EURASIP Journal on Applied Signal Processing
Advanced Signal Processing for Digital Subscriber Lines
Guest Editors: Raphael Cendrillon, Iain Collings, Tomas Nordström, Frank Sjöberg, Michail Tsatsanis, and Wei Yu
EURASIP Journal on Applied Signal Processing Advanced Signal Processing for Digital Subscriber Lines
EURASIP Journal on Applied Signal Processing Advanced Signal Processing for Digital Subscriber Lines
Guest Editors: Raphael Cendrillon, Iain Collings, Tomas Nordström, Frank Sjöberg, Michail Tsatsanis, and Wei Yu
Copyright © 2006 Hindawi Publishing Corporation. All rights reserved.
This is a special issue published in volume 2006 of “EURASIP Journal on Applied Signal Processing.” All articles are open access articles distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Editor-in-Chief Ali H. Sayed, University of California, USA
Associate Editors Kenneth Barner, USA Søren Holdt Jensen, Denmark Vitor H. Nascimento, Brazil Mauro Barni, Italy Mark Kahrs, USA Sven Nordholm , Australia Richard Barton, USA Thomas Kaiser, Germany Douglas O’Shaughnessy, Canada Ati Baskurt, France Moon Gi Kang, South Korea Montse Pardas, Spain Kostas Berberidis, Greece Matti Karjalainen, Finland Wilfried Philips, Belgium Jose C. Bermudez, Brazil Walter Kellermann, Germany Vincent Poor, USA Enis Cetin, Turkey Joerg Kliewer, USA Ioannis Psaromiligkos, Canada Jonathon Chambers, UK Lisimachos P. Kondi, USA Phillip Regalia, France Benoit Champagne, Canada Alex Kot, Singapore Markus Rupp, Austria Joe Chen, USA Vikram Krishnamurthy, Canada Bill Sandham, UK Liang-Gee Chen, Taiwan Tan Lee, Hong Kong Bulent Sankur, Turkey Huaiyu Dai, USA Geert Leus, The Netherlands Erchin Serpedin, USA Satya Dharanipragada, USA Bernard C. Levy, USA Dirk Slock, France Frank Ehlers, Italy Ta-Hsin Li, USA Yap-Peng Tan, Singapore Sharon Gannot, Israel Mark Liao, Taiwan Dimitrios Tzovaras, Greece Fulvio Gini, Italy Yuan-Pei Lin, Taiwan Hugo Van hamme, Belgium Irene Gu, Sweden Shoji Makino, Japan Bernhard Wess, Austria Peter Handel, Sweden Stephen Marshall, UK Douglas Williams, USA R. Heusdens, The Netherlands C. Mecklenbräuker, Austria Roger Woods, UK Ulrich Heute, Germany Gloria Menegaz, Italy Jar-Ferr Yang, Taiwan Arden Huang, USA Ricardo Merched, Brazil Abdelhak M. Zoubir, Germany Jiri Jan, Czech Republic Rafael Molina, Spain Sudharman K. Jayaweera, USA Marc Moonen, Belgium
Contents
Advanced Signal Processing for Digital Subscriber Lines, Raphael Cendrillon, Iain Collings, Tomas Nordström, Frank Sjöberg, Michail Tsatsanis, and Wei Yu Volume 2006 (2006), Article ID 32476, 3 pages
The Worst-Case Interference in DSL Systems Employing Dynamic Spectrum Management, Mark H. Brady and John M. Cioffi Volume 2006 (2006), Article ID 78524, 11 pages
Joint Multiuser Detection and Optimal Spectrum Balancing for Digital Subscriber Lines, Vincent M. K. Chan and Wei Yu Volume 2006 (2006), Article ID 80941, 13 pages
Spectrally Compatible Iterative Water Filling, Jan Verlinden, Etienne Van den Bogaert, Tom Bostoen, Francesca Zanier, Marco Luise, Raphael Cendrillon, and Marc Moonen Volume 2006 (2006), Article ID 58380, 10 pages
The Normalized-Rate Iterative Algorithm: A Practical Dynamic Spectrum Management Method for DSL, Driton Statovci, Tomas Nordström, and Rickard Nilsson Volume 2006 (2006), Article ID 95175, 17 pages
ADSL Transceivers Applying DSM and Their Nonstationary Noise Robustness, Etienne Van den Bogaert, Tom Bostoen, Jan Verlinden, Raphael Cendrillon, and Marc Moonen Volume 2006 (2006), Article ID 67686, 8 pages
Analysis of Iterative Waterfilling Algorithm for Multiuser Power Control in Digital Subscriber Lines, Zhi-Quan Luo and Jong-Shi Pang Volume 2006 (2006), Article ID 24012, 10 pages
Alien Crosstalk Cancellation for Multipair Digital Subscriber Line Systems, George Ginis and Chia-Ning Peng Volume 2006 (2006), Article ID 16828, 12 pages
Crosstalk Models for Short VDSL2 Lines from Measured 30 MHz Data, E. Karipidis, N. Sidiropoulos, A. Leshem, Li Youming, R. Tarafi, and M. Ouzzif Volume 2006 (2006), Article ID 85859, 9 pages
Error Sign Feedback as an Alternative to Pilots for the Tracking of FEXT Transfer Functions in Downstream VDSL, J. Louveaux and A.-J. van der Veen Volume 2006 (2006), Article ID 94105, 14 pages
Iterative Refinement Methods for Time-Domain Equalizer Design, Güner Arslan, Biao Lu, Lloyd D. Clark, and Brian L. Evans Volume 2006 (2006), Article ID 43154, 12 pages
Near-Capacity Coding for Discrete Multitone Systems with Impulse Noise, Masoud Ardakani, Frank R. Kschischang, and Wei Yu Volume 2006 (2006), Article ID 98738, 10 pages Fine-Granularity Loading Schemes Using Adaptive Reed-Solomon Coding for xDSL-DMT Systems, Saswat Panigrahi and Tho Le-Ngoc Volume 2006 (2006), Article ID 65716, 13 pages
Intra-Symbol Windowing for Egress Reduction in DMT Transmitters, Gert Cuypers, Koen Vanbleu, Geert Ysebaert, and Marc Moonen Volume 2006 (2006), Article ID 70387, 9 pages
Designing Tone Reservation PAR Reduction, Niklas Andgart, Per Ödling, Albin Johansson, and Per Ola Börjesson Volume 2006 (2006), Article ID 38237, 14 pages
Cosine-Modulated Multitone for Very-High-Speed Digital Subscriber Lines, Lekun Lin and Behrouz Farhang-Boroujeny Volume 2006 (2006), Article ID 19329, 16 pages Hindawi Publishing Corporation EURASIP Journal on Applied Signal Processing Volume 2006, Article ID 32476, Pages 1–3 DOI 10.1155/ASP/2006/32476
Editorial Advanced Signal Processing for Digital Subscriber Lines
Raphael Cendrillon,1 Iain Collings,2 Tomas Nordstrom,¨ 3 Frank Sjoberg,¨ 4 Michail Tsatsanis,5 and Wei Yu6
1 Marvell Hong Kong Ltd., Hong Kong 2 CSIRO Information Communication Technologies Center, Australia 3 Telecommunications Research Center Vienna (ftw.), Donau-City-StraBe 1, 1220 Vienna, Austria 4 Division of Signal Processing, Lulea˚ University of Technology, and Upzide Labs, Lulea,˚ Sweden 5 Aktino Inc., Irvine, California, USA 6 Electrical and Computer Engineering Department, University of Toronto, 10 King’s College Road, Toronto, ON Canada, M5S 3G4
Received 27 January 2006; Accepted 27 January 2006 Copyright © 2006 Raphael Cendrillon et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The recent deployment of digital subscriber line (DSL) tech- systems by adaptively varying transmit power-spectral den- nology around the world is rapidly making broadband access sity according to geographic locations and the crosstalk chan- for the mass consumer market a reality. The ever-growing nel characteristics of the subscribers in each bundle. This is- customer demand for higher data rates has been fueled by sue contains six papers on DSM. In “The worst-case inter- the popularity of applications like peer-to-peer (P2P) file- ference in DSL systems employing dynamic spectral man- sharing networks and video-streaming and high-definition agement,” Brady and Cioffi answer the question of what the television (HDTV). DSL technology allows telephone opera- worst-case crosstalk interference is for a given DSL line. They tors to get maximum leverage out of their existing infrastruc- characterize the performance of the system under the worst- ture by delivering broadband access over existing twisted- case noise using a game theory technique. In “Joint mul- pair telephone lines. At the heart of DSL lies a plethora of tiuser detection and optimal spectrum balancing for digital signal processing techniques which enable such high-speed subscriber lines,” Chan and Yu study the optimal spectrum transmission to be achieved over a medium originally de- management technique for a scenario in which crosstalk signed with only voice-band transmission in mind. These may also be partially cancelled using advanced crosstalk advanced signal processing techniques address many chal- cancellation techniques. In the next three papers, practi- lenges that exist in DSL networks today, such as the near-end cal spectrum management techniques are investigated. In and far-end crosstalk (NEXT/FEXT), impulse noise, peak-to- “Spectrally compatible iterative water-filling,” Verlinden et average-power ratio (PAR), intersymbol and intercarrier in- al. study a system in which spectral allocation scheme is terference (ISI/ICI), radio-frequency interference (RFI), and constrained by additional spectrum compatibility require- so forth. The goal of this special issue is to discuss the state- ments and propose a new scheme based on an earlier al- of-the-art and recent advances in signal processing tech- gorithm called iterative water-filling. In “The normalized- niques for DSL. rate iterative algorithm: a practical dynamic spectrum man- The special issue consists of fifteen papers on a range of agement method for DSL,” Statovci et al. propose a new topics. The first set of papers focuses on the area of dynamic low-complexity technique for spectrum balancing and fre- spectrum management (DSM). In a conventional DSL de- quency partition in a DSL bundle. In “ADSL transceivers ployment, the transmit spectrum for all modems in a bun- applying DSM and their nonstationary noise robustness,” dle are fixed to a predetermined level. As DSL deployment Van den Bogaert et al. report the performance of practical becomes increasingly heterogeneous, crosstalk produced by transceivers implementing dynamic spectrum management modems under a fixed spectrum can be a source of sig- and study their robustness against nonstationary noise. Fi- nificant interference. Dynamic spectrum management aims nally, from a theoretical perspective, the paper “Analysis of to improve the data rates and reaches of conventional DSL iterative water-filling algorithm for multiuser power control 2 EURASIP Journal on Applied Signal Processing in digital subscriber lines,” by Luo and Pang, takes a new Finally, we wish to take this opportunity to acknowledge look at the iterative water-filling algorithm and gives a novel and to thank all anonymous reviewers, without whom the interpretation of the algorithm based on optimization the- success of this special issue would not have been possible. ory. The next paper in the special issue deals with crosstalk cancellation. In a DSL deployment, when coordination Raphael Cendrillon among the transmit- or receive-modems is possible, further Iain Collings data improvement may be obtained via crosstalk cancella- Tomas Nordstrom¨ tion. In the paper “Alien crosstalk cancellation for multipair Frank Sjoberg¨ digital subscriber line systems,” Ginis and Peng give an Michail Tsatsanis overview of this area and propose a new crosstalk cancella- Wei Yu tion technique that takes advantage of the noise correlation among the multiple receivers. The practical success of dynamic spectrum management and crosstalk cancellation depends very much on how accu- Raphael Cendrillon was born in Mel- rately crosstalk channels may be modeled and identified in bourne, Australia, in 1978. He received an practice. Two papers of this special issue address this area. Electrical Engineering degree (honours first In “Crosstalk models for short VDSL2 lines from measured class) from the University of Queensland, 30 MHz data,” Karipidis et al. propose measurement-based Australia, in 1999, and a Ph.D. in electri- crosstalk models for VDSL. In “Error sign feedback as an al- cal engineering at the Katholieke Univer- ternative to pilots for the tracking of FEXT transfer functions siteit Leuven, Belgium, in 2004. His Ph.D. in downstream VDSL,” Louveaux and Van der Veen propose was awarded Summa Cum Laude with con- new ways of identifying the crosstalk channel using a novel gratulations of the jury, an honor given to feedback scheme. the top 5% of Ph.D. graduates. His research Equalization and coding continue to be important re- focuses on the application of multiuser communication theory to xDSL. In 2002, he was a Visiting Scholar at the Information Systems search topics in DSL. In the area of time-domain equal- Laboratory, Stanford University, with Prof. John Cioffi. In 2005, ization (TEQ), the paper “Iterative refinement methods Dr. Cendrillon was a postdoctoral Research Fellow at the Univer- for time-domain equalizer design” by Arslan et al. pro- sity of Queensland, Australia. During this period, he was also a poses a new method to reduce the implementation com- Visiting Research Fellow at Princeton University with Prof. Mung plexity of the TEQ. In the area of error-correcting coding Chiang. He is now a senior DSP engineer with Marvell Hong Kong for the DSL system, the paper “Near capacity coding for Ltd. He was awarded the Alcatel Bell Scientific Prize in 2004; IEEE discrete multitone (DMT) systems with impulse noise” by Travel Grants in 2003, 2004, and 2005; the K.U. Leuven Bursary for Ardakani et al. proposes a methodology for the design of Advanced Foreign Scholars in 2004; and the UniQuest Trailblazer the newly emerged low-density parity-check (LDPC) codes Prize for Commercialization in 2005. for a DMT system, while addressing the practical DSL de- ployment issue of impulse noise. In “Fine-granularity load- Iain Collings was born in Melbourne, Aus- ing schemes using adaptive Reed-Solomon coding for xDSL- tralia, in 1970. He received the B.E. de- DMT systems,” Panigrahi and Le-Ngoc propose a joint de- gree in electrical and electronic engineering sign of bit-loading and error-correcting code, and charac- from the University of Melbourne, in 1992, terize the performance gain made possible by fractional bit- and the Ph.D. degree in systems engineering loading. from the Australian National University, in The final set of three papers in this special issue deals 1995. Currently he is a Science Leader in the CSIRO Information Communication Tech- with the area of modulation and transmitter design. The nologies Centre, Australia. Prior to this he design of transmit window to minimize egress is studied was an Associate Professor at the Univer- in the paper by Cuypers et al. “Intra-symbol windowing sity of Sydney (1999–2005); a Lecturer at the University of Mel- for egress reduction in DMT transmitters.” The peak-to- bourne (1996–1999); and a Research Fellow in the Australian Co- average-power ratio is another important transmitter de- operative Research Centre for Sensor Signal and Information Pro- sign issue for DMT systems. This is taken up in the pa- cessing (1995). His current research interests include mobile dig- per “Designing tone reservation PAR reduction” by Andgart ital communications and broadband digital subscriber line com- et al. DMT is not the only possible multicarrier modula- munications; more specifically, synchronization, channel estima- tion scheme for DSL. An alternative is proposed and stud- tion, equalization, and multicarrier modulation, for time-varying ied in the paper “Cosine modulated multitone for very and frequency-selective channels. He currently serves as an Editor for the IEEE Transactions on Wireless Communications, and as a high-speed digital subscriber lines” by Lin and Farhang- Guest Editor for the EURASIP Journal on Advanced Signal Process- Boroujeny. ing. He has also served as the Vice Chair of the Technical Program The continued growth of digital subscriber line technol- Committee for IEEE Vehicular Technology Conf. (Spring) 2006, as ogy worldwide is in part fueled by rapid advances in signal well as serving on a number of other TPCs and organizing commit- processing techniques. We hope that the readers will enjoy tees of international conferences. He is also a founding organizer the collection of papers on this timely topic. of the Australian Communication Theory Workshops 2000–2006. Raphael Cendrillon et al. 3
Tomas Nordstrom¨ was born in Harn¨ osand,¨ Wei Yu received the B.A.S. degree in com- Sweden, in 1963. He received the M.S.E.E. puter engineering and mathematics from degree in 1988, the Licentiate degree in the University of Waterloo, Waterloo, ON, 1991, and the Ph.D. degree in 1995, all Canada, in 1997, and M.S. and Ph.D. de- from Lulea˚ University of Technology, Swe- grees in electrical engineering from Stan- den. Currently, he is a key researcher and ford University, Stanford, Calif, USA, in project manager at the Telecommunications 1998 and 2002, respectively. Since 2002, Research Center Vienna (ftw). During 1995 he has been an Assistant Professor with and 1996, he was an Assistant Professor the Electrical and Computer Engineering at Lulea˚ University of Technology research- Department at the University of Toronto, ing computer architectures, neural networks, and signal process- Toronto, ON, Canada, where he also holds a Canada Research ing. Between 1996 and 1999, he was with Telia Research (the Chair. His main research interests include multiuser information research branch of the Swedish incumbent telephone operator), theory, coding, optimization, wireline communications, and wire- where he developed broadband Internet communication over less broadband access networks. He is currently an Associate Editor twisted copper pairs. He was instrumental in the development for the IEEE Transactions on Wireless Communications. of the Zipper-VDSL concept (contributed to the standardization of VDSL in ETSI, ANSI, and ITU) and in the design of the Zipper-VDSL prototype modems. In December 1999, he joined ftw, where he is the project Manager of the “Broadband wire- line access” group. At ftw he has worked with various aspects of wireline communications like simulation of xDSL systems, ca- ble measurements, RFI suppression, and exploiting the common- mode signal in xDSL. Currently his research is focused on “ac- tive copper resource management” which includes researching op- timized usage of cable bundles and dynamic spectrum manage- ment.
Frank Sjoberg¨ wasborninUme,Sweden,in 1970. He received the M.S. degree in com- puter science and the Ph.D. degree in signal processing from Lulea˚ University of Tech- nology, Lulea,˚ Sweden, in 1995 and 2000, re- spectively. Between 2000 and 2002, he was with Telia Research AB, Lulea,˚ Sweden. He currently works as a Researcher at Upzide Labs, and holds a position as an Assistant Professor in the Division of Signal Process- ing, Lulea˚ University of Technology. His primary research interests are statistical signal processing and digital communication, with emphasis on wireline systems.
Michail Tsatsanis is a founder and Chief Scientist of Aktino, a company developing next generation DSL transceiver technol- ogy. Prior to that he was with Voyan Tech- nology, where he served as Chief Scientist and Chief Technical Officer. From 1995 to 2000, he was with Stevens Institute of Tech- nology, NJ, were he served as an Associate Professor in Electrical Engineering. He is the author of more than 80 peer reviewed papers, three book chapters, and several patents. At Aktino he is leading a technology team that was first to successfully im- plement and produce a MIMO vectored transceiver in the DSL space. He has received a number of distinctions including the Na- tional Science Foundation CAREER Award and two IEEE Best Pa- per Awards. He has served the IEEE in various capacities includ- ing the position of Associate Editor for two IEEE Transactions and Chair of workshop organizing committees. He holds M.S. and Ph.D. degrees in electrical engineering from the University of Vir- ginia. Hindawi Publishing Corporation EURASIP Journal on Applied Signal Processing Volume 2006, Article ID 78524, Pages 1–11 DOI 10.1155/ASP/2006/78524
The Worst-Case Interference in DSL Systems Employing Dynamic Spectrum Management
Mark H. Brady and John M. Cioffi
Department of Electrical Engineering, Stanford University, Stanford, CA 94305-9515, USA
Received 1 December 2004; Revised 28 July 2005; Accepted 31 July 2005 Dynamic spectrum management (DSM) has been proposed to achieve next-generation rates on digital subscriber lines (DSL). Be- cause the copper twisted-pair plant is an interference-constrained environment, the multiuser performance and spectral compati- bility of DSM schemes are of primary concern in such systems. While the analysis of multiuser interference has been standardized for current static spectrum-management (SSM) techniques, at present no corresponding standard DSM analysis has been estab- lished. This paper examines a multiuser spectrum-allocation problem and formulates a lower bound to the achievable rate of a DSL modem that is tight in the presence of the worst-case interference. A game-theoretic analysis shows that the rate-maximizing strategy under the worst-case interference (WCI) in the DSM setting corresponds to a Nash equilibrium in pure strategies of a certain strictly competitive game. A Nash equilibrium is shown to exist under very mild conditions, and the rate-adaptive waterfill- ing algorithm is demonstrated to give the optimal strategy in response to the WCI under a frequency-division (FDM) condition. Numerical results are presented for two important scenarios: an upstream VDSL deployment exhibiting the near-far effect, and an ADSL RT deployment with long CO lines. The results show that the performance improvement of DSM over SSM techniques in these channels can be preserved by appropriate distributed power control, even in worst-case interference environments.
Copyright © 2006 M. H. Brady and J. M. Cioffi. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
1. INTRODUCTION talk scenario. Such methods are useful when a reasonable es- timate of spectrum of all users can be assumed priori. How- In recent years, increased demands on data rates and compe- ever, if spectrum is instead allocated dynamically, not only is tition from other services have led to the development of new this knowledge not available priori, but also because of loop high-speed transmission standards for digital subscriber line unbundling, other users’ spectrum may not even be known (DSL) modems. Dynamic spectrum management (DSM) is even during operation. Spectral compatibility between dif- emerging as a key component in next-generation DSL stan- ferent operators using DSM is a primary concern because dards. In DSM, spectrum is allocated adaptively in response new pathologies may arise with adaptive operation. More- to channel and interference conditions, allowing mitigation over, it is not unreasonable to suspect that each competing of interference and best use of the channel. As multiuser in- service provider sharing a binder would perform DSM in a terference is the primary limiting factor to DSL performance, greedy fashion, at the possible expense of other providers’ the potential for rate improvement by exploiting its structure users. However, in DSM, a worst-case interference analysis is substantial. based on maximum allowable PSDs is overly pessimistic, so DSM contrasts with current DSL practice, known as existing spectral compatibility techniques cannot be fruit- static spectrum management (SSM). In SSM, masks are fully employed. A new paradigm is needed to assess the im- imposed on transmit power spectrum densities (PSDs) to pact of DSM on multiuser performance of the overall system. bound the amount of crosstalk induced in other lines shar- ing the same binder group [1]. As SSM masks are fixed for 1.1. Prior results all loop configurations, they can often be far from optimal or even prudent spectrum usage in typical deployments. Stan- The capacity region of the AWGN interference channel (IC) dardized tests for “spectral compatibility” [1] assess “new is in general unknown, even for the 2-user case [2]. Com- technology” by defining PSD masks and examining the im- munication in the presence of hostile interference has been pact on standardized systems using the 99th-percentile cross- studied from a game-theoretic perspective in numerous 2 EURASIP Journal on Applied Signal Processing
SMC Downstream User 1 Upstream ......
L
Victim NEXT FEXT
Figure 1: Illustration of loop plant environment showing downstream FEXT and NEXT from user 1. The victim user is shown at the bottom. applications, for example, [3, 4]. A simple and relevant IC a game-theoretic viewpoint. Certain properties of the Nash achievable region is that attained by treating interference as equilibrium of this game are explored. Section 4 considers noise [5]. Capacity results for frequency-selective interfer- numerical examples in VDSL and ADSL systems. Conclud- ence channels satisfying the strong interference condition are ing remarks are made in Section 5. also known [6]. A word on notation: vectors are written in boldface, DSM algorithms have been proposed for the cases of dis- where vk denotes the kth element of the vector v,andv 0 tributed and centralized control scenarios. This paper con- denotes that each element is nonnegative. The notation v(n) siders what has been termed “Level 0–2 DSM” [7], wherein denotes a vector corresponding to tone n. For the symmetric cooperation may be allowed to manage spectrum, but not matrix X, X 0 denotes that X is positive semidefinite. 1 for multiuser encoding and decoding. A centralized DSM is a column vector with each element equal to 1. int(X)de- center controlling multiple lines offers both higher poten- notes the (topological) interior, cl(X) the closure, and ∂X the tial performance and improved management capabilities [8]. boundary of the set X. Distributed DSM schemes based on the iterative waterfilling (IW) algorithm [9] have been presented. IW has also been 2. SYSTEM MODEL studied from a game-theoretic viewpoint [10]. Numerous al- gorithms for centralized DSM have been proposed. Reference 2.1. Channel model [11] presents a technique to maximize users’ weighted sum- A copper twisted-pair DSL binder is modelled as a frequency- rate. Rate maximization subject to frequency-division and selective multiuser Gaussian interference channel [9, 23]. fixed-rate proportions between users has been considered The binder contains a total of L + 1 twisted pairs, with one [12]. Optimal [13] and suboptimal [14] algorithms to mini- DSL line per twisted pair, as shown in Figure 1. The effect of mize transmit power have been studied. NEXT and FEXT interferences generated by L “interfering” An extensive suite of literature on upstream power- users that generate crosstalk into one “victim” user is consid- ff backo techniques to mitigate the “near-far” problem has ered. This coupling is illustrated for downstream transmis- been developed for static spectrum-management systems sion in Figure 1. [13, 15–17]. A power-backoff algorithm for DSM systems implementing iterative waterfilling has been proposed [18]. 2.2. DSL modem model In current DSL standards, upstream and downstream transmissions use either distinct frequency bands or shared 2.2.1. Modem architecture bands. In the latter case, “echo” is created between upstream and downstream transmissions [9]. As analog hybrid circuits The standardized [24] discrete-multitone (DMT)-based do not provide sufficient isolation, echo mitigation is essen- modulation scheme is employed, so that transmission over tial in practical systems [19]. Numerous echo-cancellation the frequency-selective channel may be decoupled into N in- structures have been proposed for DSL transceivers [20–22]. dependent subcarriers or tones. Both FDM and overlapping bandplans are considered. As overlapping bandplans require 1.2. Outline echo cancellation that is imperfect in practice, error that is introduced acts as a form of interference and is of concern. This paper formulates the achievable rate of a single “victim” Echo-cancellation error is modelled presuming a prevalent modem in the presence of the worst-case interference from echo-cancellation structure utilizing a joint time-frequency other interfering lines in the same binder group. The perfor- LMS algorithm [19]isemployed.1 Using the terminology mance under the WCI is a guaranteed-achievable rate that of [19], let μ denote the LMS adaptive step size parameter. can be used, for example, in studying multiuser performance The“excessMSE”foragiventoneismodelled[25,equation of DSM strategies and establishing spectral compatibility of DSM systems. Section 2 defines the channel and system models. The 1 Other models may be more applicable to different echo-cancellation WCI problem is formalized and studied in Section 3 from structures. M. H. Brady and J. M. Cioffi 3
(12.74)] as proportional to the product of the LMS adaptive in ADSL and VDSL standards [9], this represents the impor- step size parameter μ and the transmit power on that tone. tant special case of the preceding model, where βn = 0(due (n) = ≤ ≤ The constant of proportionality is absorbed by defining β as to no echo cancellation) and hl 0foralln, L +1 l 2L the ratio of excess MSE to transmitted energy on a given tone. (due to frequency division). Additional technical results will be shown to hold in the FDM setting, as detailed in Section 3. 2.2.2. Achievable rate region 3. THE WORST-CASE INTERFERENCE This section discusses an achievable rate region for a DSL modem based on the preceding channel and system model. 3.1. Game-theoretic characterization of the WCI The following analysis applies to both upstream and down- stream transmissions. For specificity, the following refers to This section introduces and motivates the concept of the downstream transmission: first, consider the case where echo worst-case interference (WCI). Suppose that a “victim” mo- cancellation is employed. Denote the victim modem’s down- dem desires to keep its data rate at some level. Such a scenario stream transmit power on tone n, n ∈{1, ..., N},asxn.Let is commonplace as carriers widely offer DSL service at fixed ∈{ } (n) ∈ R2L element l, l 1, ..., L , of the vector y + denote the data rates. The objective is to bound the impact that mul- downstream transmit power of interfering modem l on tone tiuser interference can have on this victim modem, thereby n. Similarly, let element l, l ∈{L +1,...,2L},ofy(n) denote determining whether service may be guaranteed. To this end, the upstream transmit power of interfering user l − L.Define one considers interferences that are the most harmful in the ∈{ } (n) ∈ R2L element l, l l, ..., L , of the row vector h + as the sense of minimizing the achievable rate of a “victim” modem. FEXT power gain from interfering user l on tone n (necessar- However, it is not clear what form such interferences might ily, h(n) 0). Similarly, define element l, l ∈{L +1,...,2L}, take, nor how they might be best responded to. of h(n) to be the NEXT power gain from interfering user l−L. Examining this problem from the standpoint of game ∈ RN Let element n of hn + denote the victim line’s insertion theory leads to substantial insight. Consider a worst-case in- gain on tone n (hn ≥ 0). terference game where one player jointly optimizes the spec- 2 trum of all the interfering modems, irrespective of the data Independent AWGN (thermal noise) with power σn > 0 rate they achieve in doing so, to cause the most deleteri- is present on tone n.Letβn denote the echo-cancellation ra- tio on tone n as described above. Echo-cancellation error is ous interference to the victim modem. Thus in this game, treated as AWGN. Let Γ denote the SNR gap-to-capacity [9]. all the interfering modems act as one player, while the vic- Then the following bit loading2 is achievable on tone n [9]: tim modem acts as the other player, with the channel and noise known to all. Although such an arrangement may ap- pear pathological, it will be shown numerically that such a hnxn bn = log 1+ . (1) situation is quite close to what occurs in certain loop topolo- Γ (n) (n) 2 h y + βxn + σn gies. Neither is assuming such coordination of the interferers unreasonable in practice as under “Level 2” DSM [7, 8], each Observe that if hn = 0, then it is necessarily the case that collocated carrier may individually coordinate its own lines, bn = 0, implying that tone n is never loaded. Thus, in the nor may collocated equipment be centrally controlled by a competing carrier. Channels may be estimated in the field, sequel, hn > 0foralln ∈{1, ..., N} is considered without loss of generality by removing those tones with zero direct approximated by standardized models [9], and in the future, potentially published by operators [26]. gain (h = 0). Defining α = Γ/h , β = Γβ /h ,andN = n n n n n n n A Nash equilibrium in this game may be interpreted as Γ 2 σn /hn, and substituting characterizing a worst-case interference as an optimal re- sponse (power-allocation policy) to it. The structure of the = xn Nash equilibrium lends insight into the problem as well as bn log 1+ (n) (n) ,(2) αnh y + βnxn + Nn suggesting techniques that may be implemented in practical systems. because Γ ≥ 1, it follows that αn ≥ 0, βn ≥ 0, and Nn > 0. 3.2. Formalization of the WCI game 2.2.3. Achievable rate region for FDM Consider the following two-player game: let Player 1 con- When an FDM scheme is employed, NEXT and echo can- trol the spectrum allocation of victim modem, and let Player cellation are eliminated because transmission and reception 2 control the spectrum allocations of all the interfering occur on distinct frequencies.3 As a common configuration modems. Referring again to downstream transmission for specificity, let the total (sum) downstream power of the vic- x x tim modem n xn be upper bounded by P , where 0