Modeling flexible Monitoring Systems with a Role-Based Control Loop.

Monitoring information technology systems during operation is one of the few methods available to help administrators keep track of the monitored system’s state, predict and identify errors, and assist in system repair and error avoidance. Monitoring solutions to date, however, still suffer from various trade-offs as current implementations impose architectural restrictions on monitored systems, which lead to reduced flexibility in deployment and operation. Most monitoring software is specialized to function with fixed specific data formats, protocols, and data acquisition mechanisms, further reducing their flexibility. The role-based approach for modeling and implementing software promises an intuitive way of increasing the flexibility of information technology systems’ modeling and implementation. Paired with technology on control loops from the domain of self-adaptive systems, we can create a reusable framework made from static role-playing building blocks that allow overcoming those limitations. In this paper, a novel concept and model for a flexible monitoring solution on this basis is presented, which provides functioning in most application domains while minimizing architectural and functional constraints on the monitored system. We compare the flexibility of our concept with a collection of 15 monitoring systems based on 11 criteria. We further illustrate the concept by presenting and discussing a uniform role-based model and, finally, discussing one possible implementation method.