Users don’t care about technologies; they want services that fulfi l their needs and they care about prices. As interesting and exciting technology advances develop over time, the real mobile adventure for users consists of discovering new services and running them with ease. the evolution of mobile communication systems from the 1st to the 4th generation, highlighting their respective major characteristic is shown below.
Throughout their history computers have become smaller and cheaper as a result of continuous advances in digital electronics. At the same time, processing power and storage capacity have grown at an amazing rate. The combination of these two trends has allowed the introduction of computational capabilities into places and devices that would have previously been impossible in the earlier state of technology. In addition, it can be observed that more and more everyday items possess embedded wireless interfaces, as depicted, allowing networking amongst them. This opens the door to new and interesting business opportunities, where everyday life support, healthcare, robotics and logistics are just a few of the possible application areas. In the near future people will be surrounded by ever more computers, leading to the situation in which we are ‘dwelling’ or living amongst them. Such a scenario is described as a ubiquitous service and networking environment, an execution environment in which over time users will gradually be surrounded by computational resources and tiny networked devices.
A great opportunity in leveraging the real power of mobile applications as Services with Initiative is to aim to provide the ability to detect and react to environmental variables at service provisioning. Environmental variables – usually described as the context – consist of accessible information about the user’s situation at service runtime. The assessment of the user’s situation can vary from the interpretation of a single contextual data item, such as location, to composed inputs based on multiple data items. Context categories include, but are not restricted to: computing context, for example network connectivity; time context, such as the time of the day; physical context, such as the lighting level; the relationship between the application and nearby objects; and even the user’s social situation. The context also may include aspects such as interactions previously performed between the user and the service (application history) and future expected relationships. The illustration below depicts the situation where different contextual inputs are observed and combined to create a specific context space, relevant to a particular user or user group at the time of service provisioning.
Context-Aware Service Provisioning, based on Context Spaces, allows the establishment of a framework for the creation of scoped contextual boundaries and the sharing, storage and selection of context entry objects within those boundaries. However, even the most complex context space scenario only represents a partial view of the full set of situational variables. For example, a user can move with an application from one situational context to another and in the new context be challenged with even more contextual heterogeneity. Moreover, further environmental properties representing dynamicity, continuity and nondeterminism increase the complexity of grasping the environment. The state of the environment will change dynamically, independently of the actions of the service. There is no discrete fi xed number of states forming a context. Furthermore, there is no guarantee that the available state information is accurate and the service will have only partial control of the state of the environment. Within such a diverse understanding of the context, Services with Initiative have to react within a highly dynamic scenario that is unlike any one that has existed previously. Handle the environment described above one needs effi cient ways of:
- coherent and distributed context modelling
- context search, information gathering and abstraction
- context grouping and fi ltering
In addition, the following will be needed:
- adaptive algorithms capable of gathering knowledge from samples
- ambient reasoning components to interpret the situation at hand
- generalization mechanism to apply that knowledge to new situations.
