Decentralized online social network primitives

• Main Author: Rajesh Sharma
• Other Authors: Anwitaman Datta
• Format: Theses and Dissertations
• Language: English
• Published: 2013
• Subjects: DRNTU::Engineering::Computer science and engineering::Computer systems organization::Computer-communication networks
• Online Access: https://hdl.handle.net/10356/54955
• Institution: Nanyang Technological University

Centralized online Social networks such as Facebook are highly popular [1]. Despite their success, a prominent concern about such centralized systems is privacy [2] [3] from the service providers themselves. In order to address these shortcomings, researchers have proposed decentralized architectures for social networks. Examples of such Decentralized Online Social Networks (DOSNs) include Peerson [4], Vis-a-Vis [5] and Safebook [6]. Typically in a DOSN each autonomous entity has a local view of the network based on social links it has. In other words, in the absence of global knowledge because of decentralized architecture, a node can only explore its local information, i.e., its own knowledge as well as knowledge shared by its neighbors related to it using some social relations (friends, relatives etc.). This is challenging due to the fact that problems need to be solved without global coordination of local information and local links. We explore \distributed social computing" to solve various challenges in a DOSN. Social computing is an area where the social behavior plays an important role for any computational systems. Distributed social computing is an emerging area which focuses on solving problems in distributed systems by leveraging social relations (such as interest, trust, expertise, etc.). In a centralized system, collective wisdom of the users is relatively easy to gather to solve a problem. However in a decentralized system, it is not trivial to achieve the same because of the typical characteristics of the environment of decentralized systems - (1) network is decentralized and scalable, and (2) no node can have global knowledge about the system because of the decentralized nature of the network (or because of resource constraint). In this report, we investigate issues and challenges in a DOSN that can be termed as more di cult than that in a centralized online social network as there is no central authority, which can cater users for various services. We solve various problems in DOSN in a distributed manner with constraint of social links. Speci cally we aim at social network applications like storage, information dissemination and ranking. The main contribution of our research is that in our proposed solutions, nodes exploit only their local knowledge (i.e. their own information and information gathered from their immediate neighbors) by using distributed social computing approach. We present several results which show the e ectiveness of our approach. We also highlight various future directions with respect to our current work.