원문정보
초록
영어
As the number of Web users and the amount of information produced keep growing rapidly, information systems need to be scaled up constantly. Usually scaling methods of information systems suffer from being chosen by experience, this paper studied on quantitative optimal selection of vertical/horizontal scaling methods. According to investigation and statistics, we assumed the cost growth formula of the key performance parameters reasonably. Then five key parameters of systems were chosen to construct Kiviat graphs and the areas of Kiviat graphs were computed to evaluate system performance. Next, the quantitative function relation between the performance of information system and the cost of system scaling was established, and then three optimal selection algorithms for system scaling were proposed. Finally, some case studies demonstrate that algorithms are feasible and have a great significance for guiding system scaling.
목차
1. Introduction
2. Basic Concept and Extension Requirements
3. Web System Architecture and Cost Model
3.1. Web System Architecture
3.2. Modeling
4. Comprehensive Performance Model
4.1. Area Method for Evaluating Comprehensive Performance
4.2. Calculation of Each Component
4.3. Semantics of Extension by k Times
5. Optimal Selection Algorithms
5.1. Same Extension for Each Component and Less Cost Chosen First (SELC)
5.2. Same Cost for Extension and Better Performance Chosen First (SCBP)
5.3. CPU Preference and Less Cost Chosen First (CPLC)
6. Case Studies
6.1. Actual Parameters
6.2. Applications
7. Conclusion
Acknowledgement
References
키워드
- Web information system
- vertical scaling
- horizontal scaling
- optimal selection