Design of Authentication and Privacy Schemes for a Generic Social Network

Abstract

In social networks, people (also known as actors) are often connected to each other in order to share the information. With the advancement in the information technology, social networks have in influenced actors of different regions to share information based on profile, preferences, relations, etc. The main goal of social networks is to create the information space, where actors can share information like thoughts, personal data, events, etc. They enable actors to connect with other actors of similar interests when physical barrier might have kept them apart. The popularity of social networks have increased because they enable researchers to study actors based on their characteristic features. The objective of the work is to address design of a generic social network, authentication of actors, privacy preservation and data access control of actors. We present a generic social network design by using agent technology. The generic social network enables to draw a specific real-time social network. The generic social network establishes relations among actors who share common interests, activities or connections. We have used agents (both static and mobile), where the Static Agents (SAs) reside at the Service Provider (SP) and deploy Mobile Agents (MAs) to various groups of a generic social network. The MA at a group analyses features of actors such as personally identifi able information, professional information, activity, social status and history, and contacts the SA to form hierarchical and equivalence relations among actors. As a case study, we have designed real-time agricultural social network and academic social network by using the generic social network as a tool with speci fic data tested on various conditions. We propose a novel authentication scheme by considering actors' characteristic features such as professional information, activity, relation, level, etc. In this method, we use a combination of a group leader and a centrally monitoring service provider authentication. The SA residing at the centrally monitoring service provider is responsible for authenticating a group. While MAs, in coordination with Group Leaders (GLs), are accountable for validating rest of the actors of a group in a generic social network. We designed a privacy preserving technique by considering actor's interests, preferences, history, etc. In this scheme, intelligent agents are deployed, where the SA deploys MAs for ensuring enforcement of privacy policies over a group of actors. The designed system is tested over an agriculture social network in Flickr, where we simulated different sets of actors related to the agricultural occupation. We introduce an agent based approach for an authentication based data access control scheme for generic social network actors. In this method, intelligent agents are deployed for determination of authentication of actors based on their professionalism in a generic social network. Subsequently, the data access control of actors is defi ned based on their pro le, activity, relation, etc. The local administrator is responsible for fairly enforcing data access control policies over a group of actors. The data access control studies the actor misbehaviour and invalid data access dynamically, which leads to stoppage of access of actors to data. The designed system is tested over the ASN where we simulated authentication of actors along with their data access in Flickr using OAuth 2.0. We have exhaustively tested the proposed generic social network system by allowing access to actors over soil, seed and crop database. We considered 200 actors associated with agriculture occupation and simultaneously provided them with authentication, privacy and data access control. In summary, we have designed; 1) a generic social network using agents which utilised different character- istic features that create hierarchical and equivalence relations among actors; 2) a two-level authentication scheme for generic social network actors by considering actors characteristic features such as professional information, activity, relation, level, etc., using intelligent agents (both SA and MA); 3) a privacy preserving technique for actors in a generic social network by considering actor's interests, preferences, history, etc., using agent technology; 4) data access control scheme for generic social network actors by de fining direct or indirect access using intelligent agents; and 5) testing of the generic social network system over real-time database. We simulated our proposed schemes in different social networks simulation environments. The results obtained in simulation show the importance of the approaches and the efficiency of the proposed schemes.