Adaptive Web-based monitoring and management for distributed systems

Today's business environments in which increasing number of services are being distributed on multiple computers impose new challenges on administration systems. There is a growing need for comprehensive and scalable solutions that would facilitate efficient management and monitoring operations on such distributed systems. Recent evolution in advanced Web technologies and rapid expansion of the Internet has facilitated means for development of efficient, low-cost and widely accessible administration tools. In this paper, we build upon these facts to design and prototype a scalable and efficient monitoring and management framework for real-time administration of distributed computer systems. Presented framework is based on Push technology which uses publish/subscribe mechanism for efficient real-time delivery of monitored data to web-based administration consoles. Furthermore, configurable nature of the prototype allows easy integration of additional monitoring probes at any moment without need for any software changes in the system. Our experiences with experimental use on telecom cluster running highly available AAA service along with relevant indicators which could affect service performance are pointed out at the end of the paper.