Improved Purpose-Based Concurrency Control to Reduce the Electric Energy Consumption of a Server Cluster

In order to provide context- and energy-aware information systems, the EEPO-VM (Energy-Efficient Purpose Ordering with Virtual Machines) scheduler is newly proposed to serialize transactions based on the significancy of purposes assigned to the transactions in addition to reducing the total processing electric energy consumed by a server cluster which supports multiple virtual machines. We show a transaction assigned a more significant purpose can be preferentially performed than the other transactions and the total processing electric energy (TPE) consumption of a server cluster which supports multiple virtual machines can be reduced in the EEPO-VM scheduler through a simulation.