Focused research initiatives, which dedicate resources to a specific - often collaborative - effort with clearly defined objectives, were suggested by the workshop participants as a means for advancing the field of wireless and mobile communications and networking. Four initiatives were identified as worthy of support: development of a multimegabit, multimedia software radio; demonstration of a low-power, wireless, distributed sensor network; extension of Internet services to a broadband wireless packet network; and demonstration of an order-of-magnitude increase in performance through incremental improvements in existing cellular radio infrastructure.
Transmitters and receivers are among the most critical and unique components of the wireless network. The need for higher-capacity transceivers is ongoing. Judged against the wired network, the wireless network suffers from a growing "bandwidth gap", in which the difference between the nominal data rate of the typical wirelink and that of the typical airlink is increasing over time. Since nearly all network services have been targeted at hosts connected to networks of peak capacity 10 Mb/s or greater, there is always the risk that a wireless link will be underpowered for a particular application. Moreover, users have come to expect higher levels of performance than are commonly offered by wireless networks. It is therefore crucial that radios able to deliver multimegabit data rates be developed, tested, and fielded.
The motivating need for such radios comes from multimedia applications, which have high throughput requirements and demands for multiple classes of traffic. The use of a software radio that can program a specific waveform for use in different settings will make it possible to adapt the link (perhaps in real time) to changing conditions. Such an approach is also expected to facilitate managing QoS.
Considering development of and experimentation with wireless multimedia to be of great importance, the workshop participants recommended that a focused initiative in this area be pursued to stimulate further research and provide a high-payoff yet cost-effective infrastructure for researchers in wireless and mobile communications and networking.
The development of such a software radio would be an important achievement in its own right, paving the way for commercial products that offer several services (e.g. AMPS, DAMPS, GSM, PCS, CDMA) in a single region. This could in turn stimulate new classes of research problems, such as the transparent selection and management of wireless services on behalf of a user.
The implementation of a wireless network of sensors offers an opportunity for a concentrated effort to develop communication technology, protocols, and applications for this unique, little-explored area. Such an effort would draw upon the expertise of many researchers, from device technologists to software developers. It could also provide the impetus to put into practice the crosscutting design approach advocated in Chapter 5, as the low-power, wireless, distributed-sensor network places demands on several of the design attributes identified. The requirement for rapid feedback and control, for instance, makes such a network a strong candidate for delay-reduction techniques. Likewise, the nonpermanence of both sensors and their embedding platform are relevant to mobility issues. Service integration, security, and spectral efficiency also enter into the design of this class of network. Finally, the goal of embedding large numbers of sensors and linking embedding platforms together stresses the attributes of scalability, heterogeneity, and robustness.
Many of the research priorities identified in Chapter 4 relate directly to this initiative. Energy efficiency, ad hoc network design, smart antennas, physical and link layers, signal propagation, and multicast all would play a part in this initiative.
The speed with which mobility is being adopted in the Internet suggests that wireless access to Internet services is not far behind. Indeed, many advanced users already use wireless data networks to connect within their home sites, using their internal network in a peripatetic manner. Given the pace at which Internet protocols are being deployed over new wired technologies, workshop participants felt that it is essential to foster experimental wireless networks in which to develop and evaluate Internet services.
Linking members of the Internet research community together with a broadband wireless network would spur experimentation and development of new services and protocols needed by wireless mobile network users.
The above-mentioned "bandwidth gap" between wireless and wired networks appears to be growing. Speed, reliability, and latency performance of wireless networks is often unacceptable for users, especially when judged against wired networks. A concerted effort should be made to boost the performance of state-of-the-art wireless networks. It was noted that improvements in the latency of wireless communications is a particularly important goal.