Exploiting Logical Mobility Techniques in Physically Mobile Environments

With the recent developments in wireless networks (802.11, Bluetooth) and the sales of mobile computers of any kind (such as laptop computers, Personal Digital Assistants (PDAs), mobile phones etc.) soaring, we are experiencing the availability of increasingly powerful and mobile computing environments, roaming between different types of network connectivity. We have also recently witnessed the acceptance of logical mobility techniques, or the ability to ship part of an application or even a complete process from one host to another. The increasing popularity of the Java programming language and environment has largely been correlated with the acceptance of logical mobility techniques, due to the inherent code mobility infrastructure that Java provides. To facilitate application developers programme for mobile devices, mobile middleware systems have been developed, which rise the programmer abstraction beyond the network os layer, and tackle issues of heterogeneity, communication primitives, concurrency, security etc. Whereas various mobile middleware systems have been developed the use of logical mobility has been very limited. The purpose of this work is to investigate the use of logical mobility in mobile computing environments, our hypothesis being that it can bring tangible benefits to both developers and users which cannot be supported by the current state of the art. While investigating why the use of logical mobility has not been widely adopted by the mobile industry, we have found the following problems: There is currently no design stage methodology which allows application developers to model the effects of logical mobility in their applications at the design stage. This limits the development of mobile applications powered by logical mobility to simply trial-and-error scenarios, inhibiting their popularity. Moreover, we have found that there exists no mobile middleware that offers flexible use of logical mobility primitives to application developers, whilst allowing devices to traverse dynamically different types of networks, handling heterogeneity, service advertisement and discovery, communication etc. As such, applications using logical mobility primitives have to constantly reinvent the wheel in order to implement the solutions. We believe that these deficiencies in this area of research are closely related, being aspects of the same issue, one affecting the design stage, and the other the implementation stage. This work intends to address these deficiencies, by researching a design-stage methodology, which allows developers to evaluate the use of logical mobility paradigms in their application, and developing a mobile middleware, which will be targeted by the aforementioned methodology and will offer the …