Smart Metering of Cloud Services

Smart metering, prevalent in the power grids,<bold> </bold>is a<bold> </bold>billing model in which consumers are charged variable rates for a service consumed based on the load conditions of the system providing the service. In this paper we present a similar billing model for cloud services. We develop a model where the tariff is varied based on the load prevailing on the cloud infrastructure. We obtain a mapping between the load condition for a particular time period and the pricing. The pricing for every time period is published and it is the consumers discretion to continue using the service or to suspend usage. Based on historical data, the load on the cloud infrastructure is predicted using the auto-regressive integrated moving average (ARIMA) statistical model. Monitoring the cloud infrastructure is an essential component of any form of cloud metering and we rely on tools available off-the-shelf for the purpose. The pricing information obtained is used for billing the consumers. The bill amount is a function of pricing for a time interval and the corresponding utilization of compute resource. The calculated bill is produced to the consumer.