Evaluation of service management algorithms in a distributed web search system

摘要

The World Wide Web interconnected through the internet today offers numerous specialist topic-oriented or regional search engines and systems in a largely federated heterogeneous environment. Old ones continue to exist and new ones appear in spite of the tremendous progress achieved by their generic Web-wide rival competitors, because they produce better results in their areas of specialisation. However, finding and choosing the best specialised search engines or systems for a particular information need is difficult. This is made even more complicated by the fact that these engines and systems would want to carve out a niche market that generates maximum revenue for themselves. The ADSA (Adaptive Distributed Search and Advertising) Web research project has investigated the problem at some depth and had put forward a search architecture which allows many search engines to be independently owned and controlled, offering advantages over existing centralised architectures. One aspect of the architecture has been to evaluate the service management algorithms that were designed to support competing autonomous systems in a cooperative marketplace. Here we present ADSA economic model and the service management strategies that can lead to maximum revenue generation, by making informative and intelligent decisions on search price adjustments of key quantitative parameters, as well as the results of evaluation experiments and briefly discuss the need for standardised interfaces which are required if this concept is to ease development and implementation of such a marketplace in a large scale.