Context awareness will be a major enhancement for future mobile systems. This holistic approach aims to adapt the communication system to the current environment of the user. To optimize services in a heterogeneous mobile network environment, the use of context information is essential. While second-generation mobile networks provide a very homogenous network, both in terms of network topology and network services, this will change toward a more heterogeneous environment in future mobile networks. Network services (multimedia communication, high-bandwidth data services) will be available at different access points in different ways. Thus, providing optimal services to the user will be more challenging for several reasons:
- There will be more diverse radio access networks (for example wireless LAN, second- and third-generation cellular networks and their variants, and other upcoming technologies such as ad hoc networks).
- There will be many more options concerning network services (for example QoS, security, charging, roaming, etc.).
- Applications and user preferences evolve fast and will require optimal support from the network services. _ Advanced location information and group mobility information (for example users travelling in a car) will be available to support the services.
Signal strength and the availability of radio resources are common types of information to be considered for a handover. However, they are not enough to provide efficient handovers. Even if one access point is slightly better regarding these local measurements, the decision based on them may not be the best. A significant problem is that handover decisions have to be executed fast. However, the node profile and location information is often available on a central server in the core network. Retrieving this information may be too slow for handover decisions. Furthermore, the radio conditions during handover may be poor and hence limit such information exchange. Another problem in future networks is that scanning for access points of different radio technologies can be expensive (with respect to computation power and energy consumption) or even impossible for devices that only support one mode at a time. In this case, context aware assistance of the network to avoid unnecessary search for access points can be very valuable. With knowledge about the possible access points, the number of reconfigurations to another mode can be minimized. The illustration below shows an architecture for context-aware handover.
The main entities defined in this architecture include; handover decision point, which decides on the access point to be taken for handover and context collection point in the network side, which collects and compiles the relevant context information from different sources; this is then delivered to the handover decision point.
The classical approach for context exchange is to define a context information exchange format and attributes. In addition, a protocol for exchanging context information between the context collection point and the mobile node is needed. This means that interpretation of the context information should be fixed and possibly standardized. The appropriate algorithms need to be implemented in the nodes. The algorithms to collect the context information on the network side can be specific to the network or the location.
The handover decision takes place inside the mobile node, but is possibly controlled by the network with the supplied context information and rules. This is important, because many pieces of dynamic context information are only available on the node (for example the signal strength). The level of control from the network side just depends on the logic implemented in the mobile node and how it reacts to network context information. Based on the context-aware handover concepts highlighted above a more general framework for context management can be illustrated as below;
Such a framework should be able to collect, process and distribute non-structured and short-lived information, which can be located in distinct network nodes and devices. This framework should also ensure that the right information is available at the right time to a wide range of applications. An important issue in the design of such a context management framework is the consideration of different application requirements. The requirements of applications can differ, for instance in terms of the information exchange patterns, of the data model of context information and of the granularity and amount of the transmitted data. The last of these is particularly important in mobile networks, since scarce wireless resources limit the amount of information that can be exchanged with wireless devices. Another important requirement is the capability to use cross-layer interfaces efficiently to collect information in different protocol layers.
The main goal of the framework is to facilitate the collection, pre-processing, distribution and use of distinct context information by a wide range of applications. The first obstacle to creating a general-purpose flexible framework is the application-specific nature of the context information to be collected and the distribution mechanisms to be used. To solve this problem, the context management framework is divided in two distinct parts. One subset of modules is generic to any type of application, while the other is application specific, as illustrated in above.
