Notre thèse interroge la notion de communication virale dans les espaces socionumériques de manière générale et plus particulièrement lorsque ce phénomène s’applique aux vidéos publicitaires en ligne. Nos recherches ont révélé un manque évident de clarté et de cohérence au niveau de sa définition et de son acception (Beauvisage et al., 2011) qu’il a fallu corriger par un travail d’harmonisation avant d’envisager la suite de nos travaux. De plus, notre revue de littérature a mis en avant la complexité du phénomène viral et de son appréhension ; complexité nourrie par le nombre et la nature des facteurs à son origine (Beauvisage et al., 2011). Afin de mettre en avant un de ces facteurs, nous émettons pour première hypothèse que les variations de qualité de la définition (haute ou standard) d’une vidéo influencent l’appréciation de la vidéo (H1). Par corollaire, nous pensons que le partage de cette vidéo est affecté par la qualité de la définition (H2). Plus précisément, nous pensons qu’une même vidéo sera plus partagée si elle est visionnée en haute définition plutôt qu’en définition standard. Pour répondre à ces hypothèses nous avons opté pour une approche expérimentale. / Our thesis examines the notion of viral communication in digital social spaces both in general and when applied to online video advertisement. Our research revealed a lack of clarity and coherence in its definition and meaning (Beauvisage et al., 2011) that necessited an effort of standardization before planning to pursue our work. Furthermore, our literature review pointed out the complexity of the viral phenomenon and its comprehension. This complexity is due to the number and different factors originating the viral phenomenon (Beauvisage et al., 2011).In order to highlight one of those factors, we make the first hypothesis that the variations in the quality of the definition (high or standard) of a video have an effect over the evaluation of the video (H1). Corollary, we think that the quality of the definition affects this video sharings (H2). More precisely, we think that a video will be more shared if it is watched in high definition rather than in standard definition. In order to meet those hypotheses, we opted for an experimental approach.

As humans, we tackle many problems in complex societies and manage the complexities of networked social systems. Cognition and sociability are two vital human capabilities that improve social life and complex social interactions. Adding these features to smart devices makes them capable of managing complex and networked Internet of Things (IoT) settings. Cognitive and social devices can improve their relationships and connections with other devices and people to better serve human needs. Nowadays, researchers are investigating two future generations of IoT: social IoT (SIoT) and cognitive IoT (CIoT). This study develops a new framework for IoT, called CSIoT, by using complexity science concepts and by integrating social and cognitive IoT concepts. This framework uses a new mechanism to leverage the friendships between devices to address service management, privacy, and security. The framework addresses network navigability, resilience, and heterogeneity between devices in IoT settings. This study uses a new simulation tool for evaluating the new CSIoT framework and evaluates the privacy-preserving ability of CSIoT using the new simulation tool. To address different CSIoT security and privacy issues, this study also proposes a blockchain-based CSIoT. The evaluation results show that CSIoT can effectively preserve the privacy and the blockchain-based CSIoT performs effectively in addressing different privacy and security issues.