Impact of mismatch probability on service selection in context aware service discovery

摘要

Service discovery can be enhanced by using context information by calculating a score value indicating how well the service fits to the current context of the user. In some cases context information is accessed from remote nodes, which can be done using one of the three different strategies; a reactive, proactive periodic and proactive event driven update. When the information is being accessed and used to calculate the score value, it may be outdated due to the delay in the update of the information caused by the network. Such a situation leads to an erroneous score calculation which in turns, leads to the user selecting a service based on a wrong basis. In this paper we investigate the impact of information that has changed when being accessed on the selection of services, and find that one can reduce the probability of making wrong service selections at the dispense of the accuracy of the score value.