Softswitch
`Architecture for VoIP
`
`Franklin D. Ohrtman,Jr.
`
`San Juan Seoul Singapore Sydney ‘Toronto
`
`McGraw-Hill
`New York Chicago San Francisco Lisbon
`London Madrid Mexico City Milan New Delhi
`
`5
`
`AT&T Exhibit 1018
`AT&T v. VoIP, IPR 2017-01384, Page 1
`
`|
`
`AT&T Exhibit 1018
`AT&T v. VoIP, IPR 2017-01384, Page 1
`
`

`

` Deiatnileee
`
`Cataloging-in-Publication Data is on file with the Library of Congress.
`
`Copyright © 2003 by The McGraw-Hill Companies,Inc. All rights reserved.
`Printed in the United States ofAmerica. Except as permitted under the United
`States Copyright Act of 1976, no part ofthis publication may be reproduced or
`distributed in any form or by any Means, or stored in a data base orretrieval
`system, without the prior written permission of the publisher.
`234567890 DOCMOC 0987654
`
`ISBN 0-07-140977-7
`The sponsoring editorfor this book was Marjorie Spencer and the productionsupervisor
`was Pamela A. Pelton. It was set in New Century Schoolbook by MacAllister Publishing
`Services, LLC.
`Printed and bound by RR Donnelley.
`
`McGraw-Hill books are available at special quantity discounts to use as premiums and
`sales promotions,or for use in corperste training programs. For more information,
`
`please write to the Director ofSpecial Seles, Professional Publishing, McGraw-Hill,
`Two Penn Plaza, New York, NY 10121-2298. Or contact your local bookstore.
`
`
`Information contained in this war& has been obtained by The McGraw-Hill
`Companies,Inc. (“McGraw-Hill”) Sam sources believed to be reliable. However,
`
`neither McGraw-Hill nor its
`aathers suarantee the accuracy or completeness of
`any information published b= and neither McGraw-Hill nor its authors
`shall be responsible for any ===> =issions, or damages arising out of use of
`
`this information. This work is
`paltitsed with the understanding that McGraw-
`
`
`Hill andits authors are supply=ag
`imfrmationbutare not attempting to render
`engineering or other professiamal services. If such services are required, the
`assistance of an appropriate peai=sssamalshould be sought.
`
`
`
`This book is printed on T=ape==_ S==ee paper containing a minimum of
`Teo
`a,
`©! 50 percent recycled deins== ===
`
`OOOOeee
`
`AT&T Exhibit 1018
`AT&T v. VoIP, IPR 2017-01384, Page 2
`
`AT&T Exhibit 1018
`AT&T v. VoIP, IPR 2017-01384, Page 2
`
`

`

`
`
`aie
`
`The Public
`witched
`Telephone
`Network
`
`ae
`ae
`
`Se
`anes
`
`feayar
`coe ee
`
`AT&T Exhibit 1018
`AT&T v. VoIP, IPR 2017-01384, Page 3
`
`

`

`
`
`An understanding of the workings of the Public Switched Telephone Net-
`work (PSTN)is best grasped by understanding its three major components:
`access, switching, and transport (see Figure 2-1), Each element has evolved
`over the 100-plus year history of the PSTN. Access pertains to how a user
`accesses the network. Switching refers to how a call is “switched”or routed
`through the network, and transport describes howacall travels oris “trans-
`ported”over the network.
`
`Chapter 2
`
`
`
`Access
`
`Access refers to how the user accesses the telephone network. For most
`users, access is gained to the network via a telephone handset. Transmis-
`sion and reception is via diaphragms where the mouthpiece converts theair
`pressure ofvoice into an analog electromagnetic wave for transmission to
`the switch. The earpiece performs this process in reverse. The most
`sophisticated aspect ofthe handsetis its Dual-Tone Multifrequency (DTMF)
`function, which signals the switch by tones. The handset is usually con-
`nected to the central office (where the switchis located) via copper wire
`known as twisted pair because,in most cases,it consists of a twisted pair of
`copper wire. The stretch of copper wire connects the telephone handset to
`the central office. Everything that runs between the subscriber and the cen-
`tral office is knownas outside plant. Telephone equipmentat the subscriber
`end is called customer premise equipment (CPE).
`
`
`
`AT&T Exhibit 1018
`AT&T v.VoIP, IPR 2017-01384, Page 4
`
`
`
`
`
`
`Figure 2-1
`
`
` Bae
`
`AT&T Exhibit 1018
`AT&T v. VoIP, IPR 2017-01384, Page 4
`
`

`

`The Public Switched Telephone Network (PSTN)
`
`Switching
`
`Figure 2-2
`The traditional
`
`relationship of
`Class 4, Class 5, and
`data networks
`
`Legacy Networks
`
`Web Sites
`
`The PSTN is a star network; that is, every subscriber is connected to
`another via at least one if not many hubs known asoffices. In thoseoffices
`are switches. Very simply, local offices are used for local service connections
`and tandem offices for long-distance service. Local offices, better known
`as central offices, use Class 5 switches, and tandem offices use Class
`4 switches. Figure 2-2 details the relationship between Class 4 and 5
`switches.A large city might have several central offices. Denver (population
`2 million), for example, is estimated to have almost 40 central offices. Cen-
`tral offices in a large city often take up muchofa city block and are recog-
`nizable as large brick buildings with no windows.
`Thefirst telephone switches were human.Taking a telephone handset off
`hook alerted a telephone operator of the caller’s intention to placeacall.
`The caller informed the operator of their intended called party and the
`operator set up the call by manually connecting the two parties.
`Mechanical switching is credited to Almon Stowger, an undertaker in
`Kansas City, Missouri, who realized he was losing business when families
`of the deceased picked up their telephone handset and simply asked the
`operator to connect them with “the undertaker.” The sole operator in this
`
`IP Circuits
`
`Class 4 Switch
`
`TDMCircuits
`
`AT&T Exhibit 1018
`AT&T v. VoIP, IPR 2017-01384, Page 5
`
`

`

`AT&T v. VoIP, IPR 2017-01384, Page 6
`
`town was engaged to an undertaker competing with Stowger. This compet-
`ing undertaker had promised to marry the operator once he had the finan-
`cial meansto do so. The operator, in turn, was more than willing to help him
`achieve that goal.
`Stowger, realizing he was losing business to his competitor due to the
`intercession of the telephone operator, proceeded to invent an electro-
`mechanical telephone handset and switch that enabledthecaller, by virtue
`of dialing the called party’s number, to complete the connection without
`human intervention. Telephone companies realized the enormous savings
`in manpower(or womanpoweras the majority oftelephone operators at the
`time were women) by automating the call setup and takedown process.
`Stowger switches (also known as crossbar switches) can still be found in the
`central offices of rural America and lesser developed countries.
`Stowger’s design remained the predominant telephone switching tech-
`nology until the mid-1970s. Beginning in the ‘70s, switching technology
`evolved to mainframe computers; that is, no moving parts were used and
`the computer telephony applications made such features as conferencing
`and call forwardingpossible. In 1976, AT&Tinstalledits first #4 Electronic
`Switching System (4ESS) tandem switch. This was followed shortly there-
`after with the 5ESS as a central office switch. ESS central office switches
`did not require a physical connection between incoming and outgoingcir-
`cuits. Paths betweenthecircuits consisted of temporary memory locations
`that enabled the temporary storage oftraffic. For an ESS system, a com-
`putercontrols the assignment, storage, and retrieval of memory locations so
`that a portion of an incoming line (time slot) could be stored in temporary
`memory andretrieved for insertion to an outgoingline.Thisis called a timne
`slot interchange (TSI) memory matrix. The switch control system mapsspe-
`cific time slots on an incoming communication line (such as a DS3) to
`specific time slots on an outgoing communicationline.*
`
`
`
`‘Harte, Lawrence. Telecom Made Simple. Fuquay-Varina, NC: APDG Publishing, 2002.
`
`AT&T Exhibit 1018
`
`Chapter 2
`
`‘Class 4 and 5 Switching
`
`Class 4 and 5 switches are the “brains”of the PSTN.Figure 2-3 illustrates
`the flow of a call from a handset to a Class 5 switch, which in turn hands
`the call off to a Class 4 switch for routing over a long-distance network.
`That call may be routed through other Class 4 switches before terminating
`at the Class 5 switch at the destination endofthe call before being passed
`
`AT&T Exhibit 1018
`AT&T v. VoIP, IPR 2017-01384, Page 6
`
`

`

`The Public Switched Telephone Network (PSTN)
`
`| 13 &
`
` Soeee
`Relationship of Class
`
`4 and 5 switching
`
`Class Network and Relationship to Class 5
`Switching
`Plesse) Quien St Lats
`
`Class 4 Switch Denver
`
`Class 4 Switch Chicago
`
` Class 5 Switch Chicago
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`|
`|
`
`on to the terminating handset. Class 5 switches handlelocal calling and
`Class 4 switches handle long-distance calls. The performance metrics for
`the Class 4 and 5 have beenreliability, scalability, quality of service (QoS),
`signaling, and features.
`
`Class 4 and 5 Architecture One reason for the reputation of Class 4
`and 5 switches being reliable is that they have been tested by time in the
`legacy market. Incremental improvements to the 4ESS included new inter-
`faces, hardware, software, and databases to improve Operations, Adminis-
`tration, Maintenance, and Provisioning (OAM&P). Theinclusion of the 1A
`processor improved memory in the 4 and 5ESS mainframe,allowing for
`translation databases. Ultimately, those databases were interfaced with the
`Centralized Automatic Reporting on Trunks (CAROT). Later, integrated cir-
`cuit chips replaced the magnetic core stores and improved memory and
`boosted the Busy Hour Call Attempt (BHCA)capacity to 700,000 BHCAs.”
`
`Thearchitecture ofthe Class 4 and 5 switch
`Class4and5 Components
`is the product of 25-plus years of design evolution. For the purposes ofthis
`discussion, the Nortel DMS-250, one of the most prevalent products in the
`North American Class 4 market, is used as a real-world example. The other
`
`
`
`2Chapuis, Robert, and AmosJoel. “In the United States, AT&T's Digital Switch Entry No. 4 ESS,
`First Generation Time Division Digital Switch.” Electronics, Computers, and Telephone Systems.
`New York: North Holland Publishing, 1990,p. 337-338.
`
`
`
`AT&T Exhibit 1018
`AT&T v. VoIP, IPR 2017-01384, Page 7
`
`

`

`AT&T v. VoIP, IPR 2017-01384, Page 8
`
`leading product in this marketis the 4ESS from Lucent Technologies. For
`local offices or Class 5, the most prevalent product is the 5ESS from Lucent.
`DMS-250 hardware, for example, is redundant for reliability and decreased
`downtime during upgrades.It has a reliability rating of99.999 percent (the
`five 9s), which meets the industry metric for reliability. The modular design
`ofthe hardware enables the system to scale from 480 to over 100,000 DSOs
`(individual phonelines). The density, or numberof phone lines the switch
`can handle, is one metric of scalability. The DMS$-250 is rated at 800,000
`BHCAs. Tracking BHCAs on a switch is a measure of call-processing capa-
`bility and is another metric for scalability.
`Key hardware components of the DMS-250 system include the DMS
`core, switch matrix, and trunk interface. The DMScore is the central pro-
`cessing unit (CPU) and memory ofthe system, handling high-level call pro-
`cessing, system control functions, system maintenance, andtheinstallation
`of new switch software.
`The DMS-250 switching matrix switches calls to their destinations. Its
`nonblocking architecture enables the switch to communicate with periph-
`erals through fiber optic connections. The trunk interfaces are peripheral
`modules that form a bridge between the DMS-250 switching matrix and the
`trunks it serves. They handle voice and datatraffic to and from customers
`and other switching systems. DMS-250 trunk interfaces terminate DS-1,
`Integrated Services Digital Network (ISDN) Primary Rate Interface (PRD,
`X.75/X.75 packet networking, and analog trunks. They also accommodate
`test and service circuits usedin office andfacility maintenance. It is impor-
`tant to note that the Class 4 switching matrix is a part of the centralized
`architecture ofthe Class 4. Unlike the media gateways in a softswitch solu-
`tion, it must be collocated with the other componentsof the Class 4.
`DMS-250 billing requires the maintenance of real-time, transaction-
`based billing records for many thousands of customers and scores of vari-
`ants in service pricing. The DMS-250 system automatically provides
`detailed data, formats the data into call detail records, and constructsbills.*
`
`
`
`3Nortel Networks. “Product Service Information-DMS$300/250 System Advantage.” www.
`nortel.com, 2001.
`
`AT&T Exhibit 1018
`
`Chapter 2
`
`Private Branch Exchange (PBX)
`
`As the name would imply, a private branch exchange (PBX) is a switch
`owned and maintained by a business with many (20 or more) users. A key
`
`AT&T Exhibit 1018
`AT&T v. VoIP, IPR 2017-01384, Page 8
`
`

`

`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`The Public Switched Telephone Network (PSTN)
`
`
`
`system is used by smaller offices. PBXs and key systems today are com-
`puter based and enable soft changes to be made through an administration
`terminal or PC. Unless the business has a need for technical telecommuni-
`cations personnelon staff for other reasons, the business will normally con-
`tract with their vendor for routine adds, moves, and changes of telephone
`equipment.
`PBX systems are often equipped with key assemblies and systems,
`including voice mail, call accounting, a local maintenance terminal, and a
`dial-in modem. The voice mail system is controlled by the PBX and only
`receives calls when the PBX software determines a message can beleft or
`retrieved. The call accounting system receives system message details on
`all call activities that occur within the PBX. The local terminal provides
`onsite access to the PBX for maintenanceactivities. The dial-in capability
`also provides access to the PBX for maintenanceactivities.*
`
`.
`
`Centrex
`
`After PBXs caught on in the industry, local exchange carriers began to lose
`some of their more lucrative business margins. The response to the PBX
`was Centrex. Centrex is a service offered by a local telephone service
`provider (primarily to businesses) that enables the customer to have fea-
`tures that are typically associated with a PBX. These features include
`three- or four-digit dialing, intercom features, distinctive line ringing for
`inside and outside lines, voice mail, call-waiting indication, and others. Cen-
`trex services flourished andstill have a place for many large, dispersed enti-
`ties such as large universities and major medical centers.
`Oneof the major selling points for Centrex is the lack of capital expen-
`diture up front. That, coupled with the reliability associated with Centrex
`dueto its location in the telephone companycentraloffice, has kept Centrex
`as the primary telephone system in manyof the businesses referenced pre-
`viously. PBXs, however, have cut into what was once a lucrative market for
`the telephone companies and are nowtherule rather than the exception for
`business telephoneservice. This has come about because of inventive ways
`of funding the initial capital outlay and the significantly lower operating
`cost of a PBX versus a comparable Centrex offering.
`
`
`
`
`
`‘Harte, Lawrence. Telecom Made Simple. Fuquay-Varina, NC: APDG Publishing, 2002.
`
`AT&T Exhibit 1018
`AT&T v. VoIP, IPR 2017-01384, Page 9
`
`

`

`
`
`Multiplexing
`
`Chapter 2
`
`AT&T v. VoIP, IPR 2017-01384, Page 10
`
`The earliest approach to getting multiple conversationsover one circuit was
`frequency division multiplexing (FDM). FDM was madepossible by the vac-
`uum tube wherethe rangeoffrequencies was dividedinto parcels that were
`distributed among subscribers.In the first FDM architectures, the overall
`system bandwidth was 96 kHz. This 96 kHz could be divided among a num-
`ber of subscribers into, for example, 5 kHz per subscriber, meaning almost
`20 subscribers could use this circuit.
`FDM is an analog technology and suffers from a number of shortcom-
`ings. It is susceptible to picking up noise along the transmission path. This
`FDMsignal loses its poweroverthe length of the transmission path. FDM
`requires amplifiers to strengthen the signal over that path. However, the
`amplifiers cannot separate the noise from the signal and the endresult is
`an amplified noisy signal.
`The improvement over FDM wastime division multiplexing (TDM).
`TDM was madepossible by the transistor that arrived in the marketin the
`1950s and 1960s. As the name would imply, TDM divides the “me rather
`than the frequencyof a signal over a given circuit. Although FDM wastyp-
`ified by “someof the frequencyall of the time,” TDM is “all of the frequency
`someof the time.” TDM isa digital transmission schemethat uses a small
`numberofdiscrete signal states. Digital carrier systems have only three
`valid signal values: one positive, one negative, and zero. Everythingelse is
`registered as noise. A repeater, known as a regenerator, can receive a weak
`and noisy digital signal, remove the noise, reconstruct the original signal,
`and amplify it before transmitting the signal onto the next segmentof the
`transmission facility. Digitization brings with it the advantages of better
`maintenance and troubleshooting capability, resulting in better reliability.
`Also, a digital system enables improved configuration flexibility.
`TDM has made the multiplexer, also known as the channel bank, possi-
`ble. In the United States, the multiplexer or “mux” enables 24 channels per
`single four-wire facility. This is called a T-1, DS1, or T-Carrier. Outside
`North America and Japan,it is 32 channels per facility and known as Ei.
`These systems came on the market in the early 1960s as a meansto trans-
`port multiple channels ofvoice over expensive transmissionfacilities.
`
`Voice Digitization via Pulse Code Modulation
`
`Oneof the first processes in the transmission of a telephonecall is the con-
`version of an analog signal into a digital one. This processis called pulse
`
`AT&T Exhibit 1018
`
`AT&T Exhibit 1018
`AT&T v. VoIP, IPR 2017-01384, Page 10
`
`

`

`The Public Switched Telephone Network (PSTN)
`
`Bienuaaaeee=oe
`
`417
`
`code modulation (PCM). This is a four-step process consisting of pulse
`amplitude modulation (PAM) sampling, companding, quantization, and
`encoding.
`
`Thefirst stage in PCM is known
`Pulse Amplitude Modulation (PAM)
`as PAM. In order for an analog signal to be represented as a digitally
`encoded bitstream, the analog signal must be sampled at a rate that is
`equal to twice the bandwidth of the channel over which the signal is to be
`transmitted. As each analog voice channelis allocated 4 kHz of bandwidth,
`each voice signal is sampled at twice that rate, or 8,000 samples per sec-
`ond, In a T-Carrier, the standard in North America and Japan, each chan-
`nel is sampled every one eight-thousandthof a secondin rotation, resulting
`in the generation of 8,000 pulse amplitude samples from each channel
`every second. If the sampling rate is too high, too much information is
`transmitted and bandwidth is wasted. If the samplingrateis too low, alias-
`ing mayresult. Aliasing is the interpretation of the sample points as a false
`waveform due to the lack of samples.
`
`Companding Thesecond process of PCM is companding. Compandingis
`the process of compressing the values of the PAM samples to fit the non-
`linear quantizing scale that results in bandwidth savings of more than 30
`percent.It is called companding as the sampleis compressed for transmis-
`sion and expandedfor reception.°
`
`Quantization The third stage in PCM is quantization. In quantization,
`values are assigned to each sample within a constrained range.In using a
`limited numberofbits to represent each sample, the signal is quantized.
`The difference between the actuallevel of the input analog signal and the
`digitized representation is known as quantization noise. Noise is a detrac-
`tion to voice quality and it is necessary to minimize noise. The way to do
`this is to use morebits, thus providing better granularity. In this case, an
`inevitable trade-offtakes place bewteen bandwidth and quality. More band-
`width usually improves signal quality, but bandwidth costs money. Service
`providers, whether using TDM or Voice over IP (VoIP) for voice transmis-
`sion will always have to choose between quality and bandwidth.A process
`known as nonuniform quantization involves the usage of smaller
`
`
`
`5Shepard, Steven. SONET/SDH Demystified. New York: McGraw-Hill, 2001. p. 15-21.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`AT&T Exhibit 1018
`AT&T v. VoIP, IPR 2017-01384, Page 11
`
`

`

`Chapter 2
`
`quantization steps at smaller signal levels and larger quantization steps for
`larger signallevels. This gives the signal greater granularity or quality at
`low signal levels and less granularity (quality) at high signal levels. The
`result is to spread the signal-to-noise ratio more evenly across the range of
`different signals and to enable fewer bits to be used compared to uniform
`quantization. This process results in less bandwidth being consumed than
`for uniform quantization.®
`
`AT&T v. VoIP, IPR 2017-01384, Page 12
`
`Encoding The fourth and final process in PCM is encoding the signal.
`This is performed by a codec (coder/decoder). Three types of codecs exist:
`waveform codecs, source codecs (also known as vocoders), and hybrid
`codecs. Waveform codecs sample and code an incoming analog signal
`without regard to how the signal was generated. Quantized valuesof the
`samples are then transmitted to the destination where theoriginal signal
`is reconstructed, at least to a certain approximation of the original. Wave-
`form codecs are known for simplicity with high-quality output. The disad-
`vantage of waveform codecs is that they consume considerably more
`bandwidth than the other codecs. When waveform codecs are used at low
`bandwidth, speech quality degrades markedly.
`Source codecs match an incomingsignal to a mathematical modelofhow
`speech is produced. They use the linear predictive filter model of the vocal
`tract, with a voiced/unvoicedflag to represent the excitation that is applied
`to thefilter. The filter represents the vocal tract and the voice/unvoiced flag
`represents whether a voiced or unvoiced input is received from the vocal
`chords. The information transmitted is a set of model parameters as
`opposed to the signalitself. The receiver, using the same modeling tech-
`nique in reverse, reconstructs the values received into an analog signal.
`Source codecs also operate at low bit rates and reproduce a synthetically
`sounding voice. Using higher bit rates does not result in improved voice
`quality. Vocoders (source codecs) are most widely used in private and mili-
`tary applications.
`Hybrid codecs are deployed in an attempt to derive the benefits from
`both technologies. They perform some degree of waveform matching while
`mimicking the architecture of human speech. Hybrid codecs provide better
`voice quality at low bandwidth than waveform codecs. Table 2-1 provides an
`outline of the different ITU codec standards and Table 2-2 lists the para-
`meters of the voice codecs.
`
`
`SCollins, Daniel. Carrier Grade Voice Over IP. New York: McGraw-Hill, 2001. p. 95-96.
`
`AT&T Exhibit 1018
`
`AT&T Exhibit 1018
`AT&T v. VoIP, IPR 2017-01384, Page 12
`
`

`

`The Public Switched Telephone Network (PSTN)
`
`A subjective rating system to determine the Mean Opinion Score
`(MOS)or the quality of telephone connections
`A maximum one-way delay end to end for a VoIP call (150 ms)
`Echo cancellers
`
`Digital network echo cancellers
`
`PCM of voice frequencies
`
`7 kHz audio coding within 64 Kbps
`A dual-rate speech coder for multimedia communications transmitting
`at 5.3 and 6.3 Kbps
`Coding for speech at 8 Kbps using conjugate-structure algebraic code-
`excited linear-prediction (CS-ACELP)
`Annex A reduced complexity 8 Kbps CS-ACELPspeech codec
`
`G.711
`G.721, G.723, G.726
`G.728
`G.729
`G.723.1
`
`16,24,32,40
`16
`8
`
`:
`
`MOS = Codec
`
`Waveform
`
`4.8
`4.2
`4,2
`4.2
`
`CS
`
`Popular Speech Codecs Codecs are best known for the sophisticated
`compression algorithms they introduce into a conversation. Bandwidth
`costs service providers money. The challenge for many service providers is
`to squeeze as muchtraffic as possible into one “pipe,” that is one channel.
`Most codecs allow multiple conversationsto be carried on one 64 kbps chan-
`
`Table 2-1
`
`Descriptions of
`voice codecs(ITU)
`
`
`
`Table 2-2
`
`Parameters of voice
`codecs
`
`A packet-based multimedia communications system
`Specifies a model to map actual audio signalsto their representations
`inside the human head
`Digital subscriber signaling system number 1 ISDN user-network
`interface layer 3 specification for basic call control
`aeT
`
`nel. There is an inevitable trade off in compression for voice quality in the
`
`AT&T Exhibit 1018
`AT&T v. VoIP, IPR 2017-01384, Page 13
`
`

`

`
`
`Chapter 2
`
`conversation. The challenge for service providers is to balance the econom-
`ics of compression with savings in bandwidth costs.
`
`G.711 G.711 is the best-known coding techniquein use today. It is a wave-
`form codec andis the coding technique usedin circuit-switched telephone
`networks all over the world. G.711 has a sampling rate of 8,000 Hz. If
`uniform quantization wereto be used, the signal levels commonly found in
`speech would be such that at least 12 bits per sample would be needed,giv-
`ingit a bit rate of 96 Kbps. Nonuniform quantization is used with eight bits
`used to represent each sample. This quantization leads to the well-known
`64 Kbps DSO rate. G.711 is often referred to as PCM. G.711 has twovari-
`ants: A-law and mu-law. Mu-law is used in North America and Japan where
`T-Carrier systems prevail. A-law is used everywhereelse in the world. The
`difference between the two is the way nonuniform quantization is per-
`formed. Both are symmetrical at approximately zero. Both A-law and mu-
`law offer good voice quality with a MOSof4.3, with 5 being the best and 1
`being the worst. Despite being the predominantcodec in the industry, G.711
`suffers one significant drawback; it consumes 64 Kbps in bandwidth. Car-
`riers seek to deliver voice quality using little bandwidth, thus saving on
`operating costs.
`G.728 LD-CELP Code-Excited Linear Predictor (LD-CELP) codecs imple-
`ment a filter and contain a codebook of acoustic vectors. Each vector
`contains a set of elements in which the elements represent various charac-
`teristics of the excitation signal. CELP coders transmit to the receiving end
`a set of information determiningfilter coefficients, gain, and a pointerto the
`chosen excitation vector. The receiving end contains the same code book and
`filter capabilities so that it reconstructs the original signal. G.728 is a
`backward-adaptivecoder as it uses previous speech samples to determine
`the applicablefilter coefficients. G.728 operateson five samples at one time.
`That is, 5 samples at 8,000 Hz are needed to determine a codebook vector
`and filter coefficients based upon previous and current samples. Given a
`coder operating onfive samplesat a time,a delay ofless than 1 millisecond
`is the result. Low delay equals better voice quality.
`The G.728 codebook contains 1,024 vectors, which requires a 10-bit index
`value for transmission. It also uses 5 samples at a time taken at a rate of
`8,000 per second.For each of those 5 samples, G.728 results in a transmit-
`ted bit rate of 16 Kbps. Hence, G.728 has a transmitted bit rate of 16 Kbps.
`Another advantagehereis that this coder introduces a delay of 0.625 mil-
`liseconds with an MOS of3.9. The difference from G.711’s MOSof4.3 is
`imperceptible to the human ear. The bandwidth savings between G.728’s 16
`Kbps per conversation and G.711’s 64 Kbps per conversation make G.728
`very attractive to carriers given the savings in bandwidth.
`
`AT&T Exhibit 101
`AT&T v. VoIP, IPR 2017-01384, Page 14
`
`
`
`AT&T Exhibit 1018
`AT&T v. VoIP, IPR 2017-01384, Page 14
`
`

`

`The Public Switched Telephone Network (PSTN)
`
`
`
`|
`
`G.723.1 ACELP G.723.1 ACELPcan operate at either 6.3 Kbps or 5.3 Kbps
`with the 6.3 Kbps providing highervoice quality. Bit rates are contained in
`the coder and decoder, and the transition between the two can be made dur-
`ing a conversation. The coder takes a bank-limited input speech signal that
`is sampled a 8,000 Hz and undergoes uniform PCM quantization,resulting
`in a 16-bit PCM signal. The encoder then operates on blocks or frames of
`240 samples at a time. Each frame corresponds to 30 millisecondsofspeech,
`which means that the coder causes a delay of 30 milliseconds. With a look-
`ahead delay of 7.5 milliseconds, the total algorithmic delay is 37.5 millisec-
`onds. G.723.1 gives an MOS of 3.8, which is highly advantageousin regards
`to the bandwidth used. The delay of 37.5 milliseconds one way does present
`an impediment to good quality, but the round-trip delay over varying
`aspects of a network determines the final delay and not necessarily the
`codec used.
`
`|
`
`|
`|
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
` ‘Thid.
`
`G.729 G.729 is a speech coder that operates at 8 Kbps. This coder uses
`input frames of 10 milliseconds, corresponding to 80 samples at a sampling
`rate of 8,000 Hz. This coder includes a 5-millisecond look-ahead,resulting
`in an algorithmic delay of 15 milliseconds (considerably better than
`G.723.1). G.729 uses an 80-bit frame. The transmitted bit rate is 8 Kbps.
`Given that it turns in an MOS of4.0, G.729 is perhaps the best trade-off in
`bandwidth for voice quality. The previous paragraphsprovide an overview
`of the multiple means of maximizing the efficiency of transport via the
`PSTN. Wefind today that TDM is almost synonymouswith circuit switch-
`ing. Telecommunications engineers use the term TDMto describe a circuit-
`switched solution. A 64 Kbps G.711 codec is the standard in use on the
`PSTN. The codecs described in the previous pages apply to VoIP as well.
`VoIP engineers seeking to squeeze more conversations over valuable band-
`width have found these codecs very valuable in compressing VoIP conver-
`sations over an IP circuit.”
`
`Signaling
`
`
`Signaling describes the process of how calls are set up and torn down. Gen-
`
`erally speaking, there are three main functions of signaling: supervision,
`alerting, and addressing. Supervision refers to monitoring the status of a
`
`
`line or circuit to determineif thereis traffic on the line. Alerting deals with
`the ringing of a phone indicating the arrival of an incomingcall. Address-
`
`ing is the routing of a call over a network. As telephone networks matured,
`
`
`
`
`
`AT&T Exhibit 1018
`AT&T v. VoIP, IPR 2017-01384, Page 15
`
`

`

`AT&T v. VoIP, IPR 2017-01384, Page 16
`
`Signaling System 7 (SS7) For muchof the history of circuit-switched
`networks, signaling followed the same path as the conversation. This is
`called Channel-Associated Signaling (CAS) andis still in wide use today.
`R1 Multifrequency (MF) used in North American markets and R2 Multi-
`Frequency Compelled (RFC) used elsewherein the world are the best exam-
`ples of this. Another name for this is in-channel signaling. The newer
`technology for signaling is called Common Channel Signaling (CCS), also
`knownas out-of-band signaling. CCS uses a separate transmission path for
`call signaling and not the bearer path for the call. This separation enables
`the signaling to be handled in a different mannerto the call. This enables
`signaling to be managed by a network independent of the transport net-
`work. Figure 2-4 details the difference between CAS and CCS.
`
`Stallings, William. ISDN and Broadband ISDN with Frame Relay and ATM. New York: Pren-
`tice Hall, 1995. p.292.
`
`AT&T Exhibit 1018
`
`Chapter 2
`
`individual nations developed their proprietary signaling systems. Ulti-
`mately, there become a signaling protocol for every national phoneservice
`in the world. Frankly, it is a miracle that international calls are ever com-
`pleted given the complexity of interfacing national signaling protocols.
`
`Speech and Signaling
`
`Switch
`
`,
`I
`Channel Associated Signaling
`
`Signaling
`
`Switch
`
`Switch
`
`Common Channel Signaling
`
`AT&T Exhibit 1018
`AT&T v. VoIP, IPR 2017-01384, Page 16
`
`

`

`The Public Switched Telephone Network (PSTN)
`
`
`
`| 23
`
`Signaling System 7 (SS7) is the standard for CCS with many national
`variants throughout the world (such as Mexico’s NOM-112). It routes con-
`trol messages through the network to perform call management(setup,
`maintenance, and termination) and network management functions.
`Although the network being controlled is circuit switched,the controlsig-
`naling is implemented using packet-switching technology. In effect, a
`packet-switched networkis overlaid on a circuit-switched network in order
`to operate and control the circuit-switched network. SS7 defines the func-
`tions that are performed in the packet-switched network but does not dic-
`tate any particular hardware implementation.*
`The S87 network and protocol are used for the following:
`
`= Basic call setup, management, and tear down
`= Wireless services such as pers