Social Cloud over Plug Networks, Enabling Symmetric Access to Data and Preserving Privacy – SocioPlug

Socioplug will explore how to deploy efficiently disorderly distributed programs on a federation of plugs using probabilistic approach and emergent localities. SocioPlug will allow plugs programmers the ability to perform «community-based« queries i.e. queries that preserve information symmetry. Finally, Socioplug will secure the federation by monitoring divergence evolution of streams on each node and by integrating a usage control model based on data publishers rather than service providers.

A huge work in industry and academia is or is being dedicated to cloud computing for real-time social web applications. We are convinced that cloud computing is only a footstep, the real challenges concern the evolution of those applications towards fully decentralized architectures.
The audience and the impact of this project within our scientific communities and within potential industrial customers are also great indicators of success. Finally, the degree of integration of our final demonstrator based on the Raspberry Pi cluster platform (located at LINA partner) is another measure of success. We also aim to ensure compatibility of our proposals with existing solutions (such as Mozilla model).

In fine, SocioPlug is going to propose an alternative, independent of huge data center provided by few companies. The project will give birth to a new service, which becomes precious as privacy is before a key for society commitment.

SocioPlug is basic research project. Therefore, our first success criteria are communication in top-level international conferences and journals in distributed computing, collaborative work, querying systems, privacy and safety.

Free online services such as Twitter or Google docs have come to play a pivotal role on today’s web, in both our personal and professional lives. These services consume substantial resources from their providers, who recoup their investment by monetizing the data they obtain from their users. Because this business model favors large companies, user data and their exploitation are now concentrated in the hands of few powerful players who drive the industry. Users, who have contributed to the data, have no or little control nor access to data, and must live with a fundamentally unbalanced relationship.

Many grassroots initiatives have alerted users on the threats to their privacy, and on the problematic dependency created by such services. Few users are however aware of the systematic extent to which their data is being exploited. To offer alternative choices, several initiatives such as Freedom box, Beedbox, or Soxyd have been launched to develop decentralized social networks regulated by on-line communities, with the aim to replace traditional hosted on-line services. These approaches have been further helped by the rise of novel low-cost plug computers, such as Raspberry Pi, which offer a cheap and readily available infrastructure to deploy them.

Unfortunately, and in contrast to traditional cloud services, today we do not know how to construct such a service on a federation of plugs. In particular, we lack any clear model to write and deploy applications, seek for pertinent information, and secure access in such a fully decentralized environment. How should we model the security, safety and privacy of such a decentralized social infrastructure? What consistency and programming model should we use? How can we perform efficient queries on such alternative personal cloud infrastructure? How can we valorize users’ data while preserving privacy? How can we bridge the gap between privacy and data exploitation?

SocioPlug aims to provide models for programming, querying and securing federation of plugs while preserving symmetry, fairness and privacy.

Socioplug will explore how to deploy efficiently disorderly distributed programs on a federation of plugs using probabilistic approach and emergent localities. SocioPlug will allow plugs programmers the ability to perform "community-based" queries i.e. queries that preserve information symmetry. Finally, Socioplug will secure the federation by monitoring divergence evolution of streams on each node and by integrating a usage control model based on data publishers rather than service providers.