دانلود رایگان مقاله انگلیسی کشف سرویس کارآمد در شبکه های اجتماعی آنلاین غیر متمرکز - الزویر 2018

abstract

Online social networks (OSN) have attracted millions of users worldwide over the last decade. There are a series of urgent issues faced by existing OSN such as information overload, single-point of failure and privacy concerns. The booming Internet of Things (IoT) and Cloud Computing provide paradigms for the development of decentralized OSN. In this paper, we build a self-organized decentralized OSN (SDOSN) on the overlay network of an IoT infrastructure resembling real life social graph. A user model based on homophily features is proposed considering social relationships and user interests and focuses on the key OSN functionality of efficient information dissemination. A swarm intelligence search method is also proposed to facilitate adaptive forwarding and effective service discovery. Our evaluation, performed in simulation using real-world datasets, shows that our approach achieves better performance when compared with the state-of-the-art methods in a dynamic network environment.

7. Conclusions

Cloud Computing and IoT are revolutionary technologies that are powering ubiquitous wireless communication and real-time computing to expand the Internet-connected automation into new application areas. This paper examines the problems of current OSN and proposes a self-organized decentralized architecture for future online social networks to enable efficient service discovery in dynamic social IoT environments.

This decentralized architecture fills the research gap between the traditional P2P infrastructure and the decentralized architecture of OSN, which can alleviate the rigid privacy-control problems, as well as provide adequate flexibility and capability for service discovery. Then a homophily-based user model is introduced to integrate social relationship and user interest. This model is able to identify promising neighbors having considerable similarity to the potential service provider as well as having the high number of connections. Further to achieve a bandwidth-efficient search, a novel OSS algorithm has been proposed to provide an adaptive forwarding degree in each routing hop. This algorithm utilizes the olfactory sensibility and pheromone information of swarms to discover the shortest paths with maximum desired service. Moreover, a Java based simulation platform has been designed to simulate a dynamic unstructured P2P network with knowledge-based routing protocols for service discovery in DOSN. The proposed user model and OSS algorithm have been simulated under different network scenarios to evaluate its efficiencies against a range of state-of-theart search models in unstructured P2P networks. Experimental results demonstrate that our approach achieves better performance under three dynamic network structures than the compared models.