Distributed and adaptive discovery using preference

Distributed applications produce the need of locating distributed network objects (i.e., data, application or users) that meet a certain search criteria in a large-scale and dynamic network. We propose a discovery mechanism that is fully distributed (without any centralized entity) and adaptive to dynamic network environments. The proposed discovery mechanism allows query originators to return feedbacks that describe the degree of the preference for discovery results. The preference of query originators is utilized to guide subsequent queries in a distributed manner, enhancing efficiency of discovery and improving the quality of discovery results. Furthermore an effective use of the preference of query originators enables the proposed mechanism to adapt to dynamic environment changes. We investigate the efficiency and adaptability of the proposed discovery mechanisms through simulations.