`
`THE IEEE INTERNATIONAL
`CONFERENCE ON COMMUNICATIONS
`
`Helsinki, Finland
`June 11 -14,2001
`
`Conferen
`Record
`
`+ IEEE
`
`~ ®
`
`IEEE
`COMMUNICATIONS
`SOCIETY
`
`d GLOBECOM
`
`DISH, Ex. 1020 - Page 1 of 12
`
`
`
`
`Copyright © 2001 by the In
`ctrical and Electronics Engineers, Inc.
`All rights reserved.
`
`Copyright and Reprint Permission
`Abstracting is permitted with credit to the source. Libraries are permitted to photocopy beyond the limit of
`U.S. copyright law for private use of patrons those articles in this volume that carry a code at the bottom of
`the first page, provided the per-copy fee indicated in the code is paid throught Copyright Clearance Center, 222
`Rosewood Drive, Danvers, MA 01923.
`For those other copying, reprint or republication permission, write to lEEE Copyrights Manager,
`Center, 445 Hoes Lane, PO. Box 1331, Piscataway, NJ 08855-1331.
`
`lEEE Operations
`
`M,‘mafia-Mm.»‘A
`
`lEEE Catalog Number
`
`01CH37240 (softbound)
`01CH3724OC (CD-ROM)
`
`ISBN
`
`0-7803-7097-1
`0—7803-7098—8 (Microfiche Edition)
`0-7803—7099—6 (CD—ROM)
`
`Library of Congress
`
`81—64954?
`
`
`
`Additional copies of this publication are available from
`
`lEEE Operations Center
`PO. Box 1331
`445 Hoes Lane
`Piscataway, NJ 08855-1331 USA
`
`1-800-678-IEEE
`1-732—981-1393
`1—732-981-9667 (FAX)
`ieee.or
`email: customer.services
`Internet: http://www.ieeebrg
`
`.mxmw,m..mms...“We-«Wflw
`
`
`
`
`
`
`|.mwsa~1~yswmm’mm~«wflmrflf‘tfluhvaw»“ween.“iam;.‘vaom.M
`
`DISH, Ex. 1020 - Page 2 of 12
`
`
`
`
`EOOIQQU)
`
`I:46)
`
`
`
`0.%-Q0#3736)5.00EEC)>031?)
`
`
`
`
`
`Volume 5 of 10
`
`UESDAY, JUNE 12, 2001
`13:30 - 17:00
`
`
`
`
`
`
`SESSION Ml33: MOBILE DATA
`
`Organizer: Kaveh Pahlavan, WPI, CWI
`Chair: Rajamani Ganesh, Verizon T
`
`
`gF GSM HSCSD AND GPRS ........................................................ 1330
`MI33.1 MEASURED PERFO
`
`amanen, Sonera Corp., Finland
`Jouni Korhonen, Olli Aalto, Andrei Gu
`MI33.2 TWO-STAGE DYNAMIC UPL‘INK CHANNEL ASSIGNMENT FOR GPRS ......................................... 1335
`Ying—Dar Lin, Yu-Ching Hsu, Mei-Yan Chiang, National Chiao Tung Univ., Taiwan
`MI33.3 MOBILE MULTICAST WITH ROUTING OPTIMIZATION FOR RECIPIENT MOBILITY......................... 1340
`Jiunn-Ru Lai, Wanjiun Liao, Ming-Yu Jiang, Chien-An Ke, National Taiwan Univ., Taiwan
`MI33.4 A WIRELESS NETWORK SERVICE PROVIDING WIRELESS CHANNEL INFORMATION
`FOR ADAPTIVE MOBILE APPLICATIONS, PART I:PROPOSAL .................................................................. 1345
`Byoung-Jo "J" Kim, AT&T Labs - Research, USA
`Ml33.5
`PERMISSIBLE THROUGHPUT NETWORK FEEDBACK FOR ADAPTIVE MULTIMEDIA
`IN AODV MANETS ............................................................................................................................. 1352
`Manthos Kazantzidis, Mario Gerla, UCLA, USA; SJ Lee, Hewlett—Packard Laboratories, USA
`MI33.6 A SELF-ORGANIZING APPROACH TO DATA FORWARDING IN LARGE-SCALE
`SENSOR NETWORKS ........................................................................................................................ 1357
`Jelena Mirkovic, UCLA, USA; Geeiha Priya Venkataramani, UCLA, USA; Songwu Lu, UCLA, USA; Lixia Zhang, UCLA, USA
`Ml33.7
`INTEGRATED RATE AND ERROR CONTROL IN VARIABLE SPREADING GAIN
`WCDMA SYSTEMS ............................................................................................................................ 1362
`Dong Kim, Univ. of Seoul, Korea; Ekram Hossain, Univ. of Manitoba, Canada; Vijay Bhargava, Univ. of Victoria, Canada
`MI33.8 HANDOFF OF APPLICATION SESSIONS ACROSS TIME AND SPACE .......................................... 1367
`Thomas Phan, Kaixin Xu, Richard Guy, Rajive Bagrodia, UCLA, USA
`MI33.9
`SYSTEMS ENGINEERING OF DATA SERVICES IN UMTS W-CDMA SYSTEMS .............................. 1373
`Kourosh Parsa, Golden Bridge Tech, USA; Saeed S. Ghassemzadeh, AT&T Labs, USA; Saied Kazeminejad, 880 Tech. Resources, USA
`
`51’1"
`
`SESSION G34: RADIO RESOURCE MANAGEMENT
`Organizer: Marco Chiani, University of Bologna, Italy
`___________________________._.—____—
`Chair: Roberto Verdone, CSITE-CNR, University of Bologna, Italy
`
`TIME SLOT ASSIGNMENT USING CELL PARTITIONING IN CDMA CELLULAR SYSTEMS ................. 1381
`G34.1
`Kohji Takeo, YRP Key Tech Research Laboratories, Japan
`G342 AUTONOMOUS CALL ADMISSION CONTROL WITH PRIORITIZED HANDOFF lN
`CELLULAR NETWORKS ..................................................................................................................... 1386
`El-Sayed EI-Alfy, Yu-Dong Yao, Harry Heffes, Stevens Inst. of Technology. USA
`G343 AN ADAPTIVE M-QAM MODULATOR FOR FIXED WIRELESS ATM NETWORKS .............................. 1391
`Abbas Mohammadi, AmirKabir Univ. (Tehran Polytechnic), Iran; Surinder Kumar, WaveCom Electronics Inc., Canada
`G34.4
`PERFORMANCE EVALUATION OF PERSISTENT ADMISSION CONTROL IN
`FREQUENCY-HOPPED SLOTTED RANDOM ACCESS NETWORKS .......................................................... 1396
`Katsumi Sakakibara, Tomohiro Katagiri, Hirokazu Suyama, Jiro Yamakita, Okayama Prefecture! Univ., Japan
`G34.5 CARRIER ASSIGNMENT ALGORITHMS IN WIRELESS BROADBAND NETWORKS WITH CHANNEL
`ADAPTATION .................................................................................................................................... 1301
`lordanis Koutsopoulos, Leandros Tassiulas, Univ. of Maryland at College Park, USA
`634.6 RESOURCE ALLOCATION AND SCHEDULING SCHEMES FOR WCDMA DOWNLINKS .................... 1406
`Rath Vannithamby, Elvino Sousa, Univ. of Toronto, Canada
`634.7
`PERFORMANCE OF 1XTREME SYSTEM FOR MIXED VOICE AND DATA ....................................... 1411
`Petteri Luukkanen, Zhigang Rong, Lin Ma, Nokia Research Center, USA
`G348 THE FPBA ALGORITHM WITH CONTROLLED CAPTURE ............................................................. 1416
`Mohamed Hadi Habaebi, Borhanuddin Mohd. Ali, Univ. Putra Malaysia, Malaysia
`
`xvii
`
`DISH, Ex. 1020 - Page 3 of 12
`
`DISH, Ex. 1020 - Page 3 of 12
`
`
`
`F10
`
`1330
`
`1335
`
`1340
`
`.1345
`
`1352
`
`1357
`
`1362
`
`1367
`
`1373
`
`1381
`
`1386
`
`1391
`
`1396
`
`1301
`
`1406
`
`1411
`
`1416
`
`xvii
`
`HMS
`
`Volume 5 of 10
`
`
`
`
`UESDAY, JUNE 12, 2001
`13:30 - 17:00
`
`
`
`SESSION 635: CDMA RECEIV
`
`Organizer: Kari-Pekka Estola, Nokia,
`
`Chair: Matti Latva-aho, University of I
`
`
`R 3GPP WCDMA DOWNLINK RECEIVERS WITH
`G35.1
`LMMSE CHIP EQUALI
`
`
`,
`CHANNEL CODING ..............
`.......................................................................................... 1421
`
`Peter DanNood, Paul Alexander, Ian Oppe
`, Southern Poro Communications, Australia
`6352 AN ADAPTIVE RECEIVER FOR DYNAMIC TRAFFIC DS/CDMA PACKET TRANSMISSION ................. 1426
`Zlhua GUO, Ben Letaief Khaled, Hong Kong Univ. of Science & Technology, Hong Kong
`G363 BLIND MULTIUSER DETECTOR FOR DS/CDMA CHANNELS BASED ON THE MODIFIED
`STOCHASTIC GRADIENT DESCENT ALGORITHM .......................................................................... 1431
`A. Mukherjee. Kah C Teh, E. Gunawan, Nanyang Technological Univ., Singapore
`G354 JOINT ESTIMATION OF TIME DELAYS AND DOA'S FOR DS-CDMA SYSTEM OVER
`MULTIPATH RAYLEIGH FADING CHANNELS........................................................................................ 1436
`Kun Wang, New Jersey Inst. of Technology, USA; Hongya Ge, New Jersey Inst. of Technology, USA
`G355 ADAPTIVE BEAMFORMING AND POWER CONTROL FOR DS-CDMA MOBILE RADIO
`COMMUNICATIONS........................................................................................................................... 1441
`Ying-Chang Liang, Francois Chin, Centre for Wireless Communications, Singapore; Alex Kot. Nanyang Tech. Univ., Singapore
`G356
`PERFORMANCE ANALYSIS OF FORWARD LINK DS-CDMA SYSTEMS USING
`RANDOM AND ORTHOGONAL SPREADING SEQUENCES...................................................................... 1446
`Ji-Woong Choi. Seoul National Univ., Korea; Yong-Hwan Lee, Seoul National Univ., Korea; Yong-Ho Kim, Korea Telecom, Korea
`635.7
`PULL-IN CAPABILITY OF A MODIFIED CODE TRACKING LOOP FOR DIRECT-SEQUENCE SPREAD-
`SPECTRUM COMMUNICATION ON A FREQUENCY-SELECTIVE FADING CHANNEL.................................. 1451
`Jia-Chin Lin, National Chi-Nan University, Taiwan
`635.8
`INTERFERENCE FREE DECISION FEEDBACK CHANNEL ESTIMATION TECHNIQUE
`FOR M-ARY ORTHOGONAL DS/CDMA SYSTEMS.................................................................................. 1456
`Suk-Hyon Yoon, Dae-Ki Hong, Daesik Hong, Chang-Eon Kang, Yonsei Univ., Korea
`
`SESSION G36: MULTI-CARRIER AND SPREAD-SPECTRUM COMMUNICATIONS
`Organizer: Yumin Lee, National Taiwan University, Taiwan
`Chair: Chin-Liang Wang, National Tsing-Hua University, Taiwan
`___________________________.____—————-—————————‘
`G36.1 A LOW COMPLEXITY ML CHANNEL ESTIMATOR FOR OFDM ...................................................... 1461
`Luc Deneire, Patrick Vandenameele, Liesbet van der Perre, Marc Engels, Bert Gyselinckx, IMEC, Belgium
`G362 DS-CDMA RECEIVERS BASED ON CANONICAL CHANNEL REPRESENTATION ............................. 1466
`Jan Erik Hékegérd, SINTEF Telecom and lnfonnatics, Norway
`G363 OWSS MULTIPLE ACCESS SYSTEM FOR 100 MBPS WIRELESS LANS......................................... 1471
`Vijay Jain, University of South Florida, USA
`G364
`PEAK TO MEAN ENVELOPE POWER RATIO OF OVERSAMPLED OFDM SIGNALS: AN ANALYTICAL
`APPROACH...................................................................................................................................... 1476
`Masoud Sharif, Babak H. Khalaj, Sharif Univ. of Technology, Iran
`G365 AN EFFICIENT EQUALIZATION METHOD TO MINIMIZE DELAY SPREAD IN OFDM/DMT
`SYSTEMS ......................................................................................................................................... 1481
`Romed Schur, Joachim Speidel, Univ. of Stuttgart, Germany
`G366
`IMPLEMENTATION AND FIELD TRIAL RESULTS OF A FAST FREQUENCY HOPPED FSK
`TESTBED FOR WIRELESS COMMUNICATIONS ............................................................................. 1486
`Hanli Zou, UCLA, USA; Sumit Mohan, Qualcomm, USA; Danijela Cabric, Babak Daneshrad, UCLA, USA
`G367 A "SELF-DECORRELATING" TECHNIQUE TO ENHANCE BLIND SPACE-TIME RAKE RECEIVERS
`WITH SlNGLE-USER-TYPE DS-CDMA DETECTORS....................................................................... 1491
`Kainam Thomas Wong, Guisheng Liao, Shun Keung Cheung, Pak-Chung Ching, Chinese Univ. of Hong Kong, Hong Kong;
`Michael D. Zoltowski, Purdue Univ., USA; Javier Ramos, Carlos lll Univ. of Madrid, Spain
`G368 A PEAK POWER REDUCTION METHOD FOR MULTICARRIER TRANSMISSION ............................. 1496
`Chan Soo Hwang, Samsung Advanced Inst. of Technology, Korea
`G369
`SUPERIMPOSED TRAINING ON REDUNDANT PRECODING FOR LOW-COMPLEXITY RECOVERY
`OF BLOCK TRANSMISSIONS ...................................................................................................... 1501
`Shuichi Ohno. Shimane Univ., Japan; Georgios Giannakis, Univ. of Minnesota, USA
`xviii
`
`DISH, Ex. 1020 - Page 4 of 12
`
`DISH, Ex. 1020 - Page 4 of 12
`
`
`
`
`
`
`
`
`
`Volume 5 of 10
`
`UESDAY, JUNE 12, 2001
`13:30 - 17:00
`
`
` VEILLANCE AND TRAFFIC MANAGEMENT TECHNIQUES
`
`
`Minas Gerais', Brazil
`Jose Marcos Nogueira, Federal Univ. of Minas Gerais, Brazil
`
`SESSION 637: ADVANCED N
`Organizer: Jose Marcos Nogueira,
`
`Chair: Veli Sahin, Marconi Commun
`
`
`
`I'WITH MINIMAL ADMINISTRATION ON MULTIHOMED
`637.1 A DYNAMIC TRAFFIC
`
`NETWORKS......................
`.._. ................................................................................................... 1506
`Nariyoshi Yamai, Kiyohiko Okayama, Hiroshi Shimamoto, Takuji Okamoto, Okayama Univ., Japan
`G372
`LOCATION MANAGEMENT STRATEGIES FOR WIRELESS NETWORKS - A COMPARISON
`STUDY ON SYSTEM UTILIZATION ............................................................................................... 1511 '
`Tracy Tung, Abbas Jamalipour, Univ. of Sydney, Australia
`637.3 VISUALIZATION OF TRAFFIC STRUCTURES ............................................................................. 1516
`Oliver Niggemann, Benno Stein, Univ. of Paderbom, Germany; Jens Toelle, Univ. of Bonn, Germany
`637.4 UPLINK POWER CONTROL IN UTRA TDD ................................................................................. 1522
`Janne Kurjenniemi, Tapani Ristaniemi, Univ. of Jyvaskyla, Finland; Seppo O. Hamaléinen, Nokia, Finland
`637.5 A MEASUREMENT-BASED CONGESTION ALARM FOR SELF-SIMILAR TRAFFIC ............................ 1528
`Tat-Keung Chan, Wing-Cheong Lau, Univ. of Hong Kong, Hong Kong; Victor Li, HKU, Hong Kong
`G37.6
`INFERRING LINK CHARACTERISTICS FROM END-TO-END PATH MEASUREMENTS ...................... 1534
`Masato Tsuru, TAO, Japan; Tetsuya Takine, Kyoto Univ., Japan; Yuji Oie. Kyushu Inst. of Technology, Japan
`637.7 MOBILITY ADAPTIVE PROTOCOLS FOR MANAGING LARGE AD HOC NETWORKS ........................ 1539
`Stefano Basagni, Univ. of Texas at Dallas, USA; Damla Turgut, Sajal K. Das, Univ. of Texas at Arlington, USA
`637.8 ACTIVE DISTRIBUTED MONITORING FOR DYNAMIC LARGE-SCALE NETWORKS .......................... 1544
`Antonio Liotta, Univ. of Surrey, UK; George Pavlou, Univ. of Surrey, UK; Graham Knight, Univ. College London, UK
`637.9
`PREDICTIVE BANDWIDTH CONTROL FOR MPEG VIDEO: A WAVELET APPROACH
`FOR SELF-SIMILAR PARAMETERS ESTIMATION .................................................................................. 1551
`Yen-Chieh Ouyang Ouyang, National Chung Hsing Univ., Taiwan; Li-Bin Yeh, National Chung Hsing Univ., Taiwan
`
`SESSION G38a: SPECIAL TOPICS: MPLS, DATA STRUCTURE, TRANSMISSION NETWORK
`Organizer: Horst Besier, Deutsche TelekomrT-NOVA, Germany; 6.8. Kuo, National Chengchi University, Taiwan
`Chair: Nikola Rozic, University of Split, Croatia; Richard Vickers, Nortel Networks, USA
`
`6383.1 A NEW ARCHITECTURE FOR TRANSMISSION OF MPEG—4 VIDEO ON
`MPLS NETWORKS ............................................................................................................................. 1556
`Geng-Sheng (6.8.) KUO, National Chengchi University, Taiwan; C. T. Lai, National Central University, Taiwan
`G3Ba.2 EFFICIENT CACHE STRUCTURES OF IP ROUTERS TO PROVIDE
`POLICY-BASED SERVICES ................................................................................................................ 1561
`Shingo Ata, Masayuki Murata, Hideo Miyahara, Osaka University, Japan
`G3Ba.3 RESTORATION METHODS FOR TRAFFIC ENGINEERED NETWORKS WITH LOOP-FREE
`ROUTING GUARANTEES ........................................................................................................... 1566
`Richard Rabbat, Kai-Yeung (Sunny) Siu, Massachusetts Inst. of Technology. USA
`638a.4 AN ONLINE DISTRIBUTED PROTECTION ALGORITHM IN WDM NETWORKS ............................... 1571
`Xun Su, ECE department, UT Austin, USA; Ching-Fong Su, Fujitsu Labs of America, USA
`
`xix
`
`DISH, Ex. 1020 - Page 5 of 12
`
`
`
`
`
`
`
`(DO0(7)OZC).EOIOOU’501110320130ms-nmloomQDOI:I'TIIILII><II¢DOOM
`
`
`
`
`
`DISH, Ex. 1020 - Page 5 of 12
`
`
`
`:10
`
`.1506
`
`1511
`
`1516
`
`1 522
`
`1528
`
`1534
`
`1539
`
`.1544
`
`1551
`
`.1556
`
`1561
`
`1566
`
`1571
`
`Volume 5 of 10
`
`
`
`
`TUESDAY, JUNE 12, 2001
`13:30 - 17:00
`
`
`
`
`SESSION HSN38b: QUAL
`
`
`Organizer: Jiirgen Haag, Deutsche Te
`
`
`Chair: Manfred Fetter, Deutsche Tele
`
`Sponsor: Communications 8
`
`HSN38b.1 A NOVEL EXPLICI‘
`
`Xiaolin Zhang, Jieyi Wu, Southeast Uni'ver’s
`HSN38b.2 A 008 ROUTING ALGORITHM BASED ON ANT ALGORITHM ................................................. 1568
`Zhang Subing, Liu Zemin, Beijing Univ. of Posts and Telecommunications, China
`HSN38b.3 A HIGH-THROUGHPUT SCHEDULING ALGORITHM FOR A BUFFERED CROSSBAR SWITCH
`FABRIC ................................................... . ........................................................................................ 1581
`Tara Javidi, Univ. of Michigan, USA; Robert MagiII, Terry Hrabik, TeIIabs Operations, Inc., USA
`HSN38b.4 A CAPACITY CORRELATION MODEL FOR WDM NETWORKS WITH CONSTRAINED GROOMING
`CAPABILITIES.................................................................................................................................. 1592
`Sashisekaran Thiagarajan, Arun Somani, Iowa State University, USA
`
`SESSION G39a: VIDEO ON DEMAND SYSTEMS
`Organizer: Tak Kamae, Labs. of Image Science 8. Tech, Japan
`Chair: Ryoichi Komiya, Multimedia University, Malaysia
`G39a.1
`TOKEN-TRAY/ WEIGHTED QUEUING-TIME (TTNVQT): AN ADAPTIVE BATCHING POLICY
`FOR NEAR VIDEO-ON-DEMAND SYSTEM ..................................................................................... 1597
`Wushao Wen, Biswanath Mukherjee, Univ. of California, Davis, USA;
`S.-H. Gary Chan, Hong Kong Univ. of Science and Technology, Hong Kong
`G39a.2
`FAST VIDEO PLACEMENT ALGORITHMS FOR HIERARCHICAL VOD SYSTEMS ........................... 1602
`Ren-Hung Hwang, Pin-Hao Chi, CSIE/CCU, Taiwan
`G39a.3
`PROVIDING SCALABLE ON-DEMAND INTERACTIVE VIDEO SERVICE BY MEANS OF
`MULTICASTING AND CLIENT BUFFERING........................................................................................... 1607
`S.—H. Gary Chan, Hong Kong Univ. of Science and Technology, Hong Kong;
`Edward Chang, Univ. of California at Santa Barbara, USA
`G39a.4 CHANNEL ALLOCATION FOR LARGE SCALE VIDEO-ON-DEMAND SYSTEMS ............................. 1612
`Wing-Fai Poon, Kwok-Tung Lo, Jian Feng, City Univ. of Hong Kong, Hong Kong
`
`SESSION 63%: INFORMATION INFRASTRUCTURE AND HOT TOPICS
`Organizer: Cengiz Akgun, Telcordia, USA
`Chair: Barcin Kozbe, Ericsson Inc, USA
`________________________________._____———————————-——————
`G39b.1 A ROLE BASED MODELLING APPROACH FOR THE INFORMATION INFRASTRUCTURE................ 1617
`Jiirgen Apfelbeck, University Hagen, Germany
`G39b.2 SERVICE SELECTION IN THE ACCESS NETWORK.................................................................... 1622
`Nick MarIy, Dominique Chantrain, Stephane Focant, Koen Handekyn, Koen Daenen,
`Claudine Batsleer, AIcateI Corp. Research Center, Belgium
`G39b.3 DIAGNOSIS OF SENSOR NETWORKS..................................................................................... 1627
`Chaiporn Jaikaeo, Chavalit Srisathapornphat, Chien-Chung Shen, Univ‘ of Delaware, USA
`G39b.4 DIFFSERVER: APPLICATION LEVEL DIFFERENTIATED SERVICES FOR WEB SERVERS............... 1633
`Gautam Rao, Cisco Systems, USA; Byrav Ramamurthy, UNL, USA
`
`
`
`xix
`
`XX
`
`DISH, Ex. 1020 - Page 6 of 12
`
`DISH, Ex. 1020 - Page 6 of 12
`
`
`
`Handoff of Application Sessions Across Time and Space
`
`Thomas Phan, Kaixin Xu, Richard Guy, and Rajive Bagrodia
`{phantom xkx, rguy, raj ive} @cs . ucla . edu
`
`Parallel Computing Laboratory
`Computer Science Department
`The University of California at Los Angeles
`Los Angeles, CA 90095
`http://pcl.cs.ucla.edu
`
`:9
`
`Abstract
`
`Personal computing on mobile platforms such as laptops and
`personal digital assistants, rather than in a traditional desktop
`environment, is becoming increasingly more common. In this pa-
`per we address the issue of application session transfer for unin-
`terrupted data access across this diverse range of platforms. This
`work is part of the iMASH project, a multi-year, mum-discipline
`collaborative effort focused on enabling mobile client platforms
`and incorporating them into existing legacy networked systems
`for use by medical practitioners. We have developed a tiered ar-
`chitecture that includes a middleware server layer positioned be-
`tween existing application servers and multiple clients to make
`session transfer transparent to the user. Any client application
`executing our Middleware-Aware Remote Code library can save
`and restore its session by interacting with a middleware server.
`As a proof of concept, we have implemented the transfer of book-
`marks, history, web cache, and user preferences with the Mozilla
`open source web browser. From this effort we have established
`baseline performance metrics and have found that the overhead
`is within reasonable bounds of just a few seconds of latency.
`
`1
`
`Introduction
`
`Today, as workforce professionals become more mobile. there is an
`increasing demand for computing in locations and situations that
`dramatically diverge from the traditional wired office environment.
`The popularity of laptop computers and personal digital assistants
`(PDAs) in conjunction with wireless data service further make no-
`madic computing a desirable, viable, and increasingly common work
`style. However, computing mobility opens the door to an influx of
`myriad technical issues [4] that must be resolved before nomadicity
`can be considered as a standard rather than as an exceptional be-
`havioural trait for computer users.
`
`The issue we seek to address is one of data access for the mobile
`user. Previous work has been primarily concerned with providing
`data for a user on the same client across possibly different physical
`networks.
`In contrast, our novel contribution to the field of mobile
`computing is focused on the architecture and protocols that enable
`a user to experience uninterrupted and seamless data access across
`multiple devices by performing application session handoff.
`
`We define a system as a tuple of {platform location, time} and a
`handoff as a transfer of application data from one system to another
`via the computer network. both wired and wireless. Any change in
`a given tuple represents a new system, and thus a new representative
`entity, in our environment. We consider a session to be a strict subset
`
`of the process dataspace rather than an entire binary image The ses-
`sion is defined in the context of a given application but is independent
`
`of the architecture and operating system, thus allowing a session to ‘7‘
`move from one application on one system to another implementation
`
`of the same application on another system, so long as the applica—
`
`tions handle the session uniformly. Thus, a user participating in our
`
`network will be able to move between an office desktop, a PDA, a
`
`laptop, and a home desktop with a continuous view of his or her ap-
`
`plication data. As part of the UCLA Interactive Mobile Application
`Support for Heterogeneous Clients (iMASH) project [1] providing 1
`mobile data access for medical practitioners, we have developed the _'
`hardware and software infrastructure that will enable this scenario I
`for nomadic users.
`’
`
`The key component to faciliate this session transfer in the iMASH :
`architecture is an array of middleware servers strategically placed be- »
`tween the application servers and the clients. Requests from clients ,
`are addressed to nearby middleware servers, which are the source for
`all client services. Likewise, these servers become the sole client for
`application servers, allowing the latter to better perform their role as
`a data source rather than as a supporter of a large number of heteroge-
`neous clients. In the future the middleware servers will play a larger ‘
`role in the iMASH architecture than just supporting session handoff, .
`which is the focus of this paper. The middleware layer will provide
`such essential functionalities as user authentication, information dis-
`covery, data caching, presentation conversion, location tracking, and
`profile management.
`
`
`
`
`
`5
`
`
`
`We have implemented this handoff capability into the open—source
`Mozilla web browser as a proof of concept. We posit that the func— I,
`tionality, convenience. and performance of this handoff mechanism
`in something so commonly used as a browser make a convincing ar—
`gument for the usefulness of our research effort.
`
`:xwfaw,
`The rest of this paper is organised in the following manner. In §2.
`we discuss related work in the field of nomadic computing, followed .1:
`by the system architecture in §3. In §4 we discuss our experimental
`(
`
`work on the Mozilla browser along with our results.
`
`2 Related Work
`
`
`
`Nomadic computing in a wireless environment is discussed in [3] in I
`the context of the BARWAN research project. The authors imple-
`‘
`merited a system that supports mobility by smoothing client roam— j.
`ing across physical networks with network overlays, hiding wireless 7*}
`TCP losses with a reliable transport layer, and enabling dynamic con—
`tent transformation with a network-of—workstations middle tier that
`processes data between the content servers and the clients. Clearly,
`
`-»
`5
`
`,
`
`0-7803-7097-1/01/31000 ©2001 IEEE
`
`1367
`
`DISH, Ex. 1020 - Page 7 of 12
`
`DISH, Ex. 1020 - Page 7 of 12
`
`
`
`
`
`%i
`
`,
`l'l
`l
`
`
`
`'vyW<me<<awlfimWW~/,uuvavMMW..,,,MWWWWW“...
`
`
`
`
`
`g.3
`
`lig
`
`,l
`
`i
`
`cut
`to
`ion
`ca—
`Jur
`., a
`ap—
`ion
`ing
`the
`trio
`
`SH
`be-
`:nts
`for
`for
`e as
`)ge—
`.‘ger
`off.
`Iide
`dis-
`and
`
`rrce
`lnC-
`tism
`; ar—
`
`iMASH shares similarities with BARWAN, but BARWAN does not
`support session handoff between clients; specifically, its users are
`limited to having one mobile client with independent session states.
`and its middle tier performs only content adaptation.
`
`The researchers in the Odyssey project [10] investigated the use—
`fulness of fast adaptation by modifying a UNIX kernel to alert ap-
`plications to changes in network conditions. Each application, upon
`receiving this information, individually adapts its behaviour to suit
`the available network bandwidth. Such awareness and adaptation re-
`quires applications to closely interact with a client-side portion of
`Odyssey code, similar to our Middleware-Awarc Remote Code we
`shall discuss in 3.3. and like iMASH, the Odyssey architecture con—
`tains a middleware-like layer of hardware but is largely limited to
`content adaptation.
`
`Data transfer between clients similar to our session handoff is ex-
`plored in [2]. which describes an infrastructure that manages applica—
`tion adaptations to changing network 005 using a set of cooperating
`servers, including a mobility manager and an application migration
`manager. In [12], researchers developed the Roma system for main-
`taining data across multiple clients but provide for only the transfer
`of XML-based metadata that keep file versioning information.
`In
`this architecture the actual data transfer must be accomplished by
`well-known network protocols such as HTTP, FTP, and NFS.
`
`Our work is also related to the well-known concept of process
`migration in the fields of fault—tolerance and load-balancing. Pro-
`cess migration, surveyed in [8], has primarily focused on application
`migration among workstations on LANs and generally required all
`the participating nodes to have homogeneous platforms. In contrast.
`iMASH explicitly targets a high degree of device and OS hetero-
`geneity. Additionally, the check-pointing and migration activities,
`implemented in systems such as Condor [6]. rely on the brute-force
`copying of a process’ entire binary image. An iMASH application
`session, on the other hand, is a strict subset of the dataspace and con-
`sumes much less memory. The session state at the destination should
`be adapted to the underlying device characteristics, which may be
`dramatically different from that on the source node. This additional
`goal requires the session handoff mechanism to have much more in«
`telligence than a process migration implementation.
`
`In this paper we used a web browser as one possible applica-
`tion that can take advantage of the iMASH environment by imple-
`menting a prototype session handoff capability service similarly pro—
`vided in Netscape Communicator 4.5+ and by such third parties as
`WebRoamer.com. Although these companies provide facilities that
`enable the same end result (e.g. a bookmark file transfer), such simi-
`larity is only superficial because they rely on simplistic data transfer
`via LDAP or HTTP to a fixed server. Our implementation performs
`_ dynamic service discovery with Sun’s Jini service and can execute
`much more flexible and extensible transactions using Java RMI, all
`within the context of the larger and more scalable iMASH infrastruc-
`ture.
`
`3 System Architecture
`
`. 3.1 Background
`
`
`
`Figure 1: The iMASH architecture
`
`vices; (b) balance application requirements and network conditions;
`and (c) enhance the scalability of the network.
`
`iMASH, a joint collaboration between the Computer Science De-
`partment and the Medical School,
`is developing such an adaptive
`middleware infrastructure, initially for healthcare professionals.
`lts
`objective is to provide the capability for real-time multimedia com—
`munication, such that a physician can access, on the move, patient
`records and related information. Additionally, users will be able
`to migrate ongoing application sessions seamlessly to multiple plat-
`forms that may range from a high performance diagnostic worksta-
`tion in the physician’s office to hand-held PDAs in the examination
`room.
`
`3.2 Hardware Architecture
`
`The iMASH project will deploy a hardware and software infrastruc—
`ture adapted to the healthcare domain. The planned hardware archi-
`tecture is depicted in Figure l and consists of:
`
`\/ A set of application servers connected to a high bandwidth
`wired network. These machines are typically legacy servers
`that are not well suited to support multiple mobile clients.
`
`J A collection of wired client devices that includes office work-
`stations, clinic stations, imaging equipment, etc.
`
`\/ High bandwidth wireless access within the hospital, augmented
`by one or more lower bandwidth wide area wireless systems,
`supporting a number of heterogeneous mobile client devices
`that can vary from laptop to graphical tablet to PDA. These de-
`vices have a wide variety of physical characteristics, including
`CPU power, memory, storage. battery life, weight, and display
`quality.
`
`\/ A set of strategically deployed, dedicated middleware servers.
`
`The infrastructure to support the iMASH vision must evolve to al-
`low users seamless access to data from any device on a variety of
`networks. An important challenge is to develop a middleware infras—
`tructure to: (a) help bridge the gap between application servers and
`nomadic clients as they migrate active sessions among multiple de-
`
`The key to the iMASH architecture is the distributed set of mid-
`dleware servers, which when viewed at the application level, inter-
`connect all applications and clients. Requests from clients are ex-
`plicitly addressed to nearby middleware sewers; these middleware
`servers are the source for all services for clients, simplifying service
`discovery, as well as the sole client for application servers.
`
`’
`
`| §2.
`wed
`:ntal
`
`I
`
`'
`
`
`
`
`
`
`3] in
`1ple—
`)am—
`eless
`con—
`‘ that
`:arly.
`
`,, .3,. ..
`
`0-7803—7097-1/01/$10.00 ©2001 IEEE
`
`1368
`
`DISH, Ex. 1020 - Page 8 of 12
`
`DISH, Ex. 1020 - Page 8 of 12
`
`
`
`é.
`
`
`
`
`
`
`
`
`um?".
`
`
`
`W.tmmst‘mwm
`
`”.5
`
`Figure 2: Middleware architecture
`
`The deployment of middleware servers between the clients and
`servers provides scalability in several ways. Application servers
`need not be concerned with a large number of end users but instead
`can concentrate on collecting and processing data. A client device
`need not be concerned with contacting potentially large numbers of
`servers to gather data but instead contacts only the local middleware
`server for all services. Middleware takes responsibility for getting
`the data from the right servers, and makes necessary conversion to fit
`the clients’ needs. Middleware can also help scaling by temporary
`caching, fetching each piece of data only once but serving multiple
`requests for the same data, for example, as a physician moves locally
`between a clinic workstation and a graphical tablet.
`As shown in Figure 2, middleware servers provide the following
`functionality:
`
`\/ User authentication and profiling. Middleware boxes are
`trusted agents. They authenticate end users on behalf of the
`application.
`\/ Information discovery/data caching. The middleware server
`acts as a proxy on behalf of the client to retrieve data from the
`application server.
`\/ Presentation conversion. Middleware servers fetch data based
`on user requests (or pre-fetch data based on prediction of user‘s
`near-future need) and perform conversion as needed.
`
`The preceding services would be required even in a wired en-
`vironment. Mobility brings additional requirements to middleware
`service:
`
`\/ Session handojf. When a user moves an on-going application
`session from one device to another, middleware servers act as a
`"home” for the application state (including active connections,
`cached data, etc.) to facilitate migration between devices.
`\/ Location tracking and network QoS adaptation. To maintain
`an active application session as a user moves, the server tracks
`the user’s location, adjusts to local network conditions by pre-
`allocating bandwidth and other resources along the user’s path,
`and may forward sessions to other middleware servers.
`
`0-7803-7097-1/01/$10.00 ©2001 IEEE
`
`1369
`
`DISH, Ex. 1020 - Page 9 of 12
`
`
`
`
`
`Figure 3: The MARC architecture
`
`\/ User re-authentication. When a user changes devices or
`spaWns a new branch of a session to a new device, the mid-
`dleware server authenticates the user on the new device.
`
`In this paper we focus on the issue of session data transfer while
`conserving the rest for future work.
`
`3.3 Software Architecture
`
`A principal goal of our architecture is to all