Hybrid Context Matchmaking for Fast Context-Aware Computing

The context-aware system provides customized services to the users according to the current context. The traditional context reasoning process always produces the same result for the given context, called fixed matchmaking. With this, though, the context processing time substantially increases when the number of same events becomes large. We propose to solve this problem by just-in-time (JIT) matchmaking, which finds the most appropriate agent at the moment by querying the directory facilitator. Its performance in the relatively small scale system is, however, slightly lower than fixed matchmaking due to query overhead. In this paper, thus, we propose a scheme dynamically choosing either fixed or JIT matchmaking based on the frequency of the use of any agent, called hybrid matchmaking. An experiment identifies that the proposed hybrid matchmaking scheme outperforms the other two schemes in terms of context processing time when the number of agents is over about 430. Moreover, it is almost constant regardless of the number of agents, which implies high scalability.