Systems and methods for continued edge resource demand load estimation
Managing the resource demand load for edge systems is significantly more complex than for other systems, such as cloud environments. Embodiments herein provide edge resource demand load estimation systems and methods that inform scheduling and associated edge orchestration to ensure that edge system resource capacity is appropriately utilized. Efficient utilization allows an increased number of applications to be deployed at a reduced level of reserved resources. Also presented are embodiments of assurance mechanisms for monitoring edge resource demand load characterizations. In one or more embodiments, when an estimate or estimates are deemed to not be valid (e.g., having experienced stationary drift), updated estimates may be obtained.
1 . A processor-implemented method for estimating edge resource demand load, the method comprising:
given, for each task of a set of one or more tasks, time series of statistical resource demand values for one or more edge resources for handling the task:
gauging stationarity using one or more stationarity methods; and
responsive to detecting non-stationarity for at least one statistical resource demand value for at least one edge resource for a task:
dispatching the task to a target edge system, in which the task is flagged to notify the target edge system to collect resource-related data associated with handling the task;
after receiving the resource-related data associated with handling the task, using a dataset comprising the resource-related data associated with handling the task to determine one or more resource statistics for at least one of the one or more edge resources for the task;
determining, for the task, one or more resource demand values for the at least one of the one or more edge resources using at least one of the one or more resource statistics; and
storing the one or more resource demand values.
2 . The processor-implemented method of claim 1 wherein the one or more stationarity methods comprises at least one of: Kwiatkowski-Phillips-Schmidt-Shin (KPSS) and Augmented Dickey-Fuller (ADF).
3 . The processor-implemented method of claim 1 further comprising:
responsive to not detecting non-stationarity for any statistical resource demand values:
determining whether a last time non-stationarity was detected exceeds a maximum threshold time;
responsive to the last time non-stationarity was detected not exceeding the maximum threshold time, waiting a scheduled time and returning to the step of gauging stationarity using one or more stationarity methods; and
responsive to the last time non-stationarity was detected exceeding the maximum threshold time, for each task from a group of one or more tasks:
dispatching the task to a target edge system, in which the task is flagged to notify the target edge system to collect resource-related data associated with handling the task;
after receiving the resource-related data associated with handling the task, using a dataset comprising the resource-related data associated with handling the task to determine one or more resource statistics for at least one of the one or more edge resources for the task;
determining, for the task, one or more resource demand values for the at least one of the one or more edge resources using at least one of the one or more resource statistics; and
storing the one or more resource demand values.
4 . The processor-implemented method of claim 3 further comprising:
determining a frequency of occurrence of stationary drift; and
using the frequency of occurrence of stationary drift to set the scheduled time for checking for stationary drift.
5 . The processor-implemented method of claim 1 wherein the step of using the dataset comprising the resource-related data associated with handling the task to determine one or more resource statistics for at least one of the one or more edge resources for the task comprises:
using the dataset comprising the resource-related data associated with handling the task and a resource uncertainty estimator, which uses a M-PCM-OFFD (Multivariate Probabilistic Collocation Method-Orthogonal Fractional Factorial Design) methodology, to obtain the resource statistics for the one or more edge resources for the task.
6 . The processor-implemented method of claim 1 wherein the dataset comprises resource-related data associated with handling the task collected from a plurality of instances of the target edge system handling the task over an evaluation time.
7 . The processor-implemented method of claim 1 wherein the one or more resource demand values for the task comprises, for each of a set of edge resources:
a lower control limit for the edge resource;
a mean for the edge resource; and
an upper control limit for the edge resource.
8 . One or more information handling systems collectively comprising:
one or more processors; and
one or more non-transitory computer-readable medium or media comprising one or more sets of instructions which, when executed by at least one of the one or more processors, causes steps to be performed comprising:
given, for each task of a set of one or more tasks, time series of statistical resource demand values for one or more edge resources for handling the task:
gauging stationarity using one or more stationarity methods; and
responsive to detecting non-stationarity for at least one statistical resource demand value for at least one edge resource for a task:
dispatching the task to a target edge system, in which the task is flagged to notify the target edge system to collect resource-related data associated with handling the task;
after receiving the resource-related data associated with handling the task, using a dataset comprising the resource-related data associated with handling the task to determine one or more resource statistics for at least one of the one or more edge resources for the task;
determining, for the task, one or more resource demand values for the at least one of the one or more edge resources using at least one of the one or more resource statistics; and
storing the one or more resource demand values.
9 . The one or more information handling systems of claim 8 wherein the one or more stationarity methods comprises at least one of: Kwiatkowski-Phillips-Schmidt-Shin (KPSS) and Augmented Dickey-Fuller (ADF).
10 . The one or more information handling systems of claim 8 wherein the one or more non-transitory computer-readable medium or media further comprise one or more sets of instructions which, when executed by at least one of the one or more processors, causes steps to be performed comprising:
responsive to not detecting non-stationarity for any statistical resource demand values:
determining whether a last time non-stationarity was detected exceeds a maximum threshold time;
responsive to the last time non-stationarity was detected not exceeding the maximum threshold time, waiting a scheduled time and returning to the step of gauging stationarity using one or more stationarity methods; and
responsive to the last time non-stationarity was detected exceeding the maximum threshold time, for each task from a group of one or more tasks:
dispatching the task to a target edge system, in which the task is flagged to notify the target edge system to collect resource-related data associated with handling the task;
after receiving the resource-related data associated with handling the task, using a dataset comprising the resource-related data associated with handling the task to determine one or more resource statistics for at least one of the one or more edge resources for the task;
determining, for the task, one or more resource demand values for the at least one of the one or more edge resources using at least one of the one or more resource statistics; and
storing the one or more resource demand values.
11 . The one or more information handling systems of claim 10 wherein the one or more non-transitory computer-readable medium or media further comprise one or more sets of instructions which, when executed by at least one of the one or more processors, causes steps to be performed comprising:
determining a frequency of occurrence of stationary drift; and
using the frequency of occurrence of stationary drift to set the scheduled time for checking for stationary drift.
12 . The one or more information handling systems of claim 8 wherein the step of using the dataset comprising the resource-related data associated with handling the task to determine one or more resource statistics for at least one of the one or more edge resources for the task comprises:
using the dataset comprising the resource-related data associated with handling the task and a resource uncertainty estimator, which uses a M-PCM-OFFD (Multivariate Probabilistic Collocation Method-Orthogonal Fractional Factorial Design) methodology, to obtain the resource statistics for the one or more edge resources for the task.
13 . The one or more information handling systems of claim 8 wherein the dataset comprises resource-related data associated with handling the task collected from a plurality of instances of the target edge system handling the task over an evaluation time.
14 . The one or more information handling systems of claim 8 wherein the one or more resource demand values for the task comprises, for each of a set of edge resources:
a lower control limit for the edge resource;
a mean for the edge resource; and
an upper control limit for the edge resource.
15 . A non-transitory computer-readable medium or media comprising one or more sequences of instructions which, when executed by at least one processor, causes steps to be performed comprising:
given, for each task of a set of one or more tasks, time series of statistical resource demand values for one or more edge resources for handling the task:
gauging stationarity using one or more stationarity methods; and
responsive to detecting non-stationarity for at least one statistical resource demand value for at least one edge resource for a task:
dispatching the task to a target edge system, in which the task is flagged to notify the target edge system to collect resource-related data associated with handling the task;
after receiving the resource-related data associated with handling the task, using a dataset comprising the resource-related data associated with handling the task to determine one or more resource statistics for at least one of the one or more edge resources for the task;
determining, for the task, one or more resource demand values for the at least one of the one or more edge resources using at least one of the one or more resource statistics; and
storing the one or more resource demand values.
16 . The non-transitory computer-readable medium or media of claim 15 wherein the one or more stationarity methods comprises at least one of: Kwiatkowski-Phillips-Schmidt-Shin (KPSS) and Augmented Dickey-Fuller (ADF).
17 . The non-transitory computer-readable medium or media of claim 15 further comprising one or more sequences of instructions which, when executed by at least one processor, causes steps to be performed comprising:
responsive to not detecting non-stationarity for any statistical resource demand values:
determining whether a last time non-stationarity was detected exceeds a maximum threshold time;
responsive to the last time non-stationarity was detected not exceeding the maximum threshold time, waiting a scheduled time and returning to the step of gauging stationarity using one or more stationarity methods; and
responsive to the last time non-stationarity was detected exceeding the maximum threshold time, for each task from a group of one or more tasks:
dispatching the task to a target edge system, in which the task is flagged to notify the target edge system to collect resource-related data associated with handling the task;
after receiving the resource-related data associated with handling the task, using a dataset comprising the resource-related data associated with handling the task to determine one or more resource statistics for at least one of the one or more edge resources for the task;
determining, for the task, one or more resource demand values for the at least one of the one or more edge resources using at least one of the one or more resource statistics; and
storing the one or more resource demand values.
18 . The non-transitory computer-readable medium or media of claim 17 further comprising one or more sequences of instructions which, when executed by at least one processor, causes steps to be performed comprising:
determining a frequency of occurrence of stationary drift; and
using the frequency of occurrence of stationary drift to set the scheduled time for checking for stationary drift.
19 . The non-transitory computer-readable medium or media of claim 15 wherein the step of using the dataset comprising the resource-related data associated with handling the task to determine one or more resource statistics for at least one of the one or more edge resources for the task comprises:
using the dataset comprising the resource-related data associated with handling the task and a resource uncertainty estimator, which uses a statistical technique known as M-PCM-OFFD (Multivariate Probabilistic Collocation Method-Orthogonal Fractional Factorial Design), to obtain the resource statistics for the one or more edge resources for the task.
20 . The non-transitory computer-readable medium or media of claim 15 wherein the dataset comprises resource-related data associated with handling the task collected from a plurality of instances of the target edge system handling the task over an evaluation time.