Load balancing internet-of-things (IOT) gateways
View Patent ↗Systems and methods for load balancing Internet-of-Things (IoT) gateways. In some embodiments, an Information Handling System (IHS) may include a processor and a memory coupled to the processor, the memory including program instructions stored thereon that, upon execution by the processor, cause the IHS to: receive metadata including attributes of IoT devices communicatively coupled to the IHS, wherein the attributes are collected by the IoT devices at least in part during a time interval leading up to the IHS entering an undesired state; perform feature extraction upon the metadata; identify correlations between extracted features and the undesired state; and perform load balancing based upon the correlations.
1. An Information Handling System (IHS), comprising:
one or more processors; and
a memory coupled to the one or more processors, the memory including program instructions stored thereon that, upon execution by the one or more processors, cause the IHS to:
receive metadata including attributes of Internet-of-Things (IoT) devices communicatively coupled to the IHS, wherein the attributes are collected by the IoT devices at least in part during a time interval leading up to the IHS entering an undesired state, and wherein the undesired state is characterized by at least one of: (a) an average processor utilization rate using a 2-second sampling rate across a 30-sample range is greater than 80%; (b) an average memory bus utilization rate using a 2-second sampling rate across a 30-sample range is greater than 80%; (c) an average storage utilization rate using a 2-second sampling rate across a 30-sample range is greater than 80%; or (d) a running average power using a 2-second sampling rate of package power across a 30-sample range is greater than 80% of a Thermal Design Point for a model of the one or more processors;
perform feature extraction upon the metadata;
identify correlations between extracted features and the undesired state; and
perform load balancing based upon the correlations.
2. The IHS of claim 1 , wherein the IoT devices comprise: movement sensors, power sensors, airflow sensors, temperature sensors, humidity sensors, GPS devices, RFID readers, door locks, compressor devices, valve devices, fan devices, pump devices, switch devices, thermostat devices, smoke alarm devices, or fire alarm devices.
3. The IHS of claim 1 , wherein the attributes include, for each IoT device, at least one of: an identification of the IoT device, an identification of a type of IoT device, a movement status, power status, airflow status, temperature status, humidity status, GPS status, RFID status, door lock status, compressor status, valve status, fan status, pump status, switch status, thermostat status, smoke alarm device status, or fire alarm device status.
4. The IHS of claim 1 , wherein the metadata further comprises attributes of the IHS selected from the group consisting of: number of processor cores, average processor utilization, average processor frequency, USB device count, average USB block size, average USB throughput, serial device count, and serial average throughput.
5. The IHS of claim 1 , wherein the IHS enters in an undesired state when it has at least one of: saturated processor, saturated memory, saturated storage, high average processor power, or high average processor temperature.
6. The IHS of claim 1 , wherein performing the feature extraction includes performing a principal component analysis using eigenvalue decomposition or singular value decomposition of a matrix including the metadata.
7. The IHS of claim 1 , wherein load balancing includes at least one of: rescheduling tasks within the IHS to different times, reprioritizing tasks within the IHS, or rebalancing tasks across different IHSs.
8. The IHS of claim 1 , wherein identifying the correlations includes determining that the undesirable state is reached when a subset of IoT devices are simultaneously queried, and wherein load balancing includes staggering the queries to avoid the undesired state.
9. The IHS of claim 1 , wherein identifying the correlations includes determining that the undesirable state is reached when a subset of IoT devices are simultaneously queried and the IHS is executing a given process, and wherein load balancing includes offsetting the process from the queries in time.
10. The IHS of claim 1 , wherein the program instructions, upon execution, further cause the IHS to receive updated metadata from at least one of IoT device due to a change in the IoT device.
11. The IHS of claim 10 , wherein the program instructions, upon execution, further cause the IHS to perform another load balancing based, at least in part, upon an updated feature extraction.
12. The IHS of claim 1 , wherein performing the feature extraction and identifying the correlations occur during a period of under-utilization of the IHS, and wherein the period of under-utilization is characterized by one or more of: idle processor, idle memory, idle storage, low average processor power, or low average processor temperature.
13. A hardware memory device having program instructions stored thereon that, upon execution by an Information Handling System (IHS), cause the IHS to:
receive metadata including attributes of Internet-of-Things (IoT) devices communicatively coupled to the IHS, wherein the attributes are collected by the IoT devices at least in part during a time interval leading up to the IHS entering an undesired state;
perform feature extraction upon the metadata;
identify correlations between extracted features and the undesired state in response to: (a) an average processor utilization rate using a 2-second sampling rate across a 30-sample range being less than 5%; an average memory bus utilization rate using a 2-second sampling rate across a 30-sample range being less than 15%; (c) an average storage utilization rate using a 2-second sampling rate across a 30-sample range being less than 20%; and (d) a running average power using a 2-second sampling rate of package power across a 30-sample range being less than 30% of a Thermal Design Point for a processor model; and
perform load balancing based upon the correlations.
14. The hardware memory device of claim 13 , wherein identifying the correlations includes determining that the undesirable state is reached when a subset of IoT devices are simultaneously queried, and wherein load balancing includes staggering the queries to avoid the undesired state.
15. The hardware memory device of claim 14 , wherein the subset of devices includes a door device, a temperature device, or a flow sensor.
16. The hardware memory device of claim 13 , wherein identifying the correlations includes determining that the undesirable state is reached when a subset of IoT devices are simultaneously queried and the IHS is executing a given process, and wherein load balancing includes offsetting the process from the queries in time.
17. The hardware memory device of claim 16 , wherein the process includes at least one of: the feature extraction operation, the correlation operation, or the load balancing operation.
18. A method, comprising:
receiving, by an Information Handling System (IHS), metadata including attributes of Internet-of-Things (IoT) devices communicatively coupled to the IHS, wherein the attributes are collected by the IoT devices at least in part during a time interval leading up to the IHS entering an undesired state, and wherein the undesired state is characterized by at least one of: (a) an average processor utilization rate using a 2-second sampling rate across a 30-sample range is greater than 80%; (b) an average memory bus utilization rate using a 2-second sampling rate across a 30-sample range is greater than 80%; (c) an average storage utilization rate using a 2-second sampling rate across a 30-sample range is greater than 80%; or (d) a running average power using a 2-second sampling rate of package power across a 30-sample range is greater than 80% of a Thermal Design Point for a processor model;
performing, by the IHS, feature extraction upon the metadata;
identifying, by the IHS, correlations between extracted features and the undesired state in response to: (a) an average processor utilization rate using a 2-second sampling rate across a 30-sample range being less than 5%; an average memory bus utilization rate using a 2-second sampling rate across a 30-sample range being less than 15%; (c) an average storage utilization rate using a 2-second sampling rate across a 30-sample range being less than 20%; and (d) a running average power using a 2-second sampling rate of package power across a 30-sample range being less than 30% of a Thermal Design Point for a processor model; and
performing, by the IHS, load balancing based upon the correlations.
19. The method of claim 18 , wherein identifying the correlations includes determining that the undesirable state is reached when a subset of IoT devices are simultaneously queried, and wherein load balancing includes staggering the queries to avoid the undesired state.
20. The method of claim 18 , wherein identifying the correlations includes determining that the undesirable state is reached when a subset of IoT devices are simultaneously queried and the IHS is executing a given process, and wherein load balancing includes offsetting the process from the queries in time.