A foundation for extensible and decentralized social networks

The world ascended to a new era of communications in the last decade due to the release of centralized social networks such as Facebook and Twitter. As the popularity of these platforms soared, many changes were made, new features were introduced and websites were born. Among these new platforms, a new trend emerged towards decentralized and open social networks (DOSNs). Instead of a single administrative authority collecting and storing social user data (friends, posts, comments, likes, etc.), these social networks operate as a federation of independent websites. A federation is a set of platforms where users connected to one platform can follow and interact with users from other networks while disallowing the concentration of data in a single domain. These platforms all have in common that they are open and target interoperability. The problem with all DOSNs participating in the federation is that they are built as rigid monolithic applications using a single relational database. This design leads to poor flexibility of purpose: it is difficult to adapt these DOSNs, for instance, to support new forms of social interactions. It is also difficult to isolate in these platforms the components that are critical for security and, in particular, the ones that manipulate sensitive user data. A bug or exploit in any part of the monolith can allow a malicious attacker to read the entire database, and bugs are difficult to detect due to the size of the software. The objective of this project is to investigate the use of recent advances in the engineering of cloud back-ends for building DOSNs. This research will focus on decentralization, personalization, and interoperability through the microservices architecture and event-sourcing pattern.