Access pattern analysis for a directional cache scheduling scheme on vehicular telematics networks

Abstract

This paper analyzes the data access pattern to the gateway cache for the better design of direction-based query scheduling scheme, based on the real-life trajectory data. After defining an advanced route information service in which each vehicle requests the data items associated with all links along the route in the vehicular telematics network, experiments are conducted to investigate the queue length distribution, hop count distribution of target links, and hourly pattern change in the direction of trip destinations. The observation on such factors discovers the effect of the number and length of trips to the queue length and response time, highlights the links of hops from 25 to 100 the query scheduling policy must concentrate on, and finally, obtains the spark shape in the destination pattern curve. This result makes it possible to revise the direction-based scheduling policy to maximize the cache hit ratio as well as overcome the bandwidth limitation between the gateway and information servers.