SESSION 708 Tuesda Afternoon (Cuyamaca)
`Applications of Neura Networks in Signal Processing
`an Communications - H
`ORGANIZERS: T. Aoyama, NTI‘ Telecommunication
`Networks Lab., Japan, and S.V. Kartalopoulos, AT&-T
`Laboratories, USA
`CHAIRPERSONS: T. Aoyama, S. V. Kartalopoulos
`SPONSORS: Signal Processin and Communication
`Electronics; Communications ystems Engineering;
`Data Communications
`
`708.1 Perfonnance Improvement of LMS Algorithm Using
`Hopfield Model Network
`K. Takahashi, S. Mori, Keio University, Japan
`708.2 Signal Processing and Implementation of Motion
`Detection Neurons in Optical Pathways
`S. V. Kartalopoulos, AT&'l" Bell Laboratories, USA
`708.3 Holonic Routing Scheme Based on Neural Computations
`A. Chugo, W. Sotelo, I. lida, Fujitsu Limited, Japan
`708.4 Neural Network Solution to the Link Scheduling
`Problem Using Convex Relaxation
`R. G. Ogier, D. A. Beyer, SRI International, USA
`708.5 A Neural Network Solution for Call Routing with
`Preferential Call Placement
`P. J. W. Melsa, J. B. Kenney, C. E. Rohrs, Tellabs, lnc., USA
`708.6 Integration of ATM Call Admission Control and Link
`Capacity Control By Distributed Neural Networks
`A. Hiramatsu, NIT Communication Switching Labs, Japan
`708.7 VLSI for ATM Network Management
`D. B. Schwartz, V. K. Samalam, GTE Laboratories lnc., USA
`
`SESSION 800 Wednesday Morning (Riesling)
`Internetworking
`ORGANIZERS: M. El-Zarki, University of Pennsylvania,
`USA, B. Kadava, IBM, USA
`CHAIRPERSON: M. El-Zarki
`SPONSORS: Com uter Communications,-
`Communications oftware
`
`800.1 A Proposal for Counting Access to Synchronous Time
`Slots and its Application to Wide Area LANs
`N. Kuroyanagi, Tokyo Engineering University, Japan
`800.2 A Protocol for Eraser Node Implementation Within The
`DQDB Framework
`M. Zukerman, P. G. Potter, Telecom Australia Research
`Laboratories, Australia
`800.3 Transient Phenomena in Bridged Local Area Networks
`C. Ersoy, S. S. Panwar, Polytechnic University, USA;
`R. Dalias, D. Segal, SIAC lnc., USA
`800.4 An Adaptive Token Bus Protocol for LAN Access Control
`S. T. Vuong, H. Cu, University of British Columbia, Canada
`800.5 A Fair Integrated Voice/Data Protocol for Token Ring
`Networks
`
`W. Chen, R. Lee, J. Yu, C. Huang, National Tsing Hua
`University, Taiwan
`800.6 Comparative Analysis of Virtual-Circuit Routing Control
`for ISDN Frame-Relay Networks
`I(. K. Y. I-Io, AT&T Bell Laboratories, USA
`800.7 A High Speed Protocol Processor to Boost Gateway
`Performance
`T. I-lirata, S. Matsui, T. Yokoyama, M. Mizutani, M. Terada,
`Hitachi Ltd ., Japan
`
`SESSION 801 Wednesday Morning (Chablis)
`Architecture and Protocols for Broadband ISDN
`Networks
`ORGANIZER: B. Jabb ari, George Mason Univ., USA
`CHAIRPERSONS: B. Jabbari and R. L. Pickholtz, George
`Washington Univ., USA
`SPONSORS: Data Communications; Computer
`Communication; Communications Switching
`801.1 Queueing Strategies for Multicast Packet Switching
`J. Y. Hui, Rutgers University, USA; T. Renner, University of
`Stuttgart, Germany (FR)
`801.2 Queueing Analysis for ATM Switching of Continuous-
`Bit—Rate Traffic - A Recursion Computation Method
`A. K. Wong, AT&T Bell Laboratories, USA
`801.3 Traffic Experimentation in ATM Testbed
`O. Aboul—Magd, M. Wernik, Bell-Northern Research Ltd.,
`Canada
`
`801.4 Message Waiting Times and Delays in ATDM Switching
`Elements
`H. Bruneel, Ghent State University, Belgium
`801.5 Performance Evaluation of an Input-Queued ATM Switch
`With Internal Speed-up and Finite Output Queues
`G. Bruzzi, Polytechnic of Milan, Italy; A. Pattavina,
`University La Sapienza, Italy
`801.6 A Congestion Control Framework for BISDN Using
`Distributed Source Control
`G. Ramamurthy, R. S. Dighe, AT&T Bell Laboratories, USA
`801.7 Bandwidth Sharing in Integrated Networks
`B. Kraimeche, Pratt Institute, USA; M. Schwartz, Columbia
`University, USA
`
`SESSION 802 Wednesday Morning (Colombard)
`Inte ated OAM for Network Management and Control
`OR ANIZER: S. E. Aidarous, BNR, Canada
`Cl-IAIRPERSON: S. E. Aidarous
`SPONSOR: Network Operations and Management
`802.1 A Unified Operating Environment for Customer
`Servicing
`W. Caldwell, R. Khare, Y. Lee, J. Pfaff, H. Stovall, D.
`Zuckerman, AT&T Bell Laboratories,‘ S. Colladay, J. Hunter,
`R. Krauchunas, AT&T Network Services Division, USA
`802.2 New Services‘ Operations Support in an Intelligent
`Network
`I. Dawis, C. Loberg, SBC Technology Resources, lnc., USA
`802.3 0Sl Interface for Integrated Network Management
`D. Helmy, Bell-Northern Research Ltd., Canada
`802.4 0AM&P Support Software: Current Role and Evolution
`R. Saracco, S. Barra, CSELT, Italy; C. Carrelli, P. Tiribelli,
`SIP Headquarters, Italy
`802.5 Operation, Administration and Maintenance Systems of
`the Optical Fiber Loop
`T. Uenoya, F. Ashiya, N. Tomita, K. Satoh, T. Sakai,
`N'l'l‘, Japan
`802.6 Toward the Intelligent Integrated Network Management
`T. Liao, D. Seret, ENST, France
`802.7 NEOPILOT: An Integrated ISDN Fault
`Management System
`H. Shimazaki, N. Takahashi, NTT
`Communications Switching Labs, Japan
`
`xxii
`
`ERICSSON EXHIBIT 1015
`ERICSSON EXHIBIT 1015
`ERICSSON V. IV
`ERICSSON V. IV
`|PR201 5-01 872
`IPR2015-01872
`
`

`
`SESSION 805 Wednesda Morning (Palomar)
`Broadband Switching Te nologies
`ORGANIZER: G. W. R. Luderer, AT&T Bell
`Laboratories (retired), USA
`CHAIRPERSON: G. W. R. Luderer
`SPONSOR: Communications Switching
`805.1 An Experimental Photonic ATM Switching System Using
`an InP Semiconductor Optical Switch
`Y. Takahashi, K. Ando, H. Inoue, E. Amada, Hitachi Ltd.,
`Japan
`805.2 Physical Design Issues for Very Large ATM Switching
`Systems
`K. C. Young, T. C. Banwell, S. S. Cheng, R. C. Estes,
`S. F. Habiby, G. A. Hayward, T. K. Helstern, G. R.
`Lalk, D. D. Mahoney, D. K. Wilson, Bellcore, USA
`805.3 A Distributed Modular Tera—bit/sec ATM Switch
`H. J. Chao, Bellcore, USA
`805.4 A Multi-Purpose Memory Switch LSI for ATM-Based
`Systems
`H. Yamada, H. Kai, T. Takahashi, NTT Communication
`Switching Labs, S. Yamada, NTT Electronics Technology
`Corp., Japan
`805.5 Scalability Study of Self-Routing Packet Switch Fabrics
`for Very Large Scale Broadband ISDN Central Offices
`J. N. Giacopelli, T. T. Lee, W. E. Stephens, Bellcore, USA
`805.6 Implementation of a 16 to 16 Switching Element for ATM
`Exchanges
`P. Barri, J. Goubert, Alcatel Bell, Belgium
`
`SESSION 806 Wednesday Morning (Lagunu)
`Advanced Intelligent Networks Services Concepts
`ORGANIZERS: F. J. Weisser, BellSouth Services, USA,
`R. Clendennin , Be11South Services, USA
`CHAIRPERSO S: F. J. Weisser, R. Clendenning
`SPONSOR: Multimedia Services and Terminals
`
`806.1 Advanced Intelligent Network: 1993 Snapshot
`R. K. Berman, Bellcore, USA
`806.2 Diverse Service Architectures Using Alternate AIN
`Elements
`
`V. Lapi, P. Vapheas, J. Brana, A. Ramani, A. Gopin,
`D. Marutiak, AT&T Bell Laboratories, USA
`806.3 Intelligent Network Routing Using CCS7 and ISDN
`M. Rain, 5. Chau, Bcll—Northern Research Inc., USA;
`B. Coyle, Bell Atlantic, USA
`806.4 Voice Services Concepts for Intelligent Networks
`C. B. I-lirschman, Ameritech Services, USA
`806.5 A Strategy for Studying Services and Service Interactions
`in Intelligent Networks
`R. Verstraete, H. Decuypere, Alcatel Bell Research Center,
`Belgium
`806.6 Service Verification Testing for Intelligent Network
`Services
`B. Hoang, A. Atai, Bellcore, USA
`
`SESSION 803 Wednesda Morning (Burgundy)
`Optical WDM and Fiber
`tudies
`ORGANIZERS: B. Glance, AT&T Bell Labs, USA,
`M. Maeda, Bellcore, USA, H. Kobrinsky, Bellcore, USA
`CHAIRPERSONS: B. Glance, H. Kobrinsky
`SPONSOR: Optical Communications
`803.1 Optical Wavelength Division Multiplex
`P. S. Henry, AT&T Bell Laboratories, USA
`803.2 Broadband Information Distribution Networks
`Employing Optical Frequency Division Multiplexing
`Technologies
`H. Toba, K. Oda, K. Nakanishi, N. Shibata, K. Nosu, NTT
`Transmission Systems Labs, Japan
`803.3 Monolithic Tunable Optical Active Filters and Receivers
`Based On Optical Resonant Amplifier Structure
`F. S. Choa, T. L. Koch, AT&T Bell Laboratories, USA
`803.4 A Consideration on Chromatic and Polarization
`Dispersions in Single-Mode Fibers for Large Capacity
`Transmission Systems
`M. Ohashi, M. Tsubokawa, M. Kawase, NTI‘
`Transmission Systems Labs, Japan
`803.5 Optical FDM Switch Experiments With Tunable Fiber
`Fabry-Perot Filters
`K. Y. Eng, M. A. Santoro, J. Stone, T. L. Koch, AT&T Bell
`Laboratories, USA
`803.6 10 Cbitls Modulation Performance of Distributed
`Feedback Lasers for Optical Fiber Communication
`Systems
`R. S. Vodhanel, A. F. Elrefaie, R. E. Wagner, M. Z. Iqbal,
`J. L. Gimlett, Bellcore, USA
`803.7 Wavelength-Selective Circuit and Packet Switching Using
`Acousto-Optic Tunable Filters
`K. W. Cheung, D. A. Smith,J. E. Baran, J. J. Johnson,
`Bellcore, USA
`
`SESSION 804 Wednesday Morning (Sauterne)
`Cellular Radio Networking
`ORGANIZER: P. T. Porter, Bellcore, USA
`CHAIRPERSON: P. T. Porter
`SPONSOR: Radio Communication
`
`804.1 Operation and Performance of a Self-Organizing
`Frequency Assignment Method for TDMA Portable
`Radio
`J. C-I. Chuang, Bellcore, USA
`804.2 Channel Allocation Strategies in Dual Mode Digital
`Cellular Networks
`A. K. Kakaes, The George Washington Univ., USA
`804.3 Novel Technique For Efficient Channel Utilization in a
`Mobile Radio System
`M. D. Yacoub, FEE—UNICAMP, Brazil;
`K. W. Cattermole, University of Essex, UK
`804.4 Microcellular Mobile Radio Spectrum Efficiency
`R. H. Muammar, University of Illinois, USA
`804.5 Signaling Protocols Using Mobile Application Part for
`Call Control in the Digital Mobile Network
`H. Sawada, A. Nakajima, K. Yamamoto, NIT Mobile
`Communications Division, S. Suzuki, NTI' Radio
`Communication Systems, Japan
`804.6 Mobile PR-Network in a Densely Populated Urban
`Environment
`T. Anwar, C. Jones, D. Schreib, C. Strauch, Technical
`University of Aachen, Germany (FR)
`804.7 Architecture and Protocol Design for a Car-to-
`Intrastructure Packet Radio Network
`F. Davoli, A. Giordano, S. Zappatore, University of
`Genoa, Italy
`
`

`
`SESSION 807 Wednesday Morning (San Antonio)
`Combined Equalization and Coding
`ORGANIZERS: C.-E. Sundberg, AT&T Bell Labs, USA,
`J. Ha enauer,DLR Oberpfaffenhofen, Germany (FR)
`CH
`ERSON: C.-E. Sundberg
`SPONSOR: Communication Theory
`807.1 ]oint Data and Channel Estimation Using Fast Blind
`Trellis Search Techniques
`N. Seshadri, AT&T Bell Laboratories, USA
`807.2 Combined Equalization and Coding with the Finite-
`Length Decision Feedback Equalization
`J. M. Cioffi, I’. Algoet, P. S. Chow, Stanford Univ., USA
`807.3 Baseband Trellis Coded Modulation With Combined
`Equalization/Decoding for High Bit Rate Digital
`Subscriber Loops
`D. D. Falconer, P. Mohanraj, T. A. Kwasniewski, Carleton
`University, Canada
`807.4 Quaternary Codes for Partial-Response Channels
`E. Eleftheriou, R. Cideciyan, IBM Research Division,
`Switzerland
`
`807.5 Optimum and Sub-Optimum Detection of Coded Data
`Distributed by Time-Varying Intersymbol Interference
`W. Koch, A. Baier, Philips Kornmunikations lndustrie
`AG, Germany (FR)
`807.6 A Multirate Signal Processing Approach to Block
`Decision Feedback Equalization
`R. Ramesh, California Institute of Technology, USA
`
`SESSION 808 Wednesday Morning (Cuyumaca)
`VLSI Ap lications
`ORGAN ZER: C. M. Puckette, General Electric, USA
`CHAIRPERSON: C. M. Puckette
`SPONSOR: Signal Processing and Communication
`Electronics
`
`808.1 An Optimized Digital Modern Architecture for VLSI
`Implementation
`V. Y. I-lu, LinCom Corporation, USA
`808.2 Signal Processing for a Spectrum Efficient Single
`Channel Digital Radio System
`M. Lo Curto, M. Salerno, Siemens, Italy
`808.3 Design of a Hybrid Receiver for the Olympus Spacecraft
`Beacons
`
`D. G. Sweeney, J. C. McKeeman, Virginia Polytechnic
`Institute and State University, USA
`808.4 An Experimental VLSI Implementation of Low-Overhead
`Symbol Timing and Frequency Offset Estimation for
`TDMA Portable Radio Applications
`N. R. Sollenberger, Bellcore, USA
`808.5 Minimized Method Viterbi Decoding: 600 MBIT/S Per
`Chip
`G. Fettweis, H. Dawid, H. Meyr, Aachen University of
`Technology, Germany (FR)
`808.6 Efficient DSP Based Detection of DTMF Tones
`R. A. Valenzuela, AT&T Bell Laboratories, USA
`808.7 A 200-MHz CMOS X/SIN(X) Digital Filter for
`Compensating D/A Converter Frequency Response
`Distortion in High-Speed Communication Systems
`H. Samueli, T. Lin, Univ. of California, Los Angeles, USA
`
`SESSION 900 Wednesday Afternoon (Chablis)
`Broadband Network Architecture and Standards
`ORGANIZER: R. Sinha,Bel1core, USA
`CHAIRPERSON: R. Sinha
`SPONSORS: Com uter Communications; Data
`Communications;
`tandards
`900.1 A Versatile System Architecture for ISDN Systems
`H. Mori, A. Maeda, H. Miyakita, Kokusai Denshin
`Denwa Co., Ltd., Japan: I. Shimizu, H. Murata, NEC
`Corporation, Japan
`900.2 ATMR: A Ring Architecture for Broadband Networks
`T. Honda, K. Imai, H. Kasahara, T. Ito, NT1"
`Communication Switching Labs, Japan
`900.3 An Architecture for Connectionless Data Service in B-
`ISDN
`Y. Ujihashi, T. Shikama, K. Watanabe, K. Oshima,
`S. Aoyama, Mitsubishi Electric Corporation, Japan
`900.4 Progress on Broadband ISDN User-Network Interface
`Standards
`I. Anderson, AT&T Bell Laboratories, USA
`900.5 Error Detection and Correction Options for Data Services
`in Broadband ISDN
`R. Damodaram, BellSouth Science & Technology, USA;
`S. Dravida, AT&T Bell Laboratories, USA
`900.6 Perfonnance and Reliability of DQDB Metropolitan
`Networks Under Faults
`M. Kadoch, Teleglobe Canada Inc, Canada;
`A.K. Elhakeem, Concordia University, Canada
`900.7 Some Policies for Circuit Allocation and their Packet
`Capacity in Hybrid Switching Systems like FDDI-II
`J. Wu, R. Roux, ALCATEL Business Systems, France
`
`SESSION 901 Wednesday Afternoon (Riesling)
`Signal Processing for Magnetic Recording
`ORGANIZER: P. H. Siegel, Univ. of Cali ornia, San
`Die 0, USA
`CH RPERSON: J. Wolf, Univ. of California, San Diego,
`USA
`SPONSOR: Processing for Magnetic Recording
`901.1 Bounds on the Capacity of the Peak Shift Magnetic
`Recording Channel
`E. Zehavi, S. Shamai (Shitz), Technion - Israel Institute of
`Tech, Israel
`901.2 An Improved Sliding WindowData Compression
`Algorithm Based on the Lempel-Ziv Data Compression
`Algorithm
`P. E. Bender, I. K. Wolf, University of California, San
`Diego, USA
`901.3 Combining ECC with Modulation: Performance
`Comparisons
`M. Blaum, IBM Research Di\n‘sion, USA
`901.4 Coded Partial Response Signalling with Peak Detection
`A. J. Armstrong, ]. K. Wolf, University of California, San
`Diego, USA
`901.5 Area—Efficient Architectures for the Viterbi Algorithm
`C. B. Shung, P. H. Siege], IBM Almadcn Research Center,
`USA; H. D. Lin, University of California, Berkeley, USA;
`H. K. Thapar, IBM General Product Division, USA
`901.6 Timing Recovery for Adaptive Decision Feedback
`Equalization of the Magnetic Storage Channel
`W. L. Abbott, I. M. Cioffi, Stanford University, USA
`
`xxiv
`
`

`
`SESSION 902 Wednesday Afternoon (Bur ndy)
`Topics in Mobile Satellite, Cellular and In oor
`Communications Systems
`ORGANIZER: P. T. Mathiopoulos, University of British
`Columbia, Canada
`CHAIRPERSONS: P. T. Mathiopoulos, G. Schulek,
`Ro ers Cantel, Inc., Canada
`SP NSOR: Satellite and Space Communications
`902.1 Performance Evaluation of Quasi-Synchronous Code
`Division Multiple Access (QS-CDMA) for Satellite
`Mobile Systems
`R. De Gaudenzi, C. Elia, R. Viola, European Space
`Agency, The Netherlands
`902.2 Performance Analysis of Error Control Coding for Mobile
`Channels
`D. Haccoun, F. Gagnon, Ecole Polytechnique de Montreal,
`Canada
`
`902.3 Error Protection for Low Rate Speech Transmission Over
`a Mobile Satellite Channel
`F. F. Tzeng, P. G. Sherwood, COMSAT Laboratories, USA
`902.4 Multi-Beam Land Mobile Satellite Communications
`Experiments Using ETS-V/VI
`Y. Nishi, H. Kondo, H. Komagata, NTI‘ Radio
`Communication Systems, Japan
`902.5 Rapid Acquisition Concept for Voice Activated CDMA
`Communication
`M. K. Sust, R. F. Kaufmann, F. Molitor, SCHRACK
`Aerospace, Austria; G. A. Bjornstrom, European Space
`Research and Technology Centre, The Netherlands
`902.6 Digital Intermediate Frequency Demodulation Technique
`for Cellular Communication Systems
`H. Tomita, Y. Yokoyama, T. Matsuki, NEC Corp., Japan
`902.7 Coding and Optimum Baseband Combining for
`Wideband TDMA Indoor Wireless Channels
`
`C. L. Despins, D. D. Falconer, S. A. Mahmoud, Carleton
`University, Canada
`
`SESSION 903 Wednesday Afternoon (Sauterne)
`Photonic Networking
`ORGANIZERS: P. Green, IBM, USA, A. S. Acampora,
`Columbia University, USA
`CHAIRPERSON: R. Ramaswami, IBM, USA
`SPONSORS: Optical Communications;
`Network Operations and Management
`903.1 FDDI Performance Analysis: Delay Approximations
`R. O. LaMaire, IBM Research Division, USA;
`E. M. Spiegel, Princeton University, USA
`903.2 Supervisory Signal Transmission Methods for Optical
`Amplifier Repeater Systems
`5. Matsuoka, Y. Yamabayashi, K. Aida, K. Nakagawa,
`NTT Transmission Systems Labs, Japan
`903.3 New Low-Level Flow Control Strategies for Fiber Optic
`Token Rings
`P. Martini, University of Paderborn, Germany (FR);
`R. Presty, France Telecom University, France;
`Th. Welzel, ComConsult; Germany (FR)
`903.4 An Experimental Broadband and Telephony Passive
`Optical Network
`A. M. Hill, D. B. Payne, K. J. Blyth, D. S. Forrester, A.
`Silvertown, J. W. Arkwright, D. W. Faulkner, J. W. Ballance,
`British Telecom Research Labs, UK
`903.5 On the Performance of I-‘iber—Optic CDMA Systems
`M. Y. Azizoglu, Y. Li, M.l.T., USA; J. A. Salehi,
`Bellcore, USA
`
`903.6 Linear Lightwave Networks: How Far Can They Go?
`T. E. Stern, Columbia University, USA
`903.7 Analysis of Multihop Lightwave Networks
`Z. Zhang, A. S. Acampora, Columbia University, USA
`
`SESSION 904 Wednesday Afternoon (San Antonio)
`Protocol Specification, Verification, and Testing
`ORGANIZER: C. Chow, Bellcore, USA
`CHAIRPERSON:
`SPONSOR: Communications Software
`
`904.1 A Novel Approach to Protocol Test Sequence Generation
`S. T. Vuong, K. C. K0, University of British Columbia,
`Canada
`
`904.2 An Improved Method for Automatic Validation of
`Communication Protocols
`M. T. Liu, S. H. Yu, Y. W. Yao, The Ohio State Univ., USA
`904.3 An Execution Model for LOTOS Specifications
`C. Wu, G. Bochmann, University of Montreal, Canada
`904.4 A State Transition Design Support System
`S. Kawaguchi, Y. Kaneko, C. Tanno, Hitachi Ltd ., Japan
`904.5 A Parallel Model for Protocol Converters
`M. T. Liu, Y. Yao, W. Chen, The Ohio State University, USA
`904.6 The Application Interface Model for Communication
`Software
`E. R. M. Madeira, M. de Jesus Mcndes, Universidade
`Estadual de Campinas, Brazil
`
`SESSION 905 Wednesday Afternoon (Palomar)
`ATM Traffic Performance
`ORGANIZERS: V. A. Bolotin, Bellcore, USA,
`J. G. Kap el, Bellcore (retired), USA
`CI-{AIRP RSON: S. Lederman, Bellcore, USA
`SPONSOR: Communications Switching
`905.1 Traffic Characteristics Evaluation of a Shared Buffer ATM
`Switch
`N. Endo, T. Ohuchi, T. Kozaki, H. Kuwahara, M. Mori,
`Hitachi Ltd., Japan
`905.2 Performance of Input-buffered and Output-buffered
`ATM Switches Under Bursty Traffic: Simulation Study
`S. Liew, Bellcore, USA
`905.3 Design and Performance Analysis of a Buffer Subsystem
`for the Batcher— Banyan Switch
`A. Leon-Garcia, P. S. Y. Lau, Univ. of Toronto, Canada
`905.4 A Heuristic Approach for Performance Analysis of ATM
`Systems
`K. Liao, L. G. Mason, INRS-Telecommunications, Canada
`905.5 Optimal Buffer Allocation for Packet Switches with Input
`and Output Queueing
`J. S. Chen, IBM Research Division, USA; T. E. Stern,
`Columbia University, USA
`905.6 Resource Allocation and Delay Constraints in ATM
`Networks
`P. Todorova, Research Center for Open Communication
`Systems (FOKUS), Germany (FR); D. Verma, University of
`California, Berkeley, USA
`905.7 Simulation and Performance Evaluation of an ATM
`Multiplexer Using Priority Scheduling
`D. l(. Hsing, Bellcore, USA
`
`

`
`SESSION 908 Wednesday Afternoon (Cuyamaca)
`Signal Processin in Digital RF Networks
`ORGANIZERS:
`. Chennakeshu, General Electric, USA
`I. B. Anderson, Rensselaer Polytechnic Institute, USA
`CHAIRPERSONS: S. Chennakeshu, G. I. Saulnier,
`ECSE Department, USA
`SPONSORS: Si nal Processing and Communication
`Electronics; Radgro Communication
`908.1 A DSP-Based Alternative to Direct Conversion Receivers
`for Digital Mobile Communications
`J. K. Cavers, S. I’. Stapleton, Simon Fraser University,
`Canada
`
`908.2 Nonlinear Median-Based Receiver for QAM Signals
`J. Astola, Y. Neuvo, University of Tampere, Finland
`908.3 A Decision-Feedback Equalizer with Selective Time-
`Reversal Operation for High-Rate Indoor Radio
`Communication
`3. Ariyavisitakul, Bellcore, USA
`908.4 A New Decision Feedback Equalization Method for
`Digital Microwave Radio
`I. Tsujimoto, NEC Corporation, Japan
`908.5 Real-Time ML Estimation of Very Frequency-Selective
`Multipath Channels
`J. de Weck, J. Ruprecht, Swiss I7I'I‘ General Directorate,
`Switzerland
`
`908.6 Experimental Results on Digital Decision Feedback
`Equalizers in a Multilevel QAM Digital Radio System
`M. Borgne, CNET, France
`908.7 Computation of Spectrally Efficient Pulse Shapes in
`Satellite Communication
`L. I’. Riddle, Westinghouse Electric Corp., USA
`
`SESSION 906 Wednesday Afternoon (Laguna)
`Mobile Radio — Techniques and Applications
`ORGANIZERS: H. Arnold, Bellcore, USA,
`A. Kucar, Bell Northern Research, Canada
`CHAIRPERSON: H. Arnold
`SPONSOR: Radio Communication
`
`906.1 Performance of Trellis-Coded 16QAM/TDMA System for
`Land Mobile Communications
`
`5. Sampei, Ministry of Posts & Telecom., Japan
`906.2 Linearized Saturation Amplifier with Bidirectional
`Control (LSA-BC) for Digital Mobile Radio
`K. Chiba, T. Nojima, S. Tomisato, NT1" Radio
`Communication Systems, Japan
`906.3 150K Gate General-Purpose High-Coding-Cain TCM
`VLSIs for High-Speed Multi- Level QAM Systems
`5. Aikawa, Y. Nakamura, H. Takanashi, N'I'I' Radio
`Communication Systems, Japan
`906.4 An Energy- and Bandwidth-Efficient Data Transmission
`System for 'I'ime- Selective Fading Channels
`A. Wittneben, ABB Corporate Research, Switzerland
`906.5 Local Intervehicle Radio Communication - Multiple
`Access and Error Control
`M. Dippold, A. Bottcher, E. Lutz, W. Schafer, Institute for
`Telecommunications, Germany (FR)
`906.6 Estimating Unreliable Packets in Subband Coded Speech
`W. C. Wong, National University of Singapore, Singapore;
`N. Seshadri, C. E. Sundberg, AT&T Bell Laboratories, USA
`906.7 Analysis of Packet Reservation Multiple Access: Voice
`Data Integration for Wireless Networks
`S. Nanda, Rutgers University, USA
`
`of Ottawa, Canada,
`
`SESSION 907 Wednesday Afternoon (Colombard)
`Combined Modulation and Coding
`ORGANIZERS: P. Galko, University
`F. Morales-Moreno, Telesat, Canada
`CHAIRPERSON: P. Galko
`SPONSOR: Communication Theory
`907.1 Signal Mapping and Nonlinear Encoder for Uniform
`Trellis Codes
`R. Buz, Queen's University, Canada
`907.2 Shaping Gains Using Non-Equiprobable Signaling
`J. N. Livingston, Texas A&M University, USA
`907.3 Efficient Design of Trellis-Coded CPFSK
`H. Yang, D. P. Taylor, McMaster University, Canada;
`F. Morales-Moreno, Telesat Canada, Canada
`907.4 A BISDN-Compatible Modern/Codec
`F. I-Iemmati, COMSAT Laboratories, USA
`907.5 Viterbi Receivers in the Presence of Severe Intersyrnbol
`Interference
`K. J. Kerpez, Bellcore, USA
`907.6 Trellis-Coded Partial Response Signaling for
`Power and Bandwidth Efficient Digital Transmission
`L. Lee, C. Wang, Y. Lee, National Taiwan Univ., Taiwan
`907.7 A New Tandem Source-Channel Coding Scheme
`M. R. Soleymani, McCill University, Canada
`
`xxvi
`
`

`
`Rate Controlled Servers for Very High-Speed Networks
`
`C. R. Kalrnanek and H. Kanakia
`
`AT&T Bell Laboratories
`Murray Hill, NJ 07974
`
`S. Keshav
`
`University of California, Berkeley
`Berkeley, CA 94720
`
`ABSTRACT
`
`Future high-speed networks are expected to carry
`traffic with a wide range of performance requirements.
`We describe two queue service disciplines, rate-based
`scheduling and hierarchical round robin scheduling,
`that allow some connections to receive guaranteed rate
`and jitter performance, while others receive best effort
`service. Rate-based scheduling is designed for fast
`packet networks, while hierarchical round robin is an
`extension of round robin scheduling suitable for use in
`networks based on the Asynchronous Transfer Mode
`(ATM) being defined in CCITT. Both schemes are
`feasible at rates of one Gigabit/sec. The schemes allow
`strict bounds on the buffer space required for rate con-
`trolled connections and can provide efiicient utilization
`of transmission bandwidth.
`
`Introduction
`Future high—speed networks are expected to carry
`traffic with a wide range of performance requirements.
`A classic tradeoff in network design is between provid-
`ing quality of service guarantees on one hand, and
`achieving efficient utilization of
`the transmission
`bandwidth on the other [TYM]. Synchronous circuit
`switching can provide guaranteed bandwidth and
`bounds on jitter at
`the expense of underutilizing
`bandwidth, whereas datagram networks can optimally
`utilize bandwidth but cannot provide strong perfor-
`mance guarantees.
`the design problem is
`In integrated networks,
`more difficult because different classes of traffic have
`dramatically
`different
`performance
`requirements,
`expressed in terms of the desired bandwidth, latency,
`and jitter} Real-time video requires high bandwidth,
`low latency and low jitter. Voice traffic requires low
`bandwidth,
`low latency and low jitter. Computer
`traffic spans a wide range of requirements, from appli-
`cations requiring low latency for small
`transfers to
`those which are quite insensitive to the bandwidth,
`latency, and jitter achieved.
`In addition, the operating
`range of future networks will be very large: we expect
`1 Defined below.
`
`trunk speeds from 64 Kb/sec to over 1 Gb/sec. The bur-
`den on the network designer is that where there is con-
`tention for limited transmission bandwidth or buffer
`space, control mechanisms must ensure that each
`traffic class meets its performance objectives.
`
`Rate Control
`We propose to separate the virtual circuits or
`connections
`provided
`by
`the
`network
`into two
`categories:
`rate controlled connections and best effort
`connections.
`In the literature,
`the use of the term
`"rate control" is not standardized. We use the term
`“rate control" to mean controlling both the rate and
`jitter of a connection. At call setup time, a user
`requesting a rate controlled connection specifies a
`desired average service rate and jitter bound.
`If the
`call
`is accepted,
`the network provides guaranteed
`bandwidth and jitter bounds, and there may be
`mechanisms for renegotiation if the network cannot
`provide a suitable connection. Best effort connections,
`on the other hand, do not imply (strong) performance
`guarantees.
`As in [ZHA], the rate of a connection is defined
`over an averaging interval:
`the rate is determined by
`dividing the total number of bytes transmitted during
`an averaging interval by the averaging interval. The
`value of the averaging interval will determine the
`allowed burstiness at the source.
`If the traffic source
`uses a large averaging interval, the network must have
`sufficient buffer space at the first node to accomodate a
`burst.
`
`Our notion of rate control also specifies a bound
`on the jitter introduced by a server. We first present
`an intuitive definition of jitter in order to contrast it
`with the definition that we will use. A server is
`assumed to introduce delay as packets pass through it:
`the delay is modeled as having a constant component
`and a variable component. The jitter introduced by a
`server is computed by measuring the packet departure
`time from the server and subtracting the constant com-
`ponent of the delay from the departure time for each
`packet
`to give the normalized departure time. The
`jitter for a connection is the maximum difference over
`
`0012
`
`300.3.1
`CH2827-4/90/0000-0012 $1.00 © 1990 IEEE
`
`

`
`Page 2 of9
`
`Kalmanek, Kanakia and Keshav
`
`all packets between the normalized departure time and
`the arrival time at the server. This definition is useful
`in analytic modeling of the server.
`We propose an alternative definition of jitter that
`we believe is more relevant than the above definition
`in systems design. Jitter is defined to be a short term
`average rate, where the averaging interval used in
`computing the jitter is different from the averaging
`interval used in computing the rate of a connection. We
`refer to the two averaging intervals as the rate averag-
`ing
`interval,
`and
`the
`jitter
`averaging
`interval.
`Specifically, the jitter is the maximum number of pack-
`ets transmitted by the server in the jitter averaging
`interval. This definition captures a useful property of
`our rate controlled servers, namely that they smooth
`the output stream. In principle, either averaging inter-
`val can range from the time to send one packet to the
`total connection time.
`In practice, the rate averaging
`interval might be on the order of several round trip
`times. The jitter averaging interval would be smaller
`and would depend on the size of the receive buffer at
`the destination.
`Two alternative measures of quality for real-time
`streams have been suggested in recent work [FER],
`[LL]. These two alternative measures are the source-
`to—destination delay bound for a packet and the vari-
`ance in inter-packet gaps observed at a receiver.
`In
`transporting real-time video or voice, bounds on abso-
`lute delay for a packet are not sufficient unless the
`jitter is also controlled. The absolute delay time for a
`packet in transit seems important mainly for applica-
`tions such as real-time feedback and control systems.
`The variance in inter-packet gaps would affect quality
`only when the receiver equipment does not have the
`intelligence and the memory required to smooth out
`the variation in inter—packet gaps.
`In an era of inex-
`pensive microprocessors and low cost memory devices,
`it
`seems likely that
`receiver equipment would be
`designed with the capability to smooth interpacket
`packet gaps for a burst of packets. The real factor
`affecting the cost would be the length of the burst that
`has to be stored before decoding the data, and it is this
`burst length which is captured in our definition of
`jitter. The distinction between our definition of rate
`control and these alternative definitions is important
`since we have found that neither the bounds on the
`delay in a network nor the variance in interpacket
`gaps observed at a receiver comes cheaply, especially
`for networks without a synchronous time base. How-
`ever, we contend that controlling the jitter and rate
`can be achieved at an acceptable cost.
`
`Traffic Categories
`In the paper, we use the following classification
`of sources of packet traffic. A source that injects pack-
`ets into the network at fixed time intervals is referred
`to
`as
`a Continuous Bit Rate
`(CBR)
`source.
`Uncompressed video or voice sources are CBR sources
`of packets. Sources which use compression or adaptive
`
`techniques generate packets at variable bit rates. We
`refer to these types of sources as Variable Bit Rate
`(VBR) sources. A simple model for a VBR source has
`an "on-off"
`traffic pattern:
`the source sends some
`number of packets at its peak rate for a period of time,
`and is quiet for a period of time. We expect that CBR
`and VBR sources would use rate—controlled connections.
`Sources with large amounts of data to transfer, such as
`file transfer protocols, can simulate either a CBR
`source or VBR source by using rate-based admission
`control.
`It should be noted that use of rate control
`within a network is distinct from the admission control
`scheme used. A rate-controlled network could support
`either rate-based or window-based admission control
`policies.
`A source that sends short messages at widely
`spaced intervals,
`is referred to as an intermittent
`source. Distributed computations use interprocess con-
`trol messages which are typically small and hence are
`examples of intermittent sources. For an intermittent
`source, the measure of importance is low end-to-end
`delay and small probability of data loss. We expect
`that intermittent sources would use the best effort ser-
`vice.
`
`Goals of Work
`The goal of the rate control proposed in this
`paper is to guarantee the rate and provide an upper
`bound on the jitter for the rate—controlled connections,
`while also providing the lowest possible delay and loss
`probability for best-effort connections. An equally
`important goal for the network provider is the efficient
`use of network resources: memory and bandwidth.
`In
`this paper we propose two implementations of the pro-
`posed rate control scheme. The first implementation is
`focused on fast packet networks, and the scheduling of
`packets for service can be achieved by a microprocessor
`at very high speeds using an algorithm known as rate-
`based scheduling. The second implementation focuses
`on networks based on Asynchronous Transfer Mode
`(ATM), where messages are broken up and carried in
`small, fixed length cells that are switched and multi-
`plexed in a uniform manner [CCITT]. This implemen-
`tation, known as Hierarchical Round Robin Scheduling
`(HRR), yields the desired performance guarantees by
`an enhancement of ordinary round robin scheduling.
`
`Previous Work
`In a packet switched network, the main point of
`control is the queuing di