Modern on-line collaborative learning environments need to be continuously adapted, adjusted, and personalized to each specific target learning group. Moreover, these environments are to enable and scale the involvement of an increasing large number of single/group participants who can geographically be widely distributed, and who need transparently share a huge variety of both software and hardware distributed resources. As a result, collaborative learning applications need to be designed in a way that overcome important non-functional requirements arisen in distributed contexts, such as scalability, availability, interoperability, and integration of different, heterogeneous, and legacy collaborative learning systems. In this paper, an innovative distributed-based approach is presented for increasing the overall performance of collaborative learning systems that contributes to the effectiveness of the collaborative activities, such as online discussions. The experimental results show an outstanding effect on the learning processes and outcomes by enhancing and improving the learning experience a great deal.