One of approaches to reducing energy consumption in data centers is to power down a group of servers according to the utilization rate, or the number of jobs in system. In this thesis, we consider a power management scheme for distributed parallel processing over clusters of servers. For this scheme, a part of servers in each cluster are turned on/off according to the state of a background process, or the number of jobs in system. In the former case, we model the system as a multi-server queue in which the service time of a job depends on the state of a background process at the beginning of the job service. In the latter case, on the other hand, we consider a multi-server queueing system where the service time of a job depends on the number of jobs in system. For both systems, we analyze the distribution of the number of jobs in system, deriving the mean job-response time and mean amount of energy consumption. In the latter case, we further consider the product of those two measures. In numerical examples, we investigate how performance measures are affected by the background process or thresholds which manage energy saving level.