Autonomous Information Allocation to Achieve Real Time Property in Demand Oriented Information Service System

カテゴリ: 部門大会

論文No: GS20-4

グループ名: 【C】平成14年電気学会電子・情報・システム部門大会講演論文集

発行日: 2002/09/02

タイトル(英語): Autonomous Information Allocation to Achieve Real Time Property in Demand Oriented Information Service System

著者名: Yi Zhou(Tokyo Institute of Technology),Farooq Ahmad(Tokyo Institute of Technology),Helene Arfaoui(Tokyo Institute of Technology),Kinji Mori(Tokyo Institute of Technology)

著者名(英語): Yi Zhou(Tokyo Institute of Technology),Farooq Ahmad(Tokyo Institute of Technology),Helene Arfaoui(Tokyo Institute of Technology),Kinji Mori(Tokyo Institute of Technology)

キーワード: high-assurance|information service|autonomy|mobile agent|load balancing

要約(日本語): Current information service systems are characterized by continuous changes of services from business companies promoting their special offers on-line, and unpredictable users' demand. In such system, service providers (SPs) require timely update to inform about their new products, and at the same time, users require a timely access to the information services.

This fosters an urgent need for the design of a high-assurance information service system, that is, a system can meet both providers' and users' heterogeneous requirements. Faded Information Field (FIF), sustained by push/pull mobile agents, is such an architecture that brings high-assurance through a recursive demand-oriented provision of the most popular information closer to the users to make a tradeoff between the cost of service allocation and service access. However, when users' demand changes, there is a need for dynamically restructuring the information service system to restore the balance between users and SPs. This paper proposes an autonomous information allocation technique in the FIF environment. In this technique, nodes and push mobile agents locally coordinate to estimate the workload congestion of upper nodes and determine the new allocated information volume matching the current users' needs. The effectiveness in reducing the average response time of the users has been proved through simulation.

PDFファイルサイズ: 8,202 Kバイト