Abstract
Ubiquitous computing environments (UCEs) are characterised by a myriad of heterogeneous devices fitted with computing and communication capabilities - many of them embedded in physical objects and usually imperceptible - whose main goal is to cooperate in a coordinated manner to supply users with a pool of services that facilitate their tasks. Communication among the participating devices is predominantly performed through the wireless medium. UCEs can also be partially or completely reliant on infrastructureless wireless multihop communication, which provides them with full network connectivity, great flexibility and a versatile dynamic network topology that can be desirable in many situations. However, UCEs' wireless (often multihop) nature also makes them inherently susceptible to having their communication operation disrupted due to their dependence on the cooperative packet forwarding behaviour of each individual node. For example, misbehaving nodes can cause general network disruption by not forwarding any or only forwarding some packets on behalf of other nodes in the network. Consequently, protection of the data forwarding functionality against malicious or otherwise defective nodes is an important characteristic that UCEs must support. This thesis proposes an adaptable protection scheme that can detect, accuse and penalise misbehaving nodes that disrupt the communication capabilities of a UCE by dropping data packets that they are expected to forward on behalf of their peer network nodes. This thesis presents different aspects of the design and implementation of the protection scheme. Among the most important aspects are: i) collection of information for behaviour evaluation, ii) accurate detection of misbehaving nodes, iii) accurate and effective accusation of nodes persistently misbehaving, iv) adaptability through network management policies, v) network clustering and a role-based organisational model, vi) resilience to nodes that report false metrics, and vii) resilience to colluding nodes. Concepts developed in this work are illustrated in the context of mobile ad hoc networks (MANETs) since they have emerged as an appropriate paradigm to enable the deployment of UCE technologies. The efficiency and effectiveness of each aspect of the protection scheme are evaluated and demonstrated through extensive simulations.