For Full-Text PDF, please login, if you are a member of IEICE,|
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
Performance Analysis of an Adaptive Query Processing Strategy for Mobile Databases
Hajime SHIBATA Masahiko TSUKAMOTO Shojiro NISHIO
IEICE TRANSACTIONS on Communications
Publication Date: 1997/08/25
Print ISSN: 0916-8516
Type of Manuscript: Special Section PAPER (Special Issue on Mobile Computing)
mobile computing environment, query processing, Markov chain, performance analysis,
Full Text: PDF(500.9KB)>>
Many network protocols for routing messages have been proposed for mobile computing environments. In this paper, we consider the query processing strategy which operates over these network protocols. To begin with, we introduce five fundamental location update methods based on ideas extracted from the representative network protocols. They are the single broadcast notification (SBN), the double broadcast notification (WBN), the single default notification (SDN), the double default notification (WDN), and the no notification (NN). As a network protocol, each method is strong in performance in some system enrivonment, but weak in others. In practical situations, where various kinds of applications are used for various purposes, however, it is required to use a single method. We therefore propose an adaptive query processing strategy where these five location update methods can be dynamically selected. Moreover, we analyze the performance of this adaptive query processing strategy via the Markov chain. We also use the statistical approach to estimate the traffic of individual hosts. Finally, we show the efficiency of our proposed strategy over a wide area of system environments.