Japanese
◆ Annual Review 2002

Performance Evaluation Laboratory


Kouhei Otsuyama
Assistant Professor

Hiroshi Toyoizumi
Assistant Professor

The performance evaluation laboratory is part of the Department of Software, the University of Aizu. The laboratory mission in the broad sense is to contribute to:

  • Theoretical Aspects of Performance Evaluation of Computer andCommuni- cations Systems
  • Tools andTechniques for thePerformance Evaluation ofComputerandCom- munications Systems
  • Applications to Computer and Communications Systems : Evaluations and Performance

During fiscal year 2002 the laboratory was composed of two faculty members. Current research interests include:

  • Performance evalution of network system: Using mathematical analysis we study how to evaluate network systems, sucha as WWW servers, streaming services and so on.
  • Performace evaluation of computer security system: Evaluating security system in network is becoming very important. We use mathematical model, representing such systems as public key infrustructure, computer virus and denial of service attack, to evalute the security systems on network.
  • Performace evaluation of bioinformatics: Bioinformatics deals with huge amount of data. The eAEciency of the data processing is one of the crucial points in bioinformatics. By evaluating current methods, we can improve the efficiency of the tools used in bioinformatics.
Referred Journal Papers
[toyo-01:2002]H. Toyoizumi and A. Kara. Predators: good will codes combat against computer viruses. In Christian F. Hempelmann and Editors Victor Raskin, editors, ACM SIGSAC New Security Paradigms Workshop 2002, pages 11-17, New York, Sept. 2002. ACM, ACM press.
We present a mathematical analysis of a new approach to fight against computer viruses through the use of their predators. Predators are good will mobile codes which, like viruses, travel over computer networks, and replicate and multipy themselves. The only difference is that predators are specifically designed to eliminate the viruses. We model the interaction between predators and viruses by the Lotka-Volterra equations, which are widely used in mathematical biology. Using this model, we derive a method to constrain the number of predators to be as few as possible, while maintaining their power to eliminate viruses.
Unrefereed Papers
[toyo-02:2002]H. Toyoizumi. Performance Evaluation of Group Security. In The Madrid Conference on Queueing Theory, pages 35-35, Madrid, July 2002. INFORMS.
Grants
[toyo-03:2002]H. Toyoizumi. Ministry of Education Scientific Research Fund, 2000-2002.
Ph.D and Other Thesis
[toyo-04:2002]Kentaro Arai. Graduation Thesis: Evaluation of Anti-Virus Application with KLEZ, University of Aizu, 2002.
[toyo-05:2002]Dai Azami. Graduation Thesis: Detection of the distributed denial of service attack, University of Aizu, 2002.
[toyo-06:2002]Kenshiyou Kazama. Graduation Thesis: Gene Finding with Hidden Markov Models, University of Aizu, 2002.
[toyo-07:2002]Yuka Yamada. Graduation Thesis: Performance Evaluation of Server-Clinet system with cryptograph, University of Aizu, 2002.
[toyo-08:2002]Yuuzou Kobayashi. Graduation Thesis: Modeling of Computer viruses, University of Aizu, 2002.
[toyo-09:2002]Mizutani Norimiti. Master Thesis: Computer System Performance Observation through the Computer Network, University of Aizu, 2002.
[toyo-10:2002]Matsuyoshi Takaya. Master Thesis: Performance analysis of key server of group secruity, University of Aizu, 2002.