Seasonal auto-reactive energization
Techniques for tuning resource capacity for a network device are described. A component resource manager receives an expected utilization of the resource during a utilization volatility period from an auto energizing infrastructure. The utilization volatility period is a period of time during which a chance of a utilization increase of the resource is within a predetermined threshold. The component resource manager receives a calculation of a maximum potential capacity delta for the expected utilization of the resource during the utilization volatility period from the auto energizing infrastructure. Based on the expected utilization and the maximum potential capacity delta, the component resource manager determines the right-size capacity of the resource for the utilization volatility period. The right-size capacity is at least an amount of the resource needed to ensure proper network functioning. The component resource manager provides the right-size capacity of the resource during the utilization volatility period.
1 . A method performed at least in part by a component resource manager of a network device and for provisioning a right-size capacity of a local resource for the network device, the method comprising:
determining, that a utilization volatility period is beginning, wherein the utilization volatility period is a period of time during which a chance of a utilization increase of the local resource is within a predetermined threshold;
determining, an upper capacity boundary trigger and a lower capacity boundary trigger for the utilization volatility period;
receiving, from an auto energizing infrastructure, an expected utilization of the local resource for the network device during a utilization volatility period;
receiving, from the auto energizing infrastructure, a calculation of a maximum potential capacity delta for the expected utilization of the local resource for the network device during the utilization volatility period, wherein the maximum potential capacity delta is an amount of capacity of the local resource to cover a spike in utilization of the local resource above the expected utilization of the local resource; and
algorithmically self-tune a right-size capacity of the local resource by increasing or decreasing a number of energized hardware elements providing incremental capacity during a duration of the spike based at least in part on whether local resource utilization telemetry for the network device indicates that the upper capacity boundary trigger or the lower capacity boundary trigger has been breached, wherein the right-size capacity is at least an amount of the local resource needed to ensure proper network functioning.
2 . The method of claim 1 , further comprising receiving, from a network controller, an acceptable failure rate for the maximum potential capacity delta, the acceptable failure rate based at least in part on current local traffic conditions.
3 . The method of claim 1 , wherein the maximum potential capacity delta is a maximum local resource capacity measurement increase expected during a sliding period of time having a duration equal to a capacity increase latency, wherein the capacity increase latency is an amount of time between a last measurement of capacity of the local resource that does not exceed a change in capacity trigger and a first measurement after capacity change has been instituted following a first measurement that exceeds the change in capacity trigger.
4 . The method of claim 1 , further comprising:
receiving local resource utilization telemetry for the network device,
based at least in part on the utilization telemetry, determining if an upper capacity boundary has been breached, or a lower capacity boundary has been breached and for a predetermined amount of time capacity measurements are below the lower capacity boundary;
in response to determining the upper capacity boundary has been breached, increase capacity of the local resource for the network device; and
in response to determining the lower capacity boundary has been breached and for the predetermined amount of time capacity measurements are below the lower capacity boundary, decrease capacity of the local resource for the network device.
5 . The method of claim 1 , wherein the expected utilization of the local resource for the network device during a utilization volatility period is determined using a time series forecasting model and is based at least in part on local network conditions.
6 . The method of claim 1 , wherein the right-size capacity of a local resource is at least equal to a predetermined minimum amount of capacity.
7 . The method of claim 1 , wherein determining the right-size capacity of the local resource for the utilization volatility period if further based on a local configuration change.
8 . A system comprising:
one or more processors; and
one or more computer-readable media storing computer-executable instructions that, when executed by the one or more processors, cause the one or more processors to perform operations comprising:
determining, that a utilization volatility period for is beginning, wherein the utilization volatility period is a period of time during which a chance of a utilization increase of a local resource is within a predetermined threshold;
determining, an upper capacity boundary trigger and a lower capacity boundary trigger for the utilization volatility period;
receiving, from an auto energizing infrastructure, an expected utilization of the local resource for a network device during a utilization volatility period;
receiving, from the auto energizing infrastructure, a calculation of a maximum potential capacity delta for the expected utilization of the local resource for the network device during the utilization volatility period, wherein the maximum potential capacity delta is an amount of capacity of the local resource to cover a spike in utilization of the local resource above the expected utilization of the local resource; and
algorithmically self-tune a right-size capacity of the local resource by increasing or decreasing a number of energized hardware elements providing incremental capacity during a duration of the spike based at least in part on whether local resource utilization telemetry for the network device indicates that the upper capacity boundary trigger or the lower capacity boundary trigger has been breached, wherein the right-size capacity is at least an amount of the local resource needed to ensure proper network functioning.
9 . The system of claim 8 , the operations further comprising receiving, from a network controller, an acceptable failure rate for the maximum potential capacity delta, the acceptable failure rate based at least in part on current local traffic conditions.
10 . The system of claim 8 , wherein the maximum potential capacity delta is a maximum local resource capacity measurement increase expected during a sliding period of time having a duration equal to a capacity increase latency, wherein the capacity increase latency is an amount of time between a last measurement of capacity of the local resource that does not exceed a change in capacity trigger and a first measurement after capacity change has been instituted following a first measurement that exceeds the change in capacity trigger.
11 . The system of claim 8 , the operations further comprising:
receiving local resource utilization telemetry for the network device, based at least in part on the utilization telemetry, determining if an upper capacity boundary has been breached, or a lower capacity boundary has been breached and for a predetermined amount of time capacity measurements are below the lower capacity boundary;
in response to determining the upper capacity boundary has been breached, increase capacity of the local resource for the network device; and
in response to determining the lower capacity boundary has been breached and for the predetermined amount of time capacity measurements are below the lower capacity boundary, decrease capacity of the local resource for the network device.
12 . The system of claim 8 , wherein the expected utilization of the local resource for the network device during a utilization volatility period is determined using a time series forecasting model and is based at least in part on local network conditions.
13 . The system of claim 8 , wherein the right-size capacity of a local resource is at least equal to a predetermined minimum amount of capacity.
14 . The system of claim 8 , wherein determining the right-size capacity of the local resource for the utilization volatility period if further based on a local configuration change.
15 . One or more non-transitory computer-readable media storing instructions that, when executed, cause one or more processors to perform operations comprising:
determining, that a utilization volatility period for is beginning, wherein the utilization volatility period is a period of time during which a chance of a utilization increase of a local resource is within a predetermined threshold;
determining, an upper capacity boundary trigger and a lower capacity boundary trigger for the utilization volatility period;
receiving, from an auto energizing infrastructure, an expected utilization of a local resource for a network device during a utilization volatility period;
receiving, from the auto energizing infrastructure, a calculation of a maximum potential capacity delta for the expected utilization of the local resource for the network device during the utilization volatility period, wherein the maximum potential capacity delta is an amount of capacity of the local resource to cover a spike in utilization of the local resource above the expected utilization of the local resource; and
algorithmically self-tune a right-size capacity of the local resource by increasing or decreasing a number of energized hardware elements providing incremental capacity during a duration of the spike based at least in part on whether local resource utilization telemetry for the network device indicates that the upper capacity boundary or the lower capacity boundary trigger has been breached, wherein the right-size capacity is at least an amount of the local resource needed to ensure proper network functioning.
16 . The one or more non-transitory computer-readable media of claim 15 , the operations further comprising receiving, from a network controller, an acceptable failure rate for the maximum potential capacity delta, the acceptable failure rate based at least in part on current local traffic conditions.
17 . The one or more non-transitory computer-readable media of claim 15 , wherein the maximum potential capacity delta is a maximum local resource capacity measurement increase expected during a sliding period of time having a duration equal to a capacity increase latency, wherein the capacity increase latency is an amount of time between a last measurement of capacity of the local resource that does not exceed a change in capacity trigger and a first measurement after capacity change has been instituted following a first measurement that exceeds the change in capacity trigger.
18 . The one or more non-transitory computer-readable media of claim 15 , the operations further comprising:
receiving local resource utilization telemetry for the network device,
based at least in part on the utilization telemetry, determining if an upper capacity boundary has been breached, or a lower capacity boundary has been breached and for a predetermined amount of time capacity measurements are below the lower capacity boundary;
in response to determining the upper capacity boundary has been breached, increase capacity of the local resource for the network device; and
in response to determining the lower capacity boundary has been breached and for the predetermined amount of time capacity measurements are below the lower capacity boundary, decrease capacity of the local resource for the network device.
19 . The one or more non-transitory computer-readable media of claim 15 , wherein the expected utilization of the local resource for the network device during a utilization volatility period is determined using a time series forecasting model and is based at least in part on local network conditions.
20 . The one or more non-transitory computer-readable media of claim 15 , wherein the right-size capacity of a local resource is at least equal to a predetermined minimum amount of capacity.