Key points are not available for this paper at this time.
The smart grid offers many benefits due to the bidirectional communication between the users and the utility company, which makes it possible to perform a fine-grain consumption metering. This can be used for demand response purposes with the generation and delivery of electricity in real time. It is essential to rapidly anticipate high peaks of demand or potential attacks, so as to avoid power outages and denial of service, while effectively supplying consumption areas. In this paper, we propose a novel architecture where cloud computing resources are leveraged (and tested in practice) to enable, on the one hand, the consumption prediction through time-series forecasting, as well as load balancing to uniformly distribute the demand over a set of available generators. On the other hand, it also allows the detection of connectivity losses and intrusions within the control network by using controllability concepts.
López et al. (2018) studied this question.