Pervasive networks of artefacts require scalable solutions for maintaining large heterogeneous networks of devices with limited resources, possibly connected via wireless links prone to communication failures. We strongly believe that scalability and heterogeneity in adaptive systems cannot be achieved through heuristic methods based on designer’s experience or trial and error practice: as networks become more complex and larger, and pervasive systems move towards commercial applications, new methods and algorithms are mandatory both for the self-organization of pervasive systems and for assessing their performance.
In order to achieve this it is necessary to provide methods for role de?nition and communication primitives (inter-node communication and group communication) that hide lower-level complications that arise due to mobility, failures and/or constraints on the available resources of the network and provide the required security and trust. The research carried out in this WP will try to solve this challenge; it hinges on the methodologies and the algorithms developed in WP1 and complements the activities of WP3 to provide a set of communication primitives and infrastructures on top of which higher-level applications of WP4 can then be built.
Positive solutions should be compliant with a number of requirements. The ?rst one is given by the immense size of the network and the scarcity of resources; this implies that we need scalable and ef?cient solutions. We will use the tools of algorithm analysis to formally assess such requirements. Ef?ciency should be measured not only with respect to the computation time but also with respect to other parameters such as energy, memory requirement, required communication bandwidth etc. Furthermore, future commercial pervasive systems must also be easy to program; therefore we must provide such systems with appropriate distributed algorithms and protocols that use only local information and simple rules.
Another important issue is represented by self-stabilization. In fact, future networks of artefacts will be designed to execute forever, like an operating system. Thus it is highly unlikely that they never experience a transient failure (i.e., a temporary violation of the designer assumptions), especially in highly dynamic, wireless, mobile networks. Self-stabilizing distributed systems can recover after the occurrence of transient faults. The system is designed to automatically regain its consistency from any arbitrary state that maybe the result of unexpected faults caused by the environment. We believe that self-stabilization is a useful design criterion.
The above two issues have received limited attention in the study of networks of artefacts and represent the main novelty of our approach. We split the work into three tasks of the internal, perhaps continuous, self-organizational ability of the pervasive networks:
- Adapting the communication infrastructure
- Adapting the internal structure and roles of artefacts
- Security for tiny artefacts