An extension of CloudSim toolkits for GPGPU-based cloud computing simulation

The recent technology on GPGPU (General Purpose Graphic Processor Units) has enabled Cloud computing users to run high-performance applications not only on CPUs but also on GPUs. Thus, Cloud computing technology has also been developed based on GPGPU resource, which results in much recent research about GPGPU-based Cloud resource allocation. Since little work has been done on simulation platform for this GPGPU-based Cloud computing, it is not easy to show the performance evaluation of developed algorithms for GPGPU-based Cloud computing. In this paper, we provide a Cloud simulation framework by extending the existing CloudSim toolkits which is commonly used simulation software packages for Cloud computing. We add several components to simulate GPU resource virtualization. The proposed simulation framework will be useful to develop new GPU resource allocation schemes in Cloud computing.