Abstract: Despite the high representativeness of energy costs over the cost of the data center, there are few published studies evaluating the electric consumption by the use of the processor.
This represents up to 90% of the power consumption of a server's components, and IT equipment (e.g., servers, routers) can reach 52% of a data center's power consumption. Equipment leasing in a datacenter linked to the cloud computing services model is done on the pay-as-you-go or pay-as-you-use principle. Choosing the pay-as-you-go model simplifies billing mode by charging for the allocated virtual resources (processor, memory, disk, and network), not by the resources used. Therefore, the efficient use of resources by the cloud client does not imply cost reduction. The pay-as-you-use model is based on the effective consumption of resources, motivating the cloud customer to make their application more efficient for a consequent cost reduction. Efficient use of resources is one of the principles of green computing, and reducing energy costs is interesting for both cloud customers and data center administrators. Preliminary experimental investigations indicate a strong relationship between processor consumption and server electrical consumption.
In this context, this research project aims to present a cost model for green computing in the cloud that considers the consumption of the processor and its impact on the energy consumption for fair billing of virtual machines in data center.