Focusing on Mobility

In this paper, we motivate the importance of the field of mobile computing and survey current practical and formal approaches. We argue that the existing formalisms are not sufficiently general and powerful because they do not model all necessary concepts of mobility adequately. The main contribution of the paper is, therefore, to identify and define the fundamental concepts of mobile systems by providing a precise, mathematical foundation. The model we present is an extended variant of existing, compositional network models for control and data flow of non-mobile systems, enriched by the concept of locations as places containing components. To model the migration of a component from one location to another, the containment relation may change dynamically over time. Based on this formal model, we define a number of fundamental properties and characteristics such as network transparency. Finally, we demonstrate how existing description techniques may be extended in the context of mobility, and sketch a supporting CASE tool. 1 Why Mobility is Important Recent developments in hardware and software technology have created the infrastructure for a new computation paradigm: mobile computing. According to this paradigm, both hardware and software components may dynamically migrate between distributed locations. Mobility implies a fundamental paradigm shift in the area of computing. According to Carlo Ghezzi [1], for example, “the long term vision is that computers are no more viewed as mainly autonomous and self-contained computing devices accessing local resources, occasionally communicating with each other; rather, they are part of a global computing platform, built upon a synergy of local and remote resources, whose sharing is enabled by broadband communication networks”. Without doubt, mobile computing will enable a wealth of novel applications, for example, in the area of fieldwork construction, maintenance, and life-cycle management systems, as well as smartcard-based electronic commerce applications. In the following, we sketch some of the possibilities in aviation. During the production of an airplane, the respective manufacturers may integrate small computers with sensors into the single parts of the aircraft to trace the location of the parts during construction (which is performed by workers with wearable computers, of course). Once the airplane is completed, the integrated sensors provide actual information over the state of the parts, enabling proactive maintenance and replacement on need. During flight, the communication between its computers and stationary computers on the ground (and also with the mobile computers of the maintenance staff) has to be organized. Finally, the passengers and their various mobile computing devices must eventually be taken into account, ranging from smartcards used to pay calls over the airplane’s phone system to laptop computers connected to the internet via the airplane’s local area network. Scenarios like this involve a variety of different hardware and software components with various, dynamically changing relations between them. Currently, an integrated view on mobility is still missing. Research and development is mostly focused on relatively isolated areas, like mobile hardware design, mobile data management [2, 3], and protocols for mobile and wireless networks [4]. These areas are exploited by integrated technical platforms for the development of mobile systems. Such platforms try to reach a high degree of mobile transparency, so that programmers are able to implement a mobile system as they are used to implement a non-mobile system. There are three groups of platforms: The platforms in the first group are based on traditional programming platforms for distributed systems, like CORBA [5] and DCOM [6]. Based on their facilities for transparent communication, support for mobile code and mobile data may be implemented [7]. The second group of platforms has its roots in the area of distributed operating systems. Odyssey [8], for example, enhances the distributed operating system Mach [9] with additional features like hierarchical namespaces and code as well as data resources. Mobile