1. Introduction

Over the past several years there has been a considerable amount of work conducted in the area of broadband wireless and mobile communications. The scope of this research has broadly ranged from the design of new radio, physical and MAC layer technologies to the provisioning of networking support for mobile multimedia devices, services and applications. Results from industrial and academic research laboratories are helping to define next generation wireless LANs, mobile radio networks and cellular systems. While there have been numerous conferences and workshops addressing technical developments in the field there have been very few opportunities to observe broadband wireless and mobile communications systems in operation other than visiting a number of labs dotted around the globe.

This report presents a review of the first Workshop on Wireless Broadband Testbeds (DEMO'98) [1], which was held in cooperation with the fifth International Workshop on Mobile Multimedia Communications (MoMuC'98), Rathaus Schöneberg, Berlin, October 1998. The idea behind DEMO’98 was to provide a forum for the demonstration of operational wireless broadband prototype systems.

The workshop was organized as a single-track event bringing together over 100 researchers, developers and practitioners working on the development of wireless broadband networks, middleware and mobile applications. Researchers were invited to submit a short description of their prototype system for consideration. The call for participation simply stated that the event was intended for non-commercial systems and that to qualify a prototype should be capable of demonstrating multimedia services (including voice, video and data) over a broadband wireless interface. As a result seven demonstrators were selected offering a unique forum for the demonstration of state-of-the-art wireless ATM systems, high-speed wireless LANs and next generation cellular systems.

DEMO’98 dedicated half of the day to formal presentations and the remaining time was for system demonstrations in an exhibition hall. The formal presentations included a keynote address, a session on global standardization efforts and a panel on applications, services and markets. Each project team made a workshop presentation outlining the key design features of their equipment. The formal sessions helped to add some context to the demonstrations, which started after the formal presentations concluded. Attendees freely mingled around the various booths in the exhibition hall viewing the demos at close quarters; and importantly, attendees had ample opportunity to talk with the system designers.

The format of DEMO’98 seemed to strike a good balance between hearing about the latest advances in wireless and mobile broadband communications and seeing the actual demos. In what follows, we provide a summary of that day’s events at the Rathaus Schöneberg.

 

2. Keynote Presentation

Chair: Andrew T. Campbell, Columbia University

The keynote address [2] given by Dipankar Raychaudhuri (NEC USA, C&C Research Laboratories, USA) was entitled "Broadband Wireless Networks: Current Status & Future Directions". The presentation addressed emerging services, network architecture and technology. The speaker noted that broadband wireless networks represent a key enabling technology for mobile multimedia services of the future. In the near term they will also facilitate rapid deployment of high-speed access to fixed computing devices. Future wireless networks should provide a range of generic transport services for various current and future applications. A number of systems requirements should be considered when evaluating candidate network architectures and component technologies: e.g., high-speed access capabilities, service integration and quality of service (QOS), scalability (in terms of users, traffic and coverage area), spectral efficiency and user mobility support.

Raychaudhuri discussed an integrated architectural approach suitable for the delivery of scalable ubiquitous services that would allow multiple radio air interfaces to be "plugged" into a mobility-enhanced core network. One issue that remained to be resolved was the selection of a suitable multimedia-capable core network protocol. Wireless ATM is a candidate, which assumes that the multimedia services are ATM-based, with support for service integration and quality of service and a connection-oriented framework suitable for dynamic handoff. Wireless IP presents an alternative approach given that Internet protocols evolve to support multimedia services. The major technology components of these architectures can be viewed as the core network and the necessary mobile network extensions, medium access and data link control, and finally the radio and physical layer.

A large part of the Raychaudhuri’s keynote address focussed on these technology components. Referring to the debate over standards (e.g. European Telecommunications Standards Institute (ETSI) BRAN [15]) vs. technology neutral (e.g., US U-NNI) approaches to broadband wireless spectrum, Raychaudhuri thought that the U-NNI approach seemed preferable for relatively immature high-speed radio technologies, especially since they can be designed to "plug" into standard network infrastructures. On the subject of modems he noted that in an era of exponentially increasing bandwidth, we need to start building 100 Mbps+ modems for use in 2-3 years from now! Regarding the discussion of MAC protocols and the adoption of CSMA or dynamic TDMA technologies, he thought that as with modulation, a single MAC standard is not mandatory as long as the interfaces are consistent with the core network protocols. Other subjects covered by the keynote included data link control support for stream/QOS services, integrated mobility support as a "first-class" service and IP/ATM integration with mobility. The speaker touched on a number of important systems issues, e.g., legacy radio service integration, various mobile ‘IP+ATM’ solutions, QOS and adaptive mobile application support.

NEC USA, C&C Research Laboratories had recently completed the development of a 25 Mbps/5 GHz wireless ATM system [16] that featured an equalized GMSK modem, NIC with dynamic TDMA/TDD MAC and DLC support, mobile ATM networking with integrated location management and handoff, and TCP/IP and native ATM services with QOS control. Reflecting on five years of research and development activity he said that a number of results have been achieved. First, the technical feasibility of the WATM approach has been validated, though the commercially viability has yet to be proven. In addition, key technologies (e.g.,10 Mbps+ modems, MAC, mobile ATM/ IP) are still at early product development stages but are maturing rapidly. Raychaudhuri noted that the standardization efforts for WATM is ongoing in the ATM Forum and ETSI and there was also growing interest in IP and ATM based mobile networks as a basis for migration to IMT-2000. Early commercial products are expected during 1999 on at least a trial basis (e.g., 10 Mbps+ 802.11 and 25 Mbps WATM). He thought the initial application of the emerging technology would be in vertical mobile multimedia markets gradually migrating to mass market semi-mobile broadband access. Raychaudhuri finished with his "Top 10" list of challenging work for the future. Ray said that his "David Letterman Top 10 List" included the development of a 100+ Mbps radio modem with 3+ bps/Hz with a range of 250m+, a generic MAC/DLC for ATM, IP, etc., an efficient IP/WATM mobile network solution, an IMT-2000 platform implemented as a generic mobile network, programmable mobile networks and finally new mobile terminals and applications.

3. International Broadband Wireless Initiatives

Chair: Andy Hopper, AT&T Laboratories Cambridge and Cambridge University

The next session presented three different views on standardization of broadband wireless networks. The first talk [4] was given by Masahiro Umehira (NTT Wireless Systems Laboratories, Japan). In 1996, a new study group was formed in Japan by the Multimedia Mobile Access Communication Systems (MMAC) Promotion Council [23] in cooperation with users, carriers, manufacturers and academics to propose a high performance wireless system that would go beyond the current work on IMT-2000. The goal of the Promotion Council is to realize MMAC systems through investigation of system specification, development and demonstration. Masahiro Umehira noted that three types of MMAC systems have been proposed to operate in the 5GHz band, which he noted is a common frequency band in the US, Europe and Japan. These systems cover ATM-based high-speed wireless access at 20+ Mbps with cell radius of 200m, IEEE 802.11 type WLANs and wireless home-link at 36 Mbps with a cell radius of 5m. The speaker reflected that with three global standardization activities in the area of broadband wireless access (viz. ETSI-BRAN, IEEE802.11 and MMAC) it is very important that there is agreement between the groups on common frequency spacing, physical layer, frequency sharing rules and air-interface.

The next presentation [3] was made by Jamshid Khun Jush (Ericsson Eurolab, Germany). In early 1997, ETSI created a project called Broadband Radio Access Network (BRAN) [15] to develop broadband radio local loops and other radio access systems operating in licensed and unlicensed bands at data rates between 25 -155 Mbps. BRAN leveraged work on HIPERLAN 25 Mbps ATM LAN access in the 5.15-5.3 GHz band as well as access technologies at higher frequencies (e.g., 17.1-17.3 GHz band) and data rates (e.g., 155 Mbps). Jamshid Khun Jush mentioned that BRAN is coordinating with the core network (e.g., ATM Forum and IETF) and peer (e.g., IEEE and MMAC) fora. Currently, the BRAN working group is investigating spectrum issues, broadband radio technologies and has taken input from the WAND [17] and AWACS [14] projects. The group is also working toward a common ETSI-ATM Forum reference model. The wireless ATM working group and BRAN are combining forces in this instance to develop the radio access layer specification. The speaker concluded by saying that BRAN will be the prime developer of the new radio technology that will include advanced modulation techniques for scalable performance, new scheduled TDMA techniques for assured QOS and flexible error control strategies for meeting application-level requirements.

The final presentation [5] of the session was by Naftali Chayat (Breezcom, US). In July 1997, IEEE 802.11 [24] formed two new committees to define new draft standards for 802.11 WLAN to operate at high-speeds in the 5GHz and 2.4 GHz bands. Both of these projects are developing new physical layers to work with the same 802.11 MAC. Naftali Chayat, who is the current 802.11a Chair, provided an overview of the existing 802.11 MAC and multirate support detailing its application to outdoor fixed, outdoor mobile, indoor stationary and mobile work force environments. While 802.11 MAC has been proven to support voice, higher speeds and better quality of service support is required for new bandwidth demanding services, e.g., video distribution. Naftali Chayat concluded with the comment that the first working group letter ballot for the high speed 5GHz is expected in late 1998 with final approval anticipated by the end of 1999.

4. The Demos

Chair: Adam Wolisz, Technical University, Berlin

1. Radio ATM

The Radio ATM kit [18] [6] developed by Olivetti and Oracle Research (recently acquired by AT&T), Cambridge, England was presented and demonstrated by John Porter, Alistair Massarella and David Pegler.

The current Radio ATM demonstrator evolved from an early 10 Mbps QPSK radio operating at 2.45 GHz for mobile ATM networking to a fixed wireless access network operating at 25 Mbps in the unlicensed U-NII 5 GHz band. The ORL kit demonstrated in Berlin consisted of AB-Access technology produced by a new "spin-off" company of ORL called Adaptive Broadband Ltd [22]. The AB-Access kit was used to demonstrate wireless broadband access to Internet data, video and voice applications. The Adaptive Broadband technology is highly scalable, and can be used to deploy a complete wireless network or to provide enhanced services in areas where existing wired infrastructure cannot cope with customer demand. An AB-Access network is deployed in a conventional cellular pattern, from a single cell or island sites through to coverage of a complete geographical area. The demonstrator comprised a number of subscriber terminals representing customer premises, a cell site providing communication between subscribers and a core Escalate ATM network. A covered region typically consists of a number of three to five kilometer radius cells. The cell sites provide a standard ATM connection to a wide area carrier network. Each cell site typically consists of six access points, each communicating with a set of subscribers. A single access point provides a 25 Mbps time division duplex channel, which is adaptively shared among up to 256 active service users. The AB-Access uses ATM quality of service to support traffic prioritization over the air-interface.

2. WATMnet System

The WATMnet system [16] [7] developed by the C&C Research Laboratories, NEC USA, Princeton, US was presented and demonstrated by Arup Acharya, Cesar Johnston, Max Ott and Daniel Reininger.

NEC C&C Research Laboratories demonstrated their WATMnet 2.0 systems. The prototype system represents the second generation of their WATM prototype kit operating at 25.6 Mbps in the 5GHz U-NII band. The demonstrator consisted of laptops with radio ATM interface cards, VME/v851 processor-based APs and a mobility-enhanced local area ATM switched core network. The system builds on two earlier versions of their broadband radio that initially operated at 8 Mbps in the 2.4 GHz ISM-band. The dynamic TDMA/TDD protocol used provides support for packet and quality of service stream services. Dynamic TDMA/TDD medium access control provides explicit support for the ATM CBR, VBR and ABR services over the air-interface. Both WATM and IP packet service architectures were demonstrated with system support for scalable quality of service capable applications. The mobile-capable core network uses NEC ATM switches to interconnect base stations supporting seamless handoff. The WATMnet 2.0 supports 7 channels in the 5GHz band with a transmit power of 10mw and channelisation of 25.6 MHz channels using GSMK. Mobility support includes protocol extensions providing efficient support within the access network including functions that support location management, handoff control and mobile QOS routing. The NEC demo was multifaceted showing support for adaptive QOS support and QOS-aware application software based on distributed object technology.

3. AWACS System

The AWACS systems [14] [8] developed as part of the ACTS Program was presented and demonstrated by Markku Tolonen (Elektrobit, Finland) and Cengiz Evci (Alcatel, France).

The AWACS system demonstrated wireless public access to broadband services. The system operated in the 19 GHz band supporting user bit rates of up to 34 Mbps over a radio transmission range of 50 to 100 m. The demo system was configured to support a maximum user bit-rate of 10 Mbps over a range of 50 m. Services supported included two way MPEG audio and video data between two sites. AWACS is limited to slow speed mobility in line with expectations at very high data rates but with improved portability. The AWACS system is based upon a experimental system developed by NTT. The NTT prototype system operated in the super high frequency (3-30 GHz) band for public and private access at data rates between 30-80 Mbps. Highly directive antennas helped alleviate serious shadowing effects that were apparent at these frequencies. The NTT system was designed for high-speed access with limited terminal mobility incorporating dynamic reservation-based TDMA, FEC/ARQ, and QPSK modulation with differential detection. Enhancements made by the ACTS AWACS project included supports for a fully integrated antenna solution, mobility, as well as spectrum and power efficient radio access technologies required to support ETSI-RES10 HIPERLAN specifications.

4. Magic WAND Demonstrator

The Magic WAND demonstrator [17] [9] [10] developed as part of the ACTS Program was presented and demonstrated by Juha Ala-Laurila (Nokia Wireless Business Telecommunications, Finland) and Geert Awater (Lucent Technologies Bell Labs, Netherlands).

The WAND demonstrator was shown for the first time during DEMO’98 providing wireless extensions to ATM technology for mobile users operating in an indoor pico-cellular environment. Communications between mobile/portable terminals and access points took place in the 5 GHz band with a range of up to 50 meters. The main components of the demonstrator were a radio subsystem comprising of a 20 Mbps modem using OFDM, 16 carriers and 8 PSK modulation and the MASCARA medium access scheme with networked mobility enhanced ATM signalling. The radio cell size was optimized for a range of 20m. The MASCARA medium access controller uses dynamic TDMA, a master-slave priority leaky bucket scheduling mechanism and go-back-N on a per virtual connection basis. The DLC layer performance measured with the platform delivers TCP, UDP and pure DLC at data rates between 3-7 Mbps, 8-10 Mbps and 14-17 Mbps, respectively. The demonstrator platform comprised a single mobile terminal running modified Trillium UNI 3.1 ATM signaling code, video camera and audio card, ATM NIC and radio subsystem. The access point was implemented using a 200 MHz Pentium PC with 155 Mbps ATM NICs and dedicated radio hardware. The ATM switch used for the demo was a VTT FSR ATM switch matrix with the controller running on a Sun Ultra workstation, which was attached to the switch via a 155 Mbps ATM NIC.

5. MEMO DAB/GSM Kit

The MEMO demonstrator [20] [11] developed as part of the ACTS Program was presented and demonstrated by Wolfgang Klingenberg (Robert Bosch GmbH, Germany).

The main objective of the MEMO project was to develop a generic architecture for the provision of interactive multimedia services to mobile and portable terminals. The demo provided an asymmetrical bi-directional link combining a broadband downlink of 1.8 Mbps based on a Digital Audio Broadcasting System (DAB) and a low bit rate interactive link operating at 9.6 Kbps based on GSM. Two types of services based on electronic publishing (newspaper, videotext) were demonstrated with interactive functionality. The MEMO project developed generalized service models, and hardware and software platforms implementing a wide variety of interactive datacasting services. The DAB wide band system operates at 1.5 MHz using convolutional coding and OFDM modulation for reliable mobile delivery. The demo utilized three different widely geographically separated sites and the GSM and DAB systems.

6. Mobiware Programmable Mobile Network

The Mobiware Toolkit [19] [12] developed by the COMET Group, Columbia University, US was presented and demonstrated by Andrew T. Campbell and Raymond Liao.

The COMET Group built the first open programmable mobile network in 1997 and released the source code [19] to the public in 1998. Mobiware presents a foundation for the introduction of new adaptive mobile multimedia services on ‘programmable cellular systems’. Mobiware provides a set of CORBA and Java interfaces and objects that abstract and represent network devices and resources providing a toolkit for programmable signaling, adaptation management and wireless transport services. Based on an open programmable paradigm, mobiware runs on mobile devices, base stations and mobile-capable switch/routers providing a set of open programmable interfaces and algorithms (viz. QOS controlled mobility, mobile soft-state and flow bundling) for adaptive mobile networking. The Mobiware demonstrator was used to deliver real-time scalable audio, video and web services to mobile hosts. The testbed comprised two base stations connected to a Virata ATM switch. The switch was connected to the Internet via a gateway node. The base stations were based on a set of multi-homed 300 MHz Pentium PCs that provide radio access to mobile terminals based on waveLAN operating at 2 Mbps in the 2.5 GHz band and at 25Mbps to the ATM switch.

7. Wireless Compressor

The Wireless Compressor system [21] [13] developed by Uppsala University, Sweden was presented and demonstrated by Bjorn Knutsson and Mats Bjorkman.

The demonstrator provided a transport level scheme with dynamic compression adaptation. The demo was implemented on top of TCP using Linux providing transparent adaptive compression to mobile applications. The team demonstrated a mechanism to adapt compression to currently available bandwidth and CPU resources. The simple idea is to trade CPU and memory for better bandwidth utilization across the wireless link by adaptively compressing data. Compression can double the effective throughput but it can also cut it in half. The difference between winning and losing from compression is a question of knowing when and how to use it. The wireless compressor dynamically controls compression to maximize throughput. The compression level is indirectly set based on past performance, e.g., when there is less bandwidth available the algorithms increase the compression rate. The system demonstrated used a mobile client interacting with a base station using 2 Mbps waveLAN operating in the 2.4 GHz band. The dynamic compression scheme does not require any feedback between senders and receivers.

5. Panel

Chair: Hakan Mitts, Nokia, Finland

DEMO’98 concluded with a panel entitled "Future Apps, Services and Markets" which was organized by Hakan Mitts (Nokia) with participation from Naftali Chayat (Breezcom), Max Ott (NEC) and Jorge Pereria (EC). The panel discussed the use of the technology demonstrated at the workshop. The audience raised the issue of how long before 20+ Mbps high speed radios hit the market? Was there a ready market for the technology? What were the technical barriers that limited its uptake? How would applications exploit broadband wireless access? What were the regulatory barriers to the adoption of the technology globally? What new applications would emerge? Max Ott thought that new and novel wireless applications are more likely to emerge to meet niche application environments, e.g., wearable computers. Addressing the question of standardization, Jorge Pereria thought that the results from the EC ACT Program is helping to clarify the work of ETSI BRAN and ATMF WATM which is important for global uptake of the work.

6. Concluding Remarks

As the panel finished in the exhibition hall people started to drift away after a marathon 12 hour DEMO’98 day full of exciting presentations and cool demonstrations. As researchers started to break down their equipment for shipping, organizers reflected (with some amazement) that seven projects using a variety of commercial and proprietary radios, some operating in the same spectrum, worked well with only a few minor hiccups. DEMO’98 was considered a great success by participants and attendees presenting just the right mix of interaction, hands-on experience and exposure to state-of-the art demonstrators that the organizers had hoped for.

For online proceedings see [1] or mail demo@comet.columbia.edu for a hardcopy.

7. Acknowledgements

We would like to thank the sponsors, committee and webmaster for all their efforts to make DEMO’98 a success and to all the participating researchers that transported their kit from half a dozen countries. Next, a special thanks to Prof. Adam Wolisz and his team at Technical University, Berlin. Prof. Wolisz spent some time on the phone with the German Customs trying to persuade them to let some of the DEMO’98 radio kit into the country! Many thanks to Petra Hutt for her help and a big thanks to Theodoros Assimakopoulos (Technical University, Berlin) for his support during the run up to the event and for trouble-shooting on the day itself. Finally, any errors or omissions are ours and for those, our apologies in advance.

SPONSORS

Center for Telecommunications Research, Columbia University

Nokia

European Commission

COMMITTEE

Andrew T. Campbell, Columbia U., US (Co-Chair)

Hakan Mitts, Nokia Telecomm's, Finland (Co-Chair)

Naftali Chayat, BreezeCom, US

J. Da Silva, European Commission

Cengiz Evci, ALCATEL Telecom, France

Jan Kruys, Lucent, Netherlands

Louis-Francois Pau Ericsson Utvecklings AB, Sweden

John Porter, Adaptive Broadband Ltd, UK

D. Raychaudhuri, NEC, Princeton, US

Masahiro Umehira , NTT, Japan

Bharghavan Vaduvur, UIUC, US

Olle Viktorsson, Ericsson Radio Systems, SE

Adam Wolisz, Technical U. of Berlin, Germany

WEBMASTER

Michael E. Kounavis, Columbia University

8. References

[14] AWACS Project www.uk.infowin.org/ACTS/RUS/PROJECTS/AC228.htm

[15] ETSI-BRAN Project www.etsi.fr/bran

[16] WATMnet Project www.ccrl.nj.nec.com/key/mwATM/project/detail.html

[18] Radio ATM Project www.cam-orl.co.uk/radio/

[20] MEMO Project www.uk.infowin.org/ACTS/RUS/PROJECTS/ac054.htm

[22] Adaptive Broadband Ltd. www.adaptivebroadband.com/