A continuous auction model for stochastic grid resource pricing and allocation

Grid Computing is a parallel and distributed system that facilitates the virtualization of distributed computing and enables coordinating and sharing of computing and storage resources. We develop a market-based continuous auction model that is able to allocate grid resources and discover separate prices for the different computing resources in the event of stochastic demand. We also model heterogeneity in risk aversion among jobs as allocated resources may not be enough to complete a job and the amount of risk tolerance for such an event would most likely differ from one job to the next. We formulate the problem as a continuous auction, which is operated as a set of multiple and simultaneous ascending price auctions aimed at discovering uniform unit prices for the resources in each time interval. We model jobs with deadlines and also the scenario where such deadlines can overrun the clearing horizon of a set of simultaneous auctions. We provide methods for validating the results of our model.