More and more modern computer applications from business decision support to scientific data analysis utilize visualization techniques to support exploratory activities for large datasets. Various tools have been proposed in the past decade that help users better interpret data using such display techniques. However, such exploratory visualization tools do not scale well when applied to huge datasets. Various features provided by database management systems must be applied to such applications to scale them for huge datasets. To improve the performance of such visualization systems, caching the data at client side is necessary. We exploit semantic caching for the advantages if offers over the traditional caching systems. To further improve the performance, we propose to exploit characteristics of visualization environments to prefetch the data for the visualization tools. We have incorporated these features into XmdvTool, a freeware visual tool for multivariate exploration. We also compare an array of different prefetching strategies to determine their relative effectiveness for both synthetic user traces and real users of our system. Our results show that significant improvement can be achieved for visualization applications by caching and prefetching the data on the client-side.
Doshi, Punit R.
, Rundensteiner, Elke A.
, Ward, Matthew O.
, Stroe, Daniel
(2002). Prefetching For Visual Data Exploration. .
Retrieved from: http://digitalcommons.wpi.edu/computerscience-pubs/117