Transaction-enabled systems and methods for resource acquisition for a fleet of machines
The present disclosure describes transaction-enabling systems and methods. A system can include a controller and a fleet of machines, each having at least one of a compute task requirement, a networking task requirement, and an energy consumption task requirement. The controller may include a resource requirement circuit to determine an amount of a resource for each of the machines to service the task requirement for each machine, a forward resource market circuit to access a forward resource market, and a resource distribution circuit to execute an aggregated transaction of the resource on the forward resource market.
1. A system, comprising:
a fleet of machines each having at least one of a compute task requirement, a networking task requirement, or an energy consumption task requirement;
a controller, comprising:
a data collection circuit structured to collect data from one or more data inputs, wherein the data inputs include an Internet of Things (IoT) data source,
wherein the controller prepares training data from historical data from the data inputs, wherein the training data is prepared as a result of the controller organizing the historical data from the data inputs as a stream of events and de-duplicating the historical data from the data inputs;
a resource requirement circuit structured to determine an amount of a resource for each of the machines to service at least one of the compute task requirement, the networking task requirement, or the energy consumption task requirement for each corresponding machine,
wherein the resource requirement circuit includes a first machine learning component that is trained on the training data to determine relationships between the data inputs and the at least one of the compute task requirement, the networking task requirement, or the energy consumption task requirement,
wherein the first machine learning component determines, as predictive data inputs, which of the data inputs are likely to be predictive of the amount of the resource for each of the machines to service the at least one of the compute task requirement, the networking task requirement, or the energy consumption task requirement, and
wherein the resource requirement circuit is structured to determine the amount of the resource based on the predictive data inputs determined by the first machine learning component;
a forward resource market circuit structured to access a forward resource market; and
a resource distribution circuit structured to execute an aggregated transaction of the resource on the forward resource market in response to the determined amount of the resource for each of the machines, wherein the resource distribution circuit includes a time delay neural network (TDNN) structured to process sensor data from distinct streams of the data inputs, to add delays to one or more of the data inputs to the TDNN or between one or more nodes of the TDNN to align the distinct streams in time, and to analyze the time-aligned distinct streams together to understand a change in a price pattern in the forward resource market,
wherein the resource distribution circuit is further structured to adaptively improve, using a continuous improvement circuit, an aggregate output value of the fleet of machines or decrease a total resource utilization of the fleet of machines for the aggregate output value using the executed aggregated transaction on the forward resource market,
wherein the aggregate output value includes at least one of an output volume, an output quantity, or an output quality of the fleet of machines, and
wherein the continuous improvement circuit includes at least one of a second machine learning component, an artificial intelligence component, or a neural network component.
2. The system of claim 1 , wherein the resource comprises a compute resource, and wherein the forward resource market comprises a forward market for compute resources.
3. The system of claim 1 , wherein the resource comprises a spectrum allocation resource, and wherein the forward resource market comprises a forward market for spectrum allocation.
4. The system of claim 1 , wherein the resource comprises an energy credit resource, and wherein the forward resource market comprises a forward market for energy credits.
5. The system of claim 1 , wherein the resource comprises an energy resource, and wherein the forward resource market comprises a forward market for energy.
6. The system of claim 1 , wherein the resource comprises a data storage resource, and wherein the forward resource market comprises a forward market for data storage capacity.
7. The system of claim 1 , wherein the resource comprises an energy storage resource, and wherein the forward resource market comprises a forward market for energy storage capacity.
8. The system of claim 1 , wherein the resource comprises a network bandwidth resource, and wherein the forward resource market comprises a forward market for network bandwidth.
9. The system of claim 1 , wherein the aggregated transaction of the resource on the forward resource market comprises one of buying or selling the resource.
10. A method, comprising:
collecting data from one or more data inputs, wherein the data inputs include an Internet of Things (IoT) data source,
preparing training data from historical data from the data inputs, wherein the training data is prepared as a result of organizing the historical data from the data inputs as a stream of events and de-duplicating the historical data from the data inputs;
determining an amount of a resource, for each of machine of a fleet of machines, to service at least one of a compute task requirement, a networking task requirement, or an energy consumption task requirement for each corresponding machine;
training a first machine learning component on the training data to determine relationships between the data inputs and the at least one of the compute task requirement, the networking task requirement, or the energy consumption task requirement,
the first machine learning component determining, as predictive data inputs, which of the data inputs are likely to be predictive of the amount of the resource for each of the machines to service the at least one of the compute task requirement, the networking task requirement, or the energy consumption task requirement;
determining the amount of the resource based on the predictive data inputs determined by the first machine learning component;
accessing a forward resource market;
using a time delay neural network (TDNN), processing sensor data from distinct streams of the data inputs and adding delays to one or more of the data inputs to the TDNN or between one or more nodes of the TDNN to align the distinct streams in time;
using the TDNN, analyzing the time-aligned distinct streams together to understand a change in a price pattern in the forward resource market; and
adaptively improving an aggregate output value of the fleet of machines or decreasing a total resource utilization of the fleet of machines for the aggregate output value by executing an aggregated transaction of the resource on the forward resource market in response to the determined amount of the resource for each of the machines,
wherein the aggregate output value includes at least one of an output volume, an output quantity, or an output quality of the fleet of machines.
11. The method of claim 10 , wherein the resource comprises a compute resource, and wherein the forward resource market comprises a forward market for compute resources.
12. The method of claim 10 , wherein the resource comprises a spectrum allocation resource, and wherein the forward resource market comprises a forward market for spectrum allocation.
13. The method of claim 10 , wherein the resource comprises an energy credit resource, and wherein the forward resource market comprises a forward market for energy credits.
14. The method of claim 10 , wherein the resource comprises an energy resource, and wherein the forward resource market comprises a forward market for energy.
15. The method of claim 10 , wherein the resource comprises a data storage resource, and wherein the forward resource market comprises a forward market for data storage capacity.
16. The method of claim 10 , wherein the resource comprises an energy storage resource, and wherein the forward resource market comprises a forward market for energy storage capacity.
17. The method of claim 10 , wherein the resource comprises a network bandwidth resource, and wherein the forward resource market comprises a forward market for network bandwidth.
18. The method of claim 10 , wherein executing the aggregated transaction of the resource on the forward resource market comprises one of buying or selling the resource.
19. The system of claim 1 , wherein the resource distribution circuit is further structured to execute the executed aggregated transaction using an application programming interface (API) of the forward resource market.
20. The system of claim 19 , wherein the resource distribution circuit is further structured to determine other attributes for the executed aggregated transaction based on data from the Internet of Things (IoT) data source.
21. The system of claim 20 , wherein the data from the IoT data source includes sensor data from points of use for energy or compute resources.
22. The system of claim 1 , wherein the resource distribution circuit is further structured to decrease the total resource utilization of the fleet of machines by redistributing a resource utilization of the resource for each of the machines between the compute task requirement, the networking task requirement, and the energy consumption task requirement.
23. The system of claim 1 , wherein the aggregate output value includes the output quantity, and the output quantity is a quantity of a cryptocurrency coin.
24. The system of claim 1 , wherein:
the continuous improvement circuit includes the second machine learning component that is trained on the training data to determine a favorable timing for the executed aggregated transaction based on the data from the data inputs; and
the resource distribution circuit is structured to execute the aggregated transaction of the resource on the forward resource market at the favorable timing.