Prioritized sequencing for efficient query on broadcast geographical information in mobile-computing

  • Authors:

  • Jianting Zhang;Le Gruenwald

  • Affiliations:

  • The University of Oklahoma, Norman, OK;The University of Oklahoma, Norman, OK

  • Venue:
  • Proceedings of the 10th ACM international symposium on Advances in geographic information systems
  • Year:
  • 2002

Quantified Score

Hi-index 0.00

Visualization

Abstract

Broadcasting is an effective and efficient way to disseminate public geographical information to massive users. In this paper, we use R-tree to index broadcast geographical data. Based on our observation that geographical data and their accesses are often concentrated (clustered) in real applications, we propose a novel prioritized sequencing method by considering access frequencies based on application semantics to reduce average access time (latency). The key idea is to put "hot" data items ahead of "cold" ones in the broadcast sequence while still maintain R-Tree indexing structure.We evaluate our method using MapInfo Census 2000 data sample with 576 service locations and 111 of them are hot data items. The average reduction percentages of access time are reduced by 60% for both point queries and range queries based on 10 rounds experiments, each consists 10 hot data items and 5 cold items that are randomly selected. We conclude that our simple method is very effective in improving query efficiency in geographical information broadcast systems.We also proved the linear relationship between wait time and the Aggregate Data Affinity (ADA) (Lee, 2000), a parameter measuring the "eagerness" for data experienced by a client during a query process in broadcast wireless mobile computing environment. Thus our conclusion also holds when using ADA for evaluating broadcast system performance.