Automation of System Monitoring Based on Fuzzy Logic or Rules; Comparison of Two Designed Approaches with Regard to Computational Infrastructures WlodzimierzFunika 1,2 , Filip Szura 1,2 , and Jacek Kitowski 1,2 1 AGH University of Science and Technology, ACC Cyfronet AGH, ul. Nawojki 11, 30-950 Krak´ ow, Poland 2 AGH University of Science and Technology, Faculty of Electrical Engineering, Automatics, Computer Science and Electronics, Department of Computer Science, al. Mickiewicza 30, 30-059 Krak´ ow, Poland {funika,szura,kito}@agh.edu.pl Abstract. This paper is focused on monitoring automation of distri- buted systems. In the presented research, AI-based approaches to dis- tributed monitoring related to large distributed systems such as grids, were explored. In both presented concepts knowledge is used to make de- cisions regarding management actions, using rules and fuzzy logic. The first concept is an agent-less rule-based solution, implemented in a high- level monitoring system called Saude-Net. It allows to define actions for monitored resources, using a kind of expert system. The second solu- tion, which exploits agents and fuzzy logic, is realized in a system called SAMM Compliant Agent. Both presented systems are capable of react- ing to observed failures and of modifying their knowledge to better fit possible problems with resources. We also present a short comparison of the two concepts, and an analysis of their usage. Keywords: system monitoring, automation, rules, artificial intelligence, fuzzy logic, fuzzy sets, error reporting. 1 Introduction Since the PL-Grid infrastructure [1] is meant to be a country-wide compute- and data-intensive platform, facilitating its administrator’s operations, e.g. due to the size of infrastructure, is one of the key issues. An installation of this size can be endangered by malfunctioning of its resources, which needs detection of their failures and fast reactions (responses) to them. To accomplish this task it is necessary to provide support for monitoring and reacting to observed errors. To do this the system administrator is usually assumed to check any raised alerts and respond on them. It has to be done con- tinuously. This paper presents two approaches aimed to help administrators in their daily work. These solutions use knowledge described by rules or fuzzy sets. M. Bubak, T. Szepieniec, and K. Wiatr (Eds.): PL-Grid 2011, LNCS 7136, pp. 142–156, 2012. c  Springer-Verlag Berlin Heidelberg 2012