Current system and network monitoring tools for ICT systems often only provide an overview over the current status of single system components (CPUs, memory load, available bandwidth, etc.). Just some of the available tools provide administrators with a historical overview or the development of trends and even less provide insight into possible future developments of the monitored systems. Furthermore, most of the monitoring tools single out system components and do not connect all components to a common picture, which is crucial for the analysis of complex distributed systems. Furthermore, the increasing use of cloud infrastructures – with a focus on Infrastructure-as-a-Service offerings – has led to a surge in building more dynamic, fast growing and continuously changing systems. In this environment today’s administrators are confronted more than ever with the requirement to get and keep an overview over these complex distributed infrastructures with the limited
state-of-the-art monitoring tools.
The PIPES-VS-DAMS project aims at building and providing new approaches to analyse, monitor and simulate complex distributed systems that can be found on multiple scales (ranging from web platform providers, via cloud infrastructure providers to Internet service providers). This should be achieved by applying techniques from visual analytics, data mining and dynamic network analysis.
PIPES-VS-DAMS aims at applying approaches that are already used to analyse social networks and their communication graphs to identify central communication hubs and opinion leaders in distributed computing infrastructures and communication graphs in computer systems to automatically identify central and crucial system components in complex computer infrastructures. This can be applied to identify the flow of visitors on websites, to gain insight into the communication between system components of a cloud infrastructure, or to visualize the data flow in the network infrastructure of an Internet Service Provider. In all these areas, the analysis and identification of core components of distributed infrastructures and how the single components influence each other is of high importance to provide high availability and also to optimize and further develop these complex systems.