Nowadays, a large number of collaborative editing applications have been developed. Some of them are deployed on the cloud such as Google Drive and Microsoft Office at SkyDrive. Massively used editing systems make use of operational transformation (OT), a traditional replication mechanism for concurrent document editing. Such algorithms do not scale well in peer-to-peer environments with dynamic groups. Recently, Commutative Replicated Data Types (CRDTs) were introduced as a new class of replication mechanisms whose concurrent operations are designed to be natively commutative. They ensure consistency of highly dynamic contents on peer-to-peer networks. Through this paper, we propose an architecture to take advantage of both approaches -- OT and CRDT -- and to improve the performance of collaborative editing applications. We merge both algorithms on the proposed architecture and we study their suitability.