Mobile Social P2P Networks
The social network systems like Facebook, MySpace and LinkedIn have become popular means to share personal information. People can for example live-broadcast themselves via justin.tv or share photos through Flickr. Their architecture stores the data on centralized servers consuming network bandwidth and storage space. For example YouTube stores lots of videos on the centralized servers but this architecture comes with high costs. According to Forbes, in 2006 YouTube was estimated to transfer 200 terabytes of data each day with costs of one million dollars per month. Some P2P technology developers estimate that savings of 95% in hosting costs can be achieved with a P2P architecture.
Another viewpoint is the transition that is taking place in the development of mobile devices. For a long time mobile devices have been closed. Calendar data, notes, photos etc. have been kept private in the mobile device and whenever needed transferred using short-range connections like Bluetooth to desktop computers for storage or sent via short messaging system (SMS) to other mobile users or centralized servers. However, during 2007 NOKIA has ported Raccoon Mobile Apache Web Server with Personal Apache MySQL PHP/Python (PAMP) stack to Symbian phones thus opening up the mobile device for external access from the Internet. Now it is possible to develop web services on a mobile device or provide access to different kind of content using the de facto tools of traditional web programming. This creates a huge and a rapid wave of mobile web application development, because programmers can now keep their existing development tools.
The Mobile Social Peer-to-Peer Networks research project aims to overcome the architectural difficulties of the current social network systems by distributing the content of the centralized servers to the mobile devices. In the architecture, the users of the mobile devices publish their content (files, folders, calendar etc.) locally in the devices via mobile web server and define the access rules for the content using the motto “I only display what I want to who I want”. This can be realized for example using indexing architecture similar to Google, but now locally in a mobile device with access-controlled search engine. In other words, we are developing Mobile Google.
The main benefits of the mobile social P2P networks in addition to low costs are real-time searching capabilities and a possibility to locate rare, but socially relevant data. An example of such data is information on who of my acquaintances have watched a certain movie or who of my colleagues have written about a certain topic. In addition to social relevance, other dimensions like contextual relevance can be used as a measure. An example of this is data located in nearby mobile devices. In general, the Mobile Social P2P networks architecture shortens the information flow chain significantly and complements centralized search engines. Now it takes long time when a captured information (e.g., photo) can be found from centralized search engines. With Mobile Social P2P networks architecture the users who search data reveal which content becomes popular and this content can then be added to centralized search engines making it better searchable.
The project divides into the following parts:
1. Developing an open source mobile social P2P networks prototype of a peer with PHP programming language allowing controlled publishing of content for other mobile devices and web clients in the Internet.
2. Developing a search algorithm which allows finding content in a social network consisting of the Raccoon mobile web server nodes and mobile device address book contacts.
3. Developing a result ranking algorithm to sort the results found in a mobile social P2P network.
4. Studying these algorithms using measured social network topologies and content distributions.
The project is partially funded by ICTSHOK Future Internet.