Home > Research > Publications & Outputs > Interoperating with Services in a Mobile Enviro...

Electronic data

View graph of relations

Interoperating with Services in a Mobile Environment

Research output: Working paper

Publication date2003
PublisherLancaster University
<mark>Original language</mark>English


Mobile computing is characterised by users carrying portable devices that allow communication between people and continuous access to networked services independent of their physical location. A mobile application must discover and interoperate with the required application services available to them in their present location. However, these services will be developed upon a range of middleware types (e.g. remote method invocation, publish-subscribe, message-oriented and tuple spaces) and advertised using different service discovery protocols (e.g. UPnP and SLP) unknown to the application developer. Wireless environments are also subject to changes in network QoS and connectivity, and mobile applications operate on devices with limited resources. Given these properties, existing middleware technologies are inappropriate to support mobile client-based applications; developing upon a single middleware platform restricts the possible mobile services that can be interacted with. Therefore, a middleware platform for mobile computing must adapt its behaviour to interoperate with any type of discovered service, provide the best level of service in an environment that is frequently changing and finally, be lightweight in resource use due to the constraints of mobile devices. This paper proposes reflection as a well-suited technology to implement the adaptive features of mobile middleware and identifies what the key requirements of a dynamic middleware platform are. In addition, we describe ReMMoC (Reflective Middleware for Mobile Computing), a middleware platform that dynamically adapts its structure to interoperate with a range of middleware types that may exist in the mobile environment. Finally, the use of ReMMoC in a typical mobile scenario is presented and the memory footprint cost of utilising reflection to create a mobile middleware platform is evaluated.