$gdswlyhh6fdwwhuqhww6xssruwwiru%oxhwrrwkkxvlqjj6qliii0rgh 8qlyhuvlw\\rii%rqq,qvwlwxwhhrii&rpsxwhuu6flhqfhh,95rhphuvwudvvhh'%rqq*hupdq\ ^eddw]pdwwkhzpduwlqlvfkro]`#fvxqlerqqgh

While the market for fixed, wired personal computers shows stagnating tendencies, mobile computing receives more and more attention. The terms # : 0 and 0 have moved into the center of interest. Truly mobile devices such as mobile phones, digital watches, and personal digital assistants (PDA) become more and more powerful. At the same time, notebook devices are decreasing in size and weight. As the Internet becomes a part of everyday life for many people, there is not only the need for pervasive computing but also for & 0. Up to now, users that want to connect for example their notebook to a mobile phone have to employ expensive, proprietary cabling or use infrared communication technology that requires devices to have a free line of sight between them. Bluetooth ([1], [2]) is a radio technology that promises to be a very convenient, low-cost solution for the interconnection of all kinds of mobile devices. This cheap but powerful technology will be available in a wide range of mobile devices. Thus, building Bluetooth ad hoc networks connecting all sorts of devices will be a common application for Bluetooth devices on conferences, meetings, and so on. Up to now, Bluetooth defines a simple network topology (piconet) that only supports a limited number of devices and requires all devices to be in range. However, support of the more demanding multi-hop ad hoc network topologies (so-called scatternets) is not specified in detail yet. This paper proposes a mechanism to support scatternet communication by (re)using mechanisms currently specified and thus provides the basis for Bluetooth ad-hoc networking. The paper is structured as follows. Section 2 gives a short summary of the Bluetooth technology. In section 3, basic properties of Bluetooth scatternet support are presented thereby explaining its demands and challenges. Section 4 features our adaptive approach to scatternet scheduling. Section 5 illustrates the properties of our approach by presenting simulation results in basic scenarios. Finally, section 6 presents our conclusion and describes future work.