The implementation of digital library applications is not only limited in a client-server architecture. It can also be implemented in a peer-to-peer architecture. The implementation into peer-to-peer architecture is believed to be able to lessen the workload access which is usually faced by a server previously in a client-server architecture.However this utilization does not run seamlessly whitout barrier. Several crucisl problems srise such as low document searching efficiency and low peer credibility for data exchange. In this research, we propose a method or protocol called jaris to increase document searching efficiency by using cache query and cache cluster. Jaris also can identify the level of peer quality which joins the network. In this experiment, jaris performence is compared to other method called the random voting method. The experiment shows that jaris performence is better than random voting. Keywords: Digital Library, Peer-to-peer, Jaris