ALLOW: Adaptable Pervasive Flows

Main Project Website: Allow

EU Framework 7 Project. Start Feb. 2008

Naranker Dulay, Morris Sloman

Other Partners

• University of Stuttgart - Lead Partner

• University of Passau
• Institute for Scientific and Technological Research (ITCirst), Trento
• Lancaster University

Human users are increasingly embedded in an environment consisting of uncountable computing devices and artefacts that provide various degrees of computing power and awareness. Current estimations state that ten years from now, there will be 7 billion people surrounded by 7 trillion wireless devices and sensors living on this planet. Pervasive applications are software systems that run in such environments in a massively distributed fashion and support mobile human users in their daily activities. This vision holds several scientific challenges ranging from adequate hardware architectures to communication protocols and aspects of software distribution. The most challenging question, however, is: How can a pervasive application adapt to the user in order to support him/her in an unobtrusive way? Ideally, the application shall run in the background unnoticed by the user and adapt to his actions. This requires new paradigms for programming such applications. Today, the interaction with computer systems of any kind is still done explicitly and manually in most cases. However, this is not an option for pervasive systems that shall offer support to a wide spectrum of users that may not be capable of dealing with the specificities of computing devices, even more so as these devices may not even be visible. Moreover, the necessity of having to adjust some pervasive application explicitly represents a major burden, contradicting the initial intention of pervasive computing. The same is true if users have to adjust themselves to such an application. Therefore, pervasive applications must be able to adapt themselves to the user autonomously in order to be feasible. However, there is a considerable lack of technologies for enabling this kind of adaptivity. Hence, we propose the ALLOW project that will be dedicated to developing Adaptable Pervasive Flows as a key enabling technology for pervasive applications.

Objectives

The core objective of the project is to develop a new programming paradigm for human-oriented pervasive applications. This paradigm will enable pervasive technical systems to adapt automatically and seamlessly to humans involved and embedded in them, explicitly supporting people in achieving well-defined goals in dynamically changing environments and contexts. Furthermore, the results of the ALLOW project will enable the integration of humans into pervasive business and working processes in an unobtrusive way. One of the major goals of the project is to design this technology in such a way that the resulting environments are secure and trustworthy.

Key Concepts

The key concept that will be developed within the project in order to realize these objectives is called Adaptable Pervasive Flow (APF). Many processes in real life are defined in terms of flows either implicitly or explicitly. A flow as we define it in this project, is a computer-based model that essentially consists of a set of actions that is glued together by some kind of plan (or control flow) for their execution in order to achieve some goal under a set of constraints. This resembles the well-known workflow concept which has proven successful in various fields such as business workflows and service-oriented architectures (SOA). However, APFs represent a much broader concept that ideally fits the problem of creating adaptable pervasive applications. They are situated in the real world, i.e. they can be logically or physically attached to entities like artefacts or people, moving with them through different contexts. While they are carried along, they model the behaviour intended for their entity, and adapt the entity’s environment to this behaviour. Thus, when a mobile user carries a flow that specifies his prospective actions, the pervasive computing machinery will be set up for him by the flow, wherever he goes and whatever he may intend to do there. Since people may change their minds, and since artefacts and people may be subject to changes in the environment, the flow itself may also adapt to reflect such changes. This requires flows to be context-aware. They can sense the context pertaining to their entity’s current environment as well as the entity’s actual activities. In the proposed project, this far-reaching and visionary concept of flows will be researched, new mechanisms, methods, and principles for adaptable flow-based pervasive applications will be investigated, and a respective technology will be developed as a new programming paradigm for adaptable pervasive applications.

The Imperial College focus will be on development of a language and techniques for specifying the security, trust and privacy of pervasive flows.