Skip to Main Content
In order to realize green ecosocieties, the total electric power consumption of computers and networks is required to be reduced. In applications on distributed systems, clients issue service requests to servers and then servers send replies to clients. Here, we discuss how much electric power a server consumes since the power consumption of a client is neglectable compared with a server. We classify applications into transaction- and communication-based ones. A server mainly consumes CPU resources to perform the transaction-based applications. On the other hand, a server consumes communication resources to transmit a large volume of data to a client in communication-based applications. In our previous studies, the power consumption laxity-based and extended power consumption-based algorithms are proposed to select one of servers so that the total power consumption of servers is reduced for transaction- and communication-based applications, respectively. However, most applications are mixed types, i.e., composed of both the transaction and communication processing modules. Hence, we consider the mixed types of applications in this paper. First, we integrate the power consumption models of transaction- and communication-based applications into a modified simple power consumption (MSPC) model of a server. Based on the MSPC model, we propose an algorithm to select one of servers for mixed types of applications so that the total power consumption of servers can be reduced. We show the power consumption of servers can be reduced in the algorithm through the simulation.