SocialOverlay : P2P Infrastructure for social Networks
1. Social Overlay: P2P Infrastructure for Social Networks Bipin Upadhyaya*, EunmiChoi Distributed Information Systems Lab Kookmin university Seoul
2. Outline Introduction Problem Statement Related Work Social-Content Graph System Architecture Handling FOAF Information Social P2P Layer Conclusions & Future work
3. Introduction Current search technology does not utilize the relation between the people-people and people-content. Need for more personalized search result. Peoples participate in social sites. Privacy as an important issues. Example : Google search on cute baby gives
4. Introduction A typical person is involved in Social networks like facebook for family and friends Professional networks like LinkedIn. One among many bookmarking site del.icio.us, digg, simpy, CiteULike Other book, movie recommendation sites. So your friends are clustered among different sites.
5. Introduction Facebook Which network too join ?? MySpace Imagine the situation when user wants to share multimedia, bookmarking, recommendation and other Web2.0 application LinkedIn Orkut
6. Problem Statement How can we infer the social linkage between the users in different sites? How can we identify different clusters of communities associated with certain interests? If a user participates in many social clusters, how should a search query be resolved with respect to the different clusters?
7. Related Work Tribler social community that facilitates file sharing through a peer-to-peer (p2p) networks a social P2P application where a user can share and consume the contents in Youtube, Bittorrent and liveleak Diki a social bookmarking service that allows users to encrypt and share bookmarks with friends via the XMPP real-time communication protocol. Wisebook integrated with Facebook for peer to peer file exchange but now it is defunct.
8. Social Content Graph A social content graph is a logical graph structure where the labeled nodes represent people and contents, and the labeled edges represent relations between users and items, as well as activities users perform on items or other users.
9. Social Content Graph Person-Person -friendship -subscribed -endorsed Person-Content -tagged -authored -endorsed Content-Person -authored-by -mentions Content Node Content-content -hyperlinks -message threads Person Node Linkage between same type Linkage between different type Figure 1: Social content Graph
11. System Architecture User Interface use to give basic FOAF information, publish content in both peer to peer and web and search the heterogeneous information sources. Middleware Layer for Social Network provides an interface to publish and search information in both Web and P2P Wrapper Layer for handling with other systems helps to integrate the system with the existing Web2.0 applications. Wrapper of online social networks helps to capture the relationship that already exists in Web.
14. Social P2P Layer The Social overlay is the relation between the people in P2P and Web together showing the social tie between the peers. This social overlay manages the relations between different peers, maintaining multiple identity, identifying friends, updating friends list, friend interest areas and other information that are subjected to change. Figure: Social Network Relation of Social Overlay
15. Social P2P Layer P2P Layer Figure: Lookup to find the friends in the network in P2P Figure: Process of connecting to a peer
16. Social P2P Layer Social Information in Social P2P Fig: Showing the process of query being resolved by using different sites
17. Conclusions Exploring the link between the people and the documents and maintaining the connectivity between the people Solve the problem finding of social-content that may reside in different heterogeneous sites. Combining P2P and traditional web application we tried to get the maximum benefits to retrieve the social content. P2P Infrastructures used to share the FOAF information whereas the Web is used to search for the information.
18. Future Work P2P to make server less system for Message Boards and Forums with capabilities of Web 2.0 applications. Our future research will be directed towards making social networks in P2P infrastructure.
Diki [31] is a social bookmarking service that allows users to encrypt and share bookmarks with friends via the XMPP [32] real-time communication protocol.Wisebook which is
Open socialAPIs
We cluster people based on interest. Clustering people helps to categorize people of similar interest together. When a user search for the query the system finds the clusters of people (in this case users friends/people user knows) having the same interest as the query and then search to those sites for the result.