Mobile Computing: Wireless Access to Multimedia Applications and its Implications for a Campus Network

The Network and Computing Services section of Cornell Information Technologies in partnership with the Department of Computer Science is sponsoring a trial for wireless services on campus. Our paper will present an evaluation framework and we report results available in April. Some 100 students and 12 researchers are contributing to the work and it is anticipated that this will roll out into a distinct service offering by the end of the calendar year. When users jump from a wire based architecture to a wireless mode of communication this structural adjustment has the potential to change not only the way these users work but also the way they perceive their daily life. The Technology Service Evaluation and Development Group (TSED) in conjunction with the Human Computer Interaction Lab (HCI) and the Computer Science Department at Cornell University will be developing, implementing and testing a wireless network in the Spring semester. TSED serves as a technology transfer and service development team within Cornell Information Technologies (CIT). The role of CIT in this project includes performing a service evaluation, providing technical support, configuring intra-connections, and use of the campus fiber plant. The intended outcome is to define the technical service elements of a campus wide wireless service. A number of wireless applications will become available over the next five years with a particular interest in video for many users. The merging of computers, video, and telephone over a wireless network is creating many possibilities for new patterns to emerge while pressing for rapid development of adequate infrastructure. The use of mobile devices for entertainment, communication, learning and computing is an intriguing and powerful mix of potential problems and solutions. This adds a level of complexity to the technical trial as we ask questions about how many students will be using the wireless network and for what purposes. In an effort to fully explore the opportunities this study offers TSED plans an analysis of network development, network operations and patterns of user response. The framework for the research follows. The project begins in January 2000 and will continue through the summer. At the physical layer the study will examine how the network hardware and software performs and explore the strengths and limitations of wireless services. In order to develop a full understanding of the needed service profile, spectrum management and frequency monitoring will be part of the evaluation. (The 802.11b standard protocol will be used for all devices.) The NOMAD network is designed as a web of wireless transceivers that serve as access points. Each is connected to the campus network. The access points and Aironet cards installed in each laptop exchange low power microwave signals over a limited range. Each access point tracks the wireless devices within its range, communicates with other access points and automatically assigns devices based on best reception. If the device stays in range of any of the access points, users are able to move from place-to-place. Users are not usually aware of the actual infrastructure as long as they are in range of an access point. Network design will be evaluated by robustness of the network, ability to respond to maximum demand, and any design failures experienced during the pilot. Access points will be monitored for reliability. Interference in the designated spectrum will be investigated and analyzed over the course of the semester as it occurs. Network management tools will be used to address the following questions: Are specifications met and what special circumstances may alter expectations about performance? There are various types of equipment that will be used; will compatibility be a problem? A variety of equipment types will be in use and tested for interoperability. Protocols developed at Syracuse University will be used to create an interoperability matrix. Load balancing at the access points will be tested including classroom saturation of the bandwidth. The number of access points also will be evaluated in relation to the potential number of users of the service. A major concern is security. Authentication and authorization procedures will be explored using security protocols such as PGP or Kerberos. Questions about routing, multicast, and the assignment of both fixed and dynamic IP space and the use of MAC addresses also will be explored using DHCP. Tracking and analyzing length of sessions and time based patterns of usage will offer information about adaptation and user choice. Scalability of all operations will be addressed throughout the trial. Demand for a wireless service has begun to grow as the trial progresses. In particular, campus libraries are keenly interested in collaborative opportunities within their buildings. Individual users from both the faculty and the student population are requesting access to the development network. Basic information about the number of users, the applications employed, the bandwidth used and patterns of internet use will be of special interest. Any special characteristics of the software chosen for use by students on the move will be evaluated as network operation is tracked. The users will supply a fair amount of detail at the application layer. Each will be supplied with a laptop loaded with the following software: Dell standard software, Netscape manually configured to use proxy, IE5, Shockwave, Flash plug-ins, Nomadic IE5 automatically configured to use proxy, Office 2000 (Word, Excel, PowerPoint, FrontPage, Access, Outlook, PhotoDraw, Publisher, Small Business Tools), Bear Access (Cornell University gateway), QuickTime 4, RealPlayer G2 (7?), Windows Media Player, NetMeeting 3.0, Winzip, Adobe Acrobat Reader 4.0, Aironet LinkScope, Microsoft Terminal Server Client, Driver for Linksys Ethernet PC Card EC2T, PGP Requests for additional software, software additions by users, and patterns of software usage also will be reported. The faculty, research support personnel and the technical support personnel will be interviewed for their insights regarding deployment and use of the network. Two of the faculty members are from the Department of Computer Science and one is from the Department of Communications. The Director of the Engineering Library also will contribute and patterns of use within the library will be reported. Six students will be interviewed for their perspectives on mobile computing. These reports will add much to the evaluation and the service rollout in the fall of 2000. Given the network architecture chosen for this pilot the evaluation will develop criteria for network design, development and operation in four areas. The results of the evaluation will be reported in April. The four areas are: a) basic implementation and support b) coordinating all components of the wireless service c) the capacity of the network; and d) optimization. The TSED technical evaluation is designed to reveal several aspects of development for wireless computing and communications on campus. In particular we intend to investigate the adaptive patterns of users and their ways of shaping the technology and their uses for it to suit daily activities in the campus environment. Students and faculty are well suited to such an approach as most are sophisticated users pursuing defined pedagogical and learning tasks. It is this “layer of adaptation” and its implications for network design and operations that we hope to define and analyze through interviews to determine the architecture required for campus networking. Judy Boggess is a technology transfer consultant for Cornell Information Technologies. Dan Eckstrom is a senior network engineer also at CIT Fernando Neri has recently joined a consulting firm. All three are alumni of Cornell University.