Leveraging the Grid for the Autonomic Management of Complex Infrastructures

Autonomic Management of complex IT infrastructures requires multiple different types of analysis to be performed on large data sets in order to promptly detect anomalies and attacks and take appropriate actions. The requirements in terms of computing power can be so high that new solutions must be devised in order to reduce reaction times. This paper describes how Grid technologies can be leveraged in the Autonomic Management of complex IT infrastructures, such as complex networks with tens of border firewalls. In particular, we show how an Autonomic Manager can benefit from the large computing power and storage space made available by the Grid. We designed the architecture of an Autonomic Manager which makes use of Grid resources to collect and store data and to perform analysis. We then implemented the core part of the system (LoGrid), which collects data in CBE standard format, processes them using Grid resources and feeds back the results in CBE format to the Autonomic Manager. We tested our implementation in a reference scenario and found that the use of the Grid for the Autonomic Management of complex IT infrastructure is feasible and convenient.