Processing smart meter data streams in the cloud

The ongoing integration of renewable energy sources is likely to increase the fluctuations in the ratio of produced and consumed power. Several types of Demand Response (DR) programs have been proposed to deal with the increasing volatility of power production and consumption. Many of these, such as Real Time Pricing (RTP), require intensive monitoring of the consumers' power consumption. This is one of the reasons why smart meters are currently being deployed by many utilities. Smart meters offer a two-way communication channel between the consumer and the utility, thus extending the power grid by a complex, large scale communication infrastructure. With the growing deployment of smart meters, power utilities face the problem of processing and storing the incoming data to support latency-sensitive applications such as Real-Time Pricing. In this paper we present a set of requirements for a utility-side IT infrastructure to process incoming smart meter data streams. We propose the use of Infrastructure-as-a-Service clouds and frameworks for parallel stream processing in clouds to address these requirements. Based on the Nephele cloud computing framework we demonstrate the practicality of this approach based on experiments with one million simulated smart meters and a prototypical Real-Time Pricing application deployed in our own private cloud.