Cloud-based Healthcare: Towards a SLA Compliant Network Aware Solution for Medical Image Processing
Shane Hallett, Gerard P. Parr, Sally I. McClean, Aaron McConnell, Basim Majeed · 2012
Abstract—Medical image processing in the Cloud can involve moving large data sets and/or applications across the network infrastructure. With the aim of minimizing the total processing time, the optimal placement of image data and processing algorithms on a large scale, distributed Cloud infrastructure is a challenging task. This work presents a genetic algorithm-based approach for data and application (virtual machine) placement using hypervisor and network metrics to avoid service level agreement violations. The solution involves placing medical image data and associated processing algorithms at optimized processing and compute nodes located within the Cloud. The results of initial experiments show that a genetic algorithm-based placement approach can increase Cloud-based application performance. Keywords-cloud computing; virtual machine placement; genetic algorithm; network awareness. I.