EFFICIENT MAINTENANCE FOR COMMUNICATION DEVICES
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for determining a terminal idle time. In some implementations, a server can obtain communication data from a plurality of devices in a communication network, wherein the communication data indicates levels of network traffic for the device over time. The server can generate an idle period forecasting model configured to predict occurrence of future communication idle periods in which communication activity is predicted to be below a threshold. The server can provide the idle period forecasting model to each of the plurality of devices such that the devices can respectively use the idle period forecasting model to locally predict future communication idle periods of the devices.
1 . (canceled)
2 . A method performed by a communication device, the method comprising:
receiving, by the communication device, an idle period forecasting model from one or more computers over a communication network, wherein the idle period forecasting model is configured to predict occurrences of future communication idle periods in which network traffic of the communication device over the communication network is below a threshold;
generating, by the communication device, feature data based on communication data indicating prior communication activity involving the communication device, wherein the feature data indicates amounts of network traffic over the communication network over time;
providing, by the communication device, the generated feature data as input to the idle period forecasting model;
obtaining, by the communication device, output of the idle period forecasting model that is generated in response to input of the generated feature data, wherein the output of the idle period forecasting model includes an idle period prediction that indicates a predicted future communication idle period for the communication device; and
performing, by the communication device, a maintenance operation for the communication device with timing of the maintenance operation being determined based on the idle period prediction.
3 . The method of claim 2 , wherein the communication device is a satellite terminal.
4 . The method of claim 2 , wherein the communication device provides network connectivity to one or more client devices, such that the one or more client devices send and receive network traffic over a network through the communication device;
wherein the method comprises obtaining communication data that describes network traffic over one or more network connections provided by the communication device for the one or more client devices;
wherein generating the feature data comprises generating feature data that indicates amounts of network traffic over time for the one or more network connections provided by the communication device to the one or more client devices;
wherein obtaining the output of the idle period forecasting model comprises obtaining output that indicates a predicted future communication idle period in which network traffic from connected client devices is predicted to be below the threshold.
5 . The method of claim 2 , wherein performing the maintenance operation for the communication device comprises initiating, by the communication device, the maintenance operation in a time range that the idle period prediction of the idle period forecasting model indicates that a communication channel is likely to be idle.
6 . The method of claim 2 , wherein receiving the idle period forecasting model comprises receiving the idle period forecasting model over a satellite network.
7 . The method of claim 2 , wherein receiving the idle period forecasting model comprises receiving the idle period forecasting model by multicast over a satellite network.
8 . The method of claim 2 , wherein the maintenance operation comprises at least one of a software update for the communication device or a reboot of the communication device.
9 . The method of claim 2 , further comprising determining, by the communication device, an expected duration for completion of the maintenance operation;
wherein the maintenance operation for the communication device is initiated in response to the communication device determining that the predicted future communication idle period is at least as long as the expected duration for completion of the maintenance operation.
10 . The method of claim 2 , further comprising:
based on the output of the idle period forecasting model, determining, by the communication device, a start time for the predicted future communication idle period and a duration of the predicted future communication idle period;
selecting, by the communication device, the maintenance operation to be performed and completed within the duration of the predicted future communication idle period; and
performing, by the communication device, the maintenance operation within the predicted future communication idle period.
11 . The method of claim 2 , wherein the idle period forecasting model comprises a vector autoregressive (VAR) statistical model and a long short-term memory recurrent neural network (LSTM RNN).
12 . A communication device comprising:
one or more processors; and
one or more machine-readable media storing instructions that are operable, when executed by the one or more processors, to cause the communication device to perform operations comprising:
receiving, by the communication device, an idle period forecasting model from one or more computers over a communication network, wherein the idle period forecasting model is configured to predict occurrences of future communication idle periods in which network traffic of the communication device over the communication network is below a threshold;
generating, by the communication device, feature data based on communication data indicating prior communication activity involving the communication device, wherein the feature data indicates amounts of network traffic over the communication network over time;
providing, by the communication device, the generated feature data as input to the idle period forecasting model;
obtaining, by the communication device, output of the idle period forecasting model that is generated in response to input of the generated feature data, wherein the output of the idle period forecasting model includes an idle period prediction that indicates a predicted future communication idle period for the communication device; and
performing, by the communication device, a maintenance operation for the communication device with timing of the maintenance operation being determined based on the idle period prediction.
13 . The communication device of claim 12 , wherein the communication device is a satellite terminal.
14 . The communication device of claim 12 , wherein the communication device provides network connectivity to one or more client devices, such that the one or more client devices send and receive network traffic over a network through the communication device;
wherein the operations comprise obtaining communication data that describes network traffic over one or more network connections provided by the communication device for the one or more client devices;
wherein generating the feature data comprises generating feature data that indicates amounts of network traffic over time for the one or more network connections provided by the communication device to the one or more client devices;
wherein obtaining the output of the idle period forecasting model comprises obtaining output that indicates a predicted future communication idle period in which network traffic from connected client devices is predicted to be below the threshold.
15 . The communication device of claim 12 , wherein performing the maintenance operation for the communication device comprises initiating, by the communication device, the Page. maintenance operation in a time range that the idle period prediction of the idle period forecasting model indicates that a communication channel is likely to be idle.
16 . The communication device of claim 12 , wherein receiving the idle period forecasting model comprises receiving the idle period forecasting model over a satellite network.
17 . The communication device of claim 12 , wherein receiving the idle period forecasting model comprises receiving the idle period forecasting model by multicast over a satellite network.
18 . The communication device of claim 12 , wherein the maintenance operation comprises at least one of a software update for the communication device or a reboot of the communication device.
19 . The communication device of claim 12 , further comprising determining, by the communication device, an expected duration for completion of the maintenance operation;
wherein the maintenance operation for the communication device is initiated in response to the communication device determining that the predicted future communication idle period is at least as long as the expected duration for completion of the maintenance operation.
20 . The communication device of claim 12 , further comprising:
based on the output of the idle period forecasting model, determining, by the communication device, a start time for the predicted future communication idle period and a duration of the predicted future communication idle period;
selecting, by the communication device, the maintenance operation to be performed and completed within the duration of the predicted future communication idle period; and
performing, by the communication device, the maintenance operation within the predicted future communication idle period.
21 . One or more non-transitory machine-readable media storing instructions that are operable, when executed by one or more processors of a communication device, to cause the communication device to perform operations comprising:
receiving, by the communication device, an idle period forecasting model from one or more computers over a communication network, wherein the idle period forecasting model is configured to predict occurrences of future communication idle periods in which network traffic of the communication device over the communication network is below a threshold;
generating, by the communication device, feature data based on communication data indicating prior communication activity involving the communication device, wherein the feature data indicates amounts of network traffic over the communication network over time;
providing, by the communication device, the generated feature data as input to the idle period forecasting model;
obtaining, by the communication device, output of the idle period forecasting model that is generated in response to input of the generated feature data, wherein the output of the idle period forecasting model includes an idle period prediction that indicates a predicted future communication idle period for the communication device; and
performing, by the communication device, a maintenance operation for the communication device with timing of the maintenance operation being determined based on the idle period prediction.