Distributed computing systems

TL;DR: In order to secure reliable operation of the monitoring infrastructure, two distributed service applications named Perfmon and Servmon are developed which monitors operational state of components in a distributed hardware & software system and reacts on detected problems.

Abstract: As part of our network monitoring activities, we deployed extensive monitoring systems in our network CESNET2 and in the European academic network GN2 [1]. These systems use active, passive and infrastructure monitoring. They combine multiple tools and monitoring architectures. The infrastructure consists of many hardware and software components. These include PC servers, monitoring cards, operating systems, drivers, libraries, middleware and application software. In such a complex distributed infrastructure, component failures are inevitable. It can be due to hardware failure, resource exhaustion, unexpected network changes, or even software bugs, which unfortunately happen. In order to secure reliable operation of our monitoring infrastructure, we have developed two distributed service applications named Perfmon and Servmon. Perfmon monitors operational state of components in a distributed hardware & software system and reacts on detected problems. Servmon monitors resource consumption. In this report we describe architecture of these two applications in more detail.