Section: Research Program

Intermediation technologies

DICE focuses on intermediation platforms because of the central role they play in the new economy.

Intermediation platforms connect users to one another, or users to services with a very high accuracy. They rely on innovations both technological and social, which were unthinkable only ten years ago, when Facebook started. They allow communication and interaction between billions of users, gathered in the same digital space, both producers and consumers of data and services. State-of-the-art intermediation platforms include Facebook, Google, Twitter, GitHub, as well as Wikipedia, StackOverflow or Quora. These systems share a common design and their market penetration follows the same pattern. They are built around an initial minimal viable product based on a somehow naive low-tech implementation, which evolves after a few years of improvement to Web giants. Their domination now contributes to standardize the web industry, that means in particular:

• Gatekeeping, a direct relation with users together with services satisfying users' needs;

• Continuous data flows mapped to users' profiles;

• Search engines associating, in a relevant manner, producers, consumers and services.

These common characteristics lead to new software architectural standards, which are shared by all these systems, and used in the peripheral services developed in the ecosystem around their API:

• Authentication systems: openId, OAuth, ...

• Object graphs: opengraph, follower/followee scheme, ...

• DataFlow engines: Twitter storm, Google millwheel, ...

• Databases: noSql, keyValues stores, ...

• Web Browsers: javascript, dart, MEAN (Mongo, Express, Angular, Node),...

These architectural components impact the whole digital world. DICE targets systems that use standard architecture services but preserve some aspects we consider as disruptive ones: data concentration, data symmetry and computational subsidiarity. Our current research activity includes the following directions:

• Peer-to-peer design for preserving users' primary data;

• Third parties based organic systems providing subsidiary data computation hosted at peer sites;

• In-Browser applications that impact mobile device and demonstrate instantaneous usability;

• Flow-based computing enabling a stream based exchange of information between peers at runtime.