange
carriers to evolve their plants to provide bidirectional broadband
access. Actually, the build-out of the broadband network is a
process that has been going on for nearly two decades, beginning
with the first introduction of optical glass fiber for carrying
trunk traffic in the telephone company (telco) network around 1980.
In the intervening years, fiber has replaced virtually all the
metallic cable in the interoffice plant and has begun to migrate
into the feeder portion of the distribution plant.
The most costly, but at the same time the most restrictive,
portion of the access network, the subscriber loop, at this point
remains copper. This is key in considering evolution toward a
broadband infrastructure. With the digitization of the switching
infrastructure, the state of the current network includes a totally
fiber interoffice plant, a fully digital narrowband switching
infrastructure, and a partially fiber feeder plant.
The approach to a broadband network must be formulated from this
vantage point. There are two main technological thrusts that are
enabling the digital video revolution. The first is the ability to
compress digital video with high quality to the point where one can
deliver video streams in a cost-effective way to individual
subscribers (Figure 7). Even with the high capacity of optical
fiber, uncompressed digital video would have remained a challenge
in terms of transport, transmission, and switch capacity.
The other technical event contributing to the availability of
digital video has been the development of broadband switching
technology in the form of asynchronous transfer mode (ATM). Key
features of the development of ATM include the ability to multiplex
and to switch together (in a packet or cell format) the content of
mixed streams of multimedia traffic, and, what is more, to do this
isochronously, so that the time information of each stream retains
its integrity. It is anticipated that ATM switches will become
dominant after the turn of the century. One plan, shown in Figure
8, predicts 100 percent deployment by 2015.
The first plant upgrade enabled by the digital video revolution
has been the migration of the existing CATV network to a fiber-fed
technology where fiber is brought to within two or three radio
frequency (RF) amplifiers of the subscriber. The fiber-fed bus
architecture, or so-called hybrid fiber coaxial (HFC) system,
provides capability possible to provide approximately 80 analog and
150 digital channels over a 750-MHz HFC network. If such an
Figure 6
Growth of DS-1 and DS-3 lines. SOURCE: Reprinted from Ryan et al.
2.
OCR for page 454
Page 454
Figure 7
Compression of digital video.
HFC network serves areas of 500 homes, there are clearly enough
channels available with some statistical concentration to provide
selected channels for individual subscribers.
It is GTE's plan to offer 80 analog video channels in the
initial rollout of the HFC system to about 400,000 customers in
1995, followed by 500,000 more in 1996. The cost should be in the
range of $700 per customer. Subsequent upgrades will include adding
150 digital channels of broadcast MPEG in early 1996 at an added
cost of about $200 per customer (set-top box). This will allow
delivery of near-video-on-demand. In late 1996 or 1997, switched
MPEG will be added for video-on-demand and other interactive
services for a further incremental cost of $100 to $200 per
customer.
Further HFC additions beyond 1997 will depend on results
obtained with the initial system.
With a large number of channels available, even though the bus
architecture is a shared medium, it provides most of the
functionality of a switched star-star architecture that is typical
of most telephone networks. The enchanced upstream connectivity
allows the addition of voice and data as integrated services along
with video.
The approach of the local telephone carrier to bringing
broadband services to the loop involves the evolution to a
broadband distribution network, which includes fiber that will go
closer and closer to the subscriber and ultimately to the premises
itself. The particular approach to bringing fiber to the loop is a
function of cost. At the present time, fiber to the home is too
expensive a solution. Bringing fiber to some intermediate point is
the preferred option. The hybrid fiber coaxial system, while being
implemented initially by CATV operators, is clearly one such
approach.
While several local exchange carriers have embarked on network
rollout programs with hybrid fiber coaxial technologies for their
initial thrust into video distribution, it is clear that there is
no straightforward way to integrate HFC with the existing twisted
pair copper loop plant (e.g., power, ringing the telephone, etc.).
The additional costs of managing and maintaining two networks
appear to be a distinct disadvantage for this approach. In some
cases, where aging loop plants are in need of full replacement, HFC
can be put forward for integrated services and replacement of the
existing copper plant.
The challenge is to find a mode of migration that will provide a
sufficiently robust set of services to meet customers' needs in the
coming broadband environment while maintaining an effective and
elegant migration path from the existing copper plant. One such
approach is asymmetric digital subscriber line (ADSL) technology
(Figure 9). A pair of modem transceivers on each end of the copper
loop provides digital line coding for enhancing the bandwidth of
the existing twisted pair.
Figure 8
ATM switch deployment.
OCR for page 455
Page 455
Figure 9
Very high speed ADSL (VH-ADSL) to provide video to the home.
The approach is asymmetric because the capacity or bandwidth in
the downstream direction, toward the subscriber, is greater than
that in the upstream direction. One ADSL (very high speed ADSL, or
VH-ADSL) approach for particularly high bandwidth in the 25 to 50
Mbps range involves serving areas consistent with fiber being
brought into the loop within several thousand feet of the
subscriber, and is thus consistent with the migration path that
brings fiber close to the subscriber premises.
One of the transceivers is therefore installed at a fiber node.
Eventually this may go to a fiber-to-the-curb system and
ultimately, when economically justified, to fiber to the home.
There is a continuum of serving-area sizes. But for the present,
this is an integrated network that provides all services over a
single plant and is competitive for the range of bandwidths
required.
One of the major advantages of this ADSL approach is that it can
be applied only to those customers who want it and are willing to
pay for the added services. Thus, this method allows for an
incremental buildup of a broadband capability depending on market
penetration (Figure 10).
This, then, is the infrastructure we will be looking at, but
what of the services and programming? It is clear that broadcast
services need to be provided in an ongoing way. Additionally,
various forms of on-demand or customized video programming formats
are anticipated to be important. True video-on-demand commits a
port and content to an individual subscriber, and the viewer has
VCR-like control of the content. Near-video-on-demand shows a
program frequently enough to simulate the convenience of
video-on-demand, but without the robustness of true on-demand
services. Clearly, other services will be more akin to the
interactive features that have grown up on the personal computer
(PC) platform. These include various information and transactional
services, games, shopping, and, ultimately, video telephony.
The subscriber platform is also worthy of note. There are
clearly two converging sources of services here. One is the cable
television (CATV) environment, with the set-top box and television
set as the platform, and the second is the PC. While the former has
been almost exclusively associated with the domain of entertainment
services, information and transactional services have clearly been
the domain of the PC.
It is clear that the carrier must be prepared to provide
services that are consistent with both platforms, because one or
the other will likely continue to be favored for specific
applications or classes of applications. An example of such a
service is embodied in an offering called "Main Street." Main
Street uses the TV with a set-top box to provide a visual
presentation (currently stills and sound). A telephone line is used
to send signals to a
OCR for page 456
Page 456
Figure 10
Cost comparisons for very-high-rate ADSL, fiber to the curb, and
hybrid fiber coaxial cable versus market penetration.
Figure 11
Main Street.
OCR for page 457
Page 457
head-end server (Figure 11). Services offered include scholastic
aptitude test study, home shopping, access to encyclopedias, games,
etc. The service is currently offered at a few locations around the
United States.
These new networks of the future are much more complex and
require a corresponding increase in intelligence. Intelligence
(defined here as the ability to self-inventory, collect performance
data, self-diagnose, correct faults, respond to queries, etc.) no
longer resides exclusively in the central office but is spreading
into the local access network, thanks to the plummeting cost and
increasing reliability of processing power (Figure 12). The new
distributed intelligent network elements will enable a revolution
in the efficiency with which telcos can perform core business
functions needed to administer customer services. For example,
telcos have historically billed for service based on time of
connection, bandwidth, and distance. This approach has little
meaning in the case of connectionless data transmission, and new
approaches need to be formulated.
Similarly, monitoring and testing network performance will
attain new levels of efficiency as digital performance monitoring
and automatic alarm collection/correlation move down to the
individual line card level. Probably the most exciting aspect of
these new intelligent access elements is the new services they will
make cost effective. Reducing the cost of digital services such as
ISDN and frame relay, and of higher-bandwidth services such as
interactive multimedia, will require intelligent elements in the
access networks. Dynamic service provisioning, whereby services are
delivered in real time on an as-needed basis, will similarly rely
on intelligent network elements. As these examples show, the
addition of intelligence to the access network represents a new
paradigm for the rapid, efficient, and cost-effective addition of
new services.
Figure 12
Future network.
