Intelligent cache management for mobile data warehouse systems.(Research Note)

ABSTRACT

This research proposes an intelligent cache mechanism for a data warehouse system in a mobile environment. Because mobile devices can often be disconnected from the host server and due to the low bandwidth of wireless networks, it is more efficient to store query results from a mobile device in the cache. For more personal use of mobile devices, we use a data mining technique to determine the pattern from a record of previous queries. Then the data, which will be retrieved by the user, are prefetched and stored in the cache, thus, improving the query efficiency. We demonstrate the feasibility of the proposed approach with experiments using simulation. Comparison of our approach with a standard approach indicates that there is a significant advantage to using mobile data warehouse systems.

Keywords: data mining; data warehouse; mobile technology; wireless technology

INTRODUCTION

Given the current keenly competitive business environment, a Business Intelligence (BI) system should provide data access for managers, analyzers, and decision makers with respect to speed and lower cost. BI provides a set of technologies and products to supply users with the information to answer business questions and make tactical and strategic business decisions (White, 2000). BI systems include Data Warehouse, On-Line Analytical Processing (OLAP), data mining, and so forth. Significant research has been carried out in the development of wireless data warehousing that can serve as the base for a BI system (Segev, Shao & Zhao, 2003). As the power and storage evolution of mobile device and environment, the mobile data warehouse environment will become practical very soon. However, a recent survey indicated that the commercial use of such mobile BI systems is still limited since 2001. Currently, only about 10% of suppliers of data warehouses noted that they provide or developed their wireless environment application solution (Teige, 2001). Thus, there is still neglect for development within the scope of wireless BI systems.

In this domain, possible issues have been mentioned in Siau and Shen (2003), including killer applications, usability, user interfaces, mobile access to databases, and agent technologies. For improving mobile access to databases, it is stated that part of the database systems must be able to reside on the mobile devices. A mobile database system may only manage a portion of a large central database, pulling in additional data on demand. With such respect, query processing for mobile database systems must be significantly improved. Our research follows this issue and discusses means of improving the query performance in such a mobile environment.

However, querying at a stationary server by a mobile client is expensive due to the limited bandwidth of a wireless channel and the instability of the wireless network (Jiao & Hurson, 2004; Lee, Leong & Si, 2002). This research is concerned with the mobile data warehouse, focusing on characteristics of connection unreliability, limited bandwidth, and limited storage space in the mobile device. Traditionally, information extracted from different resources can become integrated information stored in a data warehouse (Hsieh & Lin, 2003). Mobile users can send queries to the data warehouse through a wireless network, and the query result is presented to users with multidimensional schema.