Design decision alternatives for agent-based monitoring of distributed server environments

Abstract

In every IT department it is crucial for the management to be able to have a good overview of their service infrastructure to be able to charge for their service operation cost. With distributed services that do not always run on dedicated servers many services can be located on the same server and therefore it is needed to split the operation cost between the services. To be able to split this cost in a correct way a lot of information needs to be available, such as the resource usage and audit information for the systems so the cost for the system can be divided between the users. In this thesis are two design alternatives for agent frameworks developed and evaluated to see if it possible to create more accurate cost model by either of them using agents to monitor the service located on servers.