Abstract: Methods, systems, and computer storage media are disclosed for determining electric energy flow predictions for electric systems including photovoltaic solar systems. In some examples, a method is performed by a computer system and includes supplying a consumption time series and a predicted production time series for an electric system to a machine-learning predictor trained during a prior training phase using electric energy consumption training data and photovoltaic production training data. The consumption time series has a first data resolution, and the electric energy consumption training data and the photovoltaic production training data have a second data resolution greater than the first data resolution. The method includes determining, using an output of the machine-learning predictor, a predicted import time series of electric import values each specifying an amount of electric energy predicted to be imported by the electric system with a prospective photovoltaic solar system installed.

Abstract: Methods are provided for communicating between devices in a network and remote servers, which may be located behind intermediate devices such as load balancers, by encapsulating messages sent by those devices and, in one implementation, to a load balancer in a transport header that may be understood by that load balancer; decapsulating the message from the transport header; re-encapuslating the message in a GRE tunnel and passing the message to a server, where the GRE tunnel is removed. Methods are also provided for communicating between devices in a network and local gateways by encapsulating messages sent by those devices and, in one implementation, to a load balancer in a transport header that may be understood by that gateway, and decapsulating the message from the transport header at the gateway.

Abstract: Methods, systems, and computer storage media are disclosed for determining electric energy flow predictions for electric systems including photovoltaic solar systems. In some examples, a method is performed by a computer system and includes supplying a consumption time series and a predicted production time series for an electric system to a machine-learning predictor trained during a prior training phase using electric energy consumption training data and photovoltaic production training data. The consumption time series has a first data resolution, and the electric energy consumption training data and the photovoltaic production training data have a second data resolution greater than the first data resolution. The method includes determining, using an output of the machine-learning predictor, a predicted import time series of electric import values each specifying an amount of electric energy predicted to be imported by the electric system with a prospective photovoltaic solar system installed.

Abstract: Methods for bridging computers functioning as Publish/Subscribe brokers are provided to build a distributed cluster for scalability by enabling publishers to connect to one broker and subscribers to that publisher to connect to a different broker. In one aspect, a method includes listening to an internal bridge port on one broker for receiving bridge connections from other brokers in the cluster, and in parallel, connecting to the internal bridge ports of each of the other brokers in the cluster, waiting for all such incoming and outgoing bridge connections to be completed, and then opening up the external port for serving actual clients. The method also includes creating separate bridges for each direction of traffic to make the broker code completely symmetric and also supporting the creation of an arbitrary number of bridge connections between each pair of brokers for each direction to reduce latency.

Abstract: Methods, systems, and computer storage media are disclosed for determining electric energy flow predictions for electric systems including photovoltaic solar systems. In some examples, a method is performed by a computer system and includes supplying a consumption time series and a predicted production time series for an electric system to a machine-learning predictor trained during a prior training phase using electric energy consumption training data and photovoltaic production training data. The consumption time series has a first data resolution, and the electric energy consumption training data and the photovoltaic production training data have a second data resolution greater than the first data resolution. The method includes determining, using an output of the machine-learning predictor, a predicted import time series of electric import values each specifying an amount of electric energy predicted to be imported by the electric system with a prospective photovoltaic solar system installed.

Abstract: Methods are provided for communicating between devices in a network and remote servers, which may be located behind intermediate devices such as load balancers, by encapsulating messages sent by those devices and, in one implementation, to a load balancer in a transport header that may be understood by that load balancer; decapsulating the message from the transport header; re-encapuslating the message in a GRE tunnel and passing the message to a server, where the GRE tunnel is removed. Methods are also provided for communicating between devices in a network and local gateways by encapsulating messages sent by those devices and, in one implementation, to a load balancer in a transport header that may be understood by that gateway, and decapsulating the message from the transport header at the gateway.

Abstract: Methods, systems, and computer storage media are disclosed for determining electric energy flow predictions for electric systems including photovoltaic solar systems. In some examples, a method is performed by a computer system and includes supplying a consumption time series and a predicted production time series for an electric system to a machine-learning predictor trained during a prior training phase using electric energy consumption training data and photovoltaic production training data. The consumption time series has a first data resolution, and the electric energy consumption training data and the photovoltaic production training data have a second data resolution greater than the first data resolution. The method includes determining, using an output of the machine-learning predictor, a predicted import time series of electric import values each specifying an amount of electric energy predicted to be imported by the electric system with a prospective photovoltaic solar system installed.

Abstract: Methods for bridging computers functioning as Publish/Subscribe brokers (e.g. MQTT brokers) are provided as a means to build a distributed cluster for scalability by enabling publishers to connect to one broker and subscribers to that publisher to connect to a different broker. In one aspect, a method includes listening to an internal bridge port on one broker for receiving bridge connections from other brokers in the cluster, and in parallel, connecting to the internal bridge ports of each of the other brokers in the cluster, waiting for all such incoming and outgoing bridge connections to be completed, and then opening up the external port for serving actual clients. The method also includes monitoring that the bridge connection is alive and healthy throughout the operation of the broker and upon detection of connection failure, storing the messages in a persistent store and replaying them on the bridge connection once it is restored.