EFFICIENT STATE MACHINES FOR REAL-TIME DATAFLOW PROGRAMMING
A system for a smart building includes sensors configured to provide measurements using a plurality of communication protocols and an edge device, wherein the edge device provides a data bus, a plurality of protocol agents configured to ingest data from the plurality of sensors using the plurality of protocols and publish the data to the data bus, an analytics engine subscribed to the data on the data bus, and a plurality of containerized applications connected to the data bus.
1 . A system for a smart building, comprising:
sensors configured to provide measurements using a plurality of communication protocols; and
an edge device, wherein the edge device is programmed to provide:
a data bus;
a plurality of protocol agents configured to ingest data from the plurality of sensors using the plurality of protocols and publish the data to the data bus;
an analytics engine subscribed to the data on the data bus; and
a plurality of containerized applications connected to the data bus.
2 . The system of claim 1 , wherein the data bus is configured to provide access to the data and flow of control messages between components of the edge device.
3 . The system of claim 1 , wherein the data bus is configured to enrich the data from the sensors by decoding raw data into a consumable data format and enriching with metadata.
4 . The system of claim 1 , further comprising a cloud computing system communicable with the edge device, wherein the edge device is programmed to provide a data publisher connected to the data bus, wherein the data publisher is configured to subscribe to information on the data bus to be stored in the cloud computing system and publish the information to the cloud computing system.
5 . The system of claim 1 , wherein the analytics engine is configured to determine an output event by:
processing tokens in an input queue without backtracking to identify patterns in a sequence of the tokens matching a predetermined input pattern, wherein the predetermined input patter is stored in a state table format and the tokens correspond to the data from the sensors; and
generating the output event in response to identifying a match between the sequence of tokens and the one or more input patterns.
6 . The system of claim 1 , wherein the edge device comprises computer memory providing a time-series database.
7 . The system of claim 1 , further comprising a computing device remote from the edge device, wherein the plurality of containerized applications are dynamically mobile between the edge device and the computing device.
8 . The system of claim 1 , wherein the edge device is further programmed to provide a virtual sensor configured to provide data onto the data bus.
9 . The system of claim 1 , wherein the data bus is configured for streaming of the data on the data bus and the analytics engine is configured to perform pattern recognition on the streaming data in a first-in-first-out queue without backtracking.
10 . The system of claim 1 , wherein the analytics engine comprises a set of state machine arranged as a dataflow program.
11 . An edge device for a smart building, the edge device comprising circuitry programmed to provide:
a data bus;
a plurality of protocol agents configured to ingest data from a plurality of sensors using a plurality of communication protocols of the plurality of sensors and publish the data to the data bus;
an analytics engine subscribed to the data on the data bus; and
a plurality of containerized applications connected to the data bus.
12 . The edge device of claim 11 , wherein the data bus is configured to provide access to the data and flow of control messages between components of the edge device.
13 . The edge device of claim 11 , wherein the data bus is configured to enrich the data from the sensors by decoding raw data into a consumable data format and enriching with metadata.
14 . The edge device of claim 11 , wherein the analytics engine is configured to determine an output event by:
processing tokens in an input queue without backtracking to identify patterns in a sequence of the tokens matching a predetermined input pattern, wherein the predetermined input patter is stored in a state table format and the tokens correspond to the data from the sensors; and
generating the output event in response to identifying a match between the sequence of tokens and the one or more input patterns.
15 . The edge device of claim 11 , wherein the edge device comprises computer memory providing a time-series database.
16 . The edge device of claim 11 , wherein the plurality of containerized applications are dynamically mobile between the edge device and a computing device remote from the edge device.
17 . The edge device of claim 11 , wherein the circuitry is further programmed to provide a virtual sensor configured to provide data onto the data bus.
18 . The edge device of claim 11 , wherein the data bus is configured for streaming of the data on the data bus and the analytics engine is configured to perform pattern recognition on the streaming data in a first-in-first-out queue without backtracking.
19 . The edge device of claim 11 , wherein the analytics engine comprises a set of state machine arranged as a dataflow program.
20 . The edge device of claim 11 , wherein the circuitry is further programed to provide a time-series database locally on the edge device.