Section: Software

Treedoc

Participants : Marc Shapiro [correspondent] , Marek Zawirski, Nuno Preguiça.

A Commutative Replicated Data Type (CRDT) is one where all concurrent operations commute. The replicas of a CRDT converge automatically, without complex concurrency control. We designed and developed a novel CRDT design for cooperative text editing, called Treedoc. It is designed over a dense identifier space based on a binary trees. Treedoc also includes an innovative garbage collection algorithm based on tree rebalancing. In the best case, Treedoc incurs no overhead with respect to a linear text buffer. The implementation has been validated with performance measurements, based on real traces of social text editing in Wikipedia and SVN.

Work in 2010 has focused on studying large-scale garbage collection for Treedoc, and design improvements. Future work includes engineering a large-scale collaborative Wiki, and studying CRDTs more generally.

TreeDoc is supported by the Prose, Streams and ConcoRDanT ANR projects (Sections  8.1.7 , 8.1.6 and  8.1.4 respectively) and by a Google European Doctoral Fellowship (Section  8.2.2.1 ).

The code is freely available on http://gforge.inria.fr/ under a BSD license.