3.3 Research Programme Themes
The M-Zone research programme is structured into a number of research streams, which
are described here as research themes. These research themes represent the main scientific
and technical issues to be addressed in the development of ubiquitous management for
smart spaces. Ubiquitous management serves as an umbrella for investigations targeting
the integration of separated or already converging technologies. This includes network
connectivity and configuration, managing a smart space, and interoperability and cooperation
between M-Zones. The research themes are intended to provide concepts and
solutions for the usage, operation, control, administration, and maintenance of smart
spaces, in other words the establishment of managed zones.
The research themes, enable researchers from the three partners to interact and cooperate
to particular research topics and scientific challenges without the strong boundaries of
formal work packages. In the current stage of the project, we have identified three major
research themes. These themes will be reviewed and if necessary revised within the
project life cycle in order to adapt them to the developments of the international research
community. On the other hand, these themes can be used as an instrument to influence
international research activities. Therefore, researchers from the international
collaborating research centres will also contribute to a number of the research themes.
The identified research themes are the following:
Managing Dynamic Environments
The focus of managing dynamic environments theme is to develop a framework for the management of dynamic environments where different technologies inter-work to give smart connectivity to users, user applications or devices. The theme will be driven by concepts broadly in the areas of policy-based management, context aware computing and networking technologies. Scenarios demonstrating the application of smart environments will be developed in theory and possible implementations thereof will be investigated.
The objective of the theme is not limited to development of architectures to deliver Internet connectivity to users. It extends beyond conventional connectivity in that it aims at providing not just connectivity, but communication, between devices that users encounter in their day-to- day life. The environment needs to be aware of what the user wants and also what it can deliver. It should be able to receive and analyse the feedback it receives from the user. The feedback should enable the environment to decide if its analysis of the needs of the user was correct. If the analysis of the user needs was inaccurate, then the environment should learn and improve on its knowledge of user needs. To achieve this, sensors, processors and actuators will need to embedded in everyday products and devices. An Ultra WideBand (UWB) transceiver system is being designed and implemented as a system on a chip that will offer the possibility of interworking the millions of sensors that will be used to provide information about the environment. There will also be a need to design interfaces, using which, the user will be able to interact with its environment. The interfaces will need to be simple to interact with and may have the capability to understand speech, actions etc., though the active interaction from the user side is aimed to be kept at its minimum. The interface will adapt to the user's context, learn through use and automate trivial tasks. The current theme on managing dynamic environments will have an interaction with theme 1 which focuses on adopting to users' needs and will be able to define and design such interfaces. The environment will need to hold information about its users and structure it so that the data collected from one source can be reused for a separate task, thus narrowing the gap between what it offered to the user and what the user actually wanted.
The support for smart environment will be enabled by intelligent software and involve using a heterogeneous network of devices with varying levels of intelligence. The network medium will be both wired and wireless. Many communication technologies like bluetooth, WLAN, UMTS etc. will have to be used depending upon the environment. For example, appliances may communicate with each other using bluetooth, while WLAN could be used to give the user high speed connectivity to the Internet when the user is near a hot spot or, using UMTS if WLAN is not available.
The theme will also investigate the impact of the high volume of traffic that will result from the environment aiming to provide advanced context aware services to its users. Research on network specific issues like intelligent call admission control strategies will enable the environment to choose, for example, the air interface to use when more than one are available. Such decisions, whether network or environment initiated, will need to take into account the network perspective as much as the user perspective, for the network to operate efficiently. The analysis of these issues will also be based on network specific strategies such as call admission control, which the network could decide using policy based rules.
Adapting to a User's Needs
This theme investigates techniques and models that support the autonomous adaptation of systems to match user needs to available service capabilities and the current context. Autonomous adaptation by system is essential in smart spaces, where the user's context and the operating environment are constantly changing due to movement of the user, movement of the user's peers and volatile load and connectivity characteristics. This theme focuses on aspects of adaptivity that are driven by implicit rather than explicit direction from the user and which therefore need to rely heavily on context information and learnt user patterns.
Mechanisms for adaptation may reside within a single software application, however smart space users will often not have continuous access to a stable, fully functional set of applications, and will have to work in environment with disparate sets of available application functionality. One way this functional heterogeneity is addressed in this theme through taking a service-oriented approach. This involves modelling and accessing of application functionality through well defined interfaces expressed in a common format such as the Web Service Description Language (WSDL), which despite it name is compatible with protocols in addition to HTTP. A service-oriented approach allows for the dynamic assembly of available services into composite services that may be designed on the fly to meet current user needs.
Service interfaces at the level of WSDL provide only syntactic interoperability information. For services to be involved in any autonomous adaptation they must expose richer semantic information about the functionality they offer, in a form that can be exchanged and reasoned about by automated reasoners and other forms of adaptive agents. Many previous efforts to model such semantic have suffered a focus on one domain and hence a limited scope of applicability. The expansive array of user tasks and application domains that can benefit from smart spaces requires an approach where domain specific models are not restricted by implicit assumptions. The use of ontologies is therefore of interest in this theme, since they provide taxonomical information on a domain accompanied by inference rules which can be used reasoning agents.
The context in which adaptivity is performed is of vital importance in ensuring adaptive behaviour accurately reflects the user's needs. The interpretation of user actions is highly dependent on a good understanding of the context in which those actions are performed. Context can also be historical, based on past user behaviour and feedback from users on related adaptive actions, thus allowing adaptive systems to improve their accuracy through learning. As context information must present a view of the real world, this is an ideal area to exploit the ability of ontologies to model aspects of the real world.
The management of adaptive services is addressed in the theme "Managing Dynamic Environment", which addresses the mechanisms for exerting explicit human control over the collective behaviour of adaptive systems and the technological mechanisms that support them. The engineering concerns related to the development of adaption-ready services and the ontologies which they use are addressed in the theme "Seamless Engineering of Open Smart Spaces".
Seamless Enineering of Open Smart Spaces
As Managed Smart Spaces are interconnected, the challenge is to ensure that a seamless flow of users (people and devices) across the Smart Spaces is provided to give seamless M-Zone interoperability. The creation of such an environment is a major challenge for the future direction of smart spaces.
At present, the smart space research community has not addressed the issue of smart space interoperability. However, the telecommunications management community has addressed in considerable detail the issue of inter-domain management systems for cooperating management domains. Analyses of the requirements for the creation and management of such interoperability/co-operation may be discussed in the context of the functional areas of fault configuration accounting, security and performance. While there are important differences between the systems, the telecommunications management approach is a valid starting point for the development of solutions for the seamless interoperability of M-Zones.
Security Requirements: All smart spaces and the information contained therein must be managed against unauthorised access and manipulation through the application of an appropriate security policy. Users entering the M-Zone must be validated and requests for access to information must be authenticated.
Information Management: The seamless interoperability of M-Zones must include a framework of access and manipulation and sharing of information between the cooperating zones.
Adaptive Information Delivery: using knowledge of people, place and devices is an important issue for the development of seamless M-Zone interoperability. The issue of what information is accessed from an external M-Zone and how this information is presented to the end user is an important management consideration. When roaming, information such as user profile information may be made available to the visiting MZone to support user service requirements. The transformation of information from one zone to another and the management of the persistence and integrity of information is a major management challenge. In many cases, information may be integrated to provide additional context to the end user. The management of information modification in response to external operation must also be addressed.
Performance Management: M-Zones must co-operate to support user performance requirements as they migrate from one M-Zone to another. A common example is the mobile phone scenario where users who roam expect that there is no loss of service quality as they move from one operator domain to another. The problem of mobility between M-Zones is more challenging. Unlike the single service scenario of today’s mobile phone network, the user will require access to a set of services. In addition, access to mobile information is greater and more complex.
Configuration Management: As in the mobile phone environment M-Zones need to be configured to support user (people and devices) into and out of M-Zones. Issues such as IP address allocation and Bandwidth allocation need to be addressed against the user service requirements.
Business Process Models: Just as is the case in the development of solutions for the development of interoperable telecommunications networks a business process model can be developed to support the interoperability of M-Zones. The TeleManagement Forum Business Process Model is a good starting point for the development of such a model.
Several infrastructure concepts that underpin inter-domain management in telecommunications are also very valuable in realising co-operating M-Zones. These concepts include Service Level Agreement negotiation, policing and enforcement, as well as real time selection, deployment and enforcement of operational policies within a managed area.
Standardisation work on interoperability issues for telecommunications network is a valid starting point for addressing the M-Zone requirements. However, there are a number of important differences that make the M-Zone environment particularly different. The first of course is the non-hierarchical nature of any future M-Zone environments. Also, a more lightweight solution will probably be more appropriate.
Summarizing this research theme, the major challenges are to handle information access (across smart spaces), personal communication (e.g. user-centric approaches) and to enable remote facility control. At the current stage, information access and management is provided by the exchange and the evaluation of profiles. Personal Communications require investigations on how the relationships of a user to one or more smart spaces can be modelled and what management services are needed to support this very user.
|
