A quadtree replication method to optimize web and mobile applications dealing with large sets of objects on a map
Patrick Munroe, Jonathan Milot, Éric Beaudry, Naouel Moha, Francois R Grondin · 2017
This paper presents an original caching method called Minimal Quadtree Replication (MQTR) for Web and Mobile Applications (WMA) to query and display large sets of objects on a map. This method pre-builds a quadtree at the server side that is incrementally replicated at the client side. Our method tends to avoid useless queries and to minimize the overlaps of content when similar queries to remote services are done. This caching method offers benefits that increase the quality of service (QoS) provided by the WMA using it, when measured in terms of CPU time, bandwidth, and response time. We compared our MQTR approach with common approaches that are currently used by some popular web sites. Results show that MQTR significantly outperforms the other approaches. The best variant of our method reduces in average the number of requests by a factor of 5.5, transfers 3.9 times less data and requires 2.4 times less CPU time at the server side.