Selective modification of power states based on conditions

• US 10,642,341 B2
• Filed: 03/23/2018
• Issued: 05/05/2020
• Est. Priority Date: 03/23/2018
• Status: Active Grant

• Current Assignee: Juniper Networks Incorporated
• Original Assignee: Juniper Networks Incorporated
• Inventors: Srinivasan, Eswaran
• Primary Examiner(s): Choudhury, Zahid
• Application Number:
  US15/934,684
• Publication Number:
  US 20190294233A1
• Time in Patent Office: 774 Days
• CPC Class Codes: G06F 1/3206 Monitoring of events, devices or parameters that trigger a change in power modality
• CPC Class Codes: G06F 1/325 Power saving in peripheral device
• CPC Class Codes: G06F 1/3287 by switching off individual functional units in the computer system
• CPC Class Codes: G06F 1/3296 by lowering the supply or operating voltage
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: H04L 12/10 Current supply arrangements
• CPC Class Codes: H04L 41/0654 using network fault recovery ring fault isolation or reconfiguration in loop networks without recovery actions by a network management system H04L12/437
• CPC Class Codes: H04L 41/12 Discovery or management of network topologies
• CPC Class Codes: H04L 41/147 for predicting network behaviour
• CPC Class Codes: H04L 41/16 using machine learning or artificial intelligence
• CPC Class Codes: H04L 43/0817 by checking functioning
• CPC Class Codes: H04L 49/30 Peripheral units, e.g. input or output ports
• CPC Class Codes: H04L 49/405 Physical details, e.g. power supply, mechanical construction or backplane of ATM switches
• CPC Class Codes: H04L 49/555 Error detection

Abstract

A network device may identify a configuration of resources that are to support attachable line cards. The configuration may include a power supply configuration that is used to provide power to packet processing components that are supported by the line cards, and a resource distribution configuration indicating whether resources in the line cards are shared between the packet processing components. The network device may determine whether to modify a power state of a packet processing component based on whether one or more power modification conditions are satisfied. The network device may modify the power state of the packet processing component based on determining that the power modification condition is satisfied. The power state of the packet processing component may be able to be modified to a particular power state based on the configuration of resources.

Claims

1. A network device, comprising:

one or more line cards,

wherein the one or more line cards support one or more groups of packet processing components; and

one or more processors to;

execute a scanning technique to identify a configuration of resources that are to support the one or more groups of packet processing components,

wherein the configuration includes;

a power supply configuration that is used to provide power to the one or more groups of packet processing components, and

a resource distribution configuration indicating whether resources in the one or more line cards are shared between packet processing components included in the one or more groups of packet processing components;

train, before identifying the configuration of resources that are to support the one or more groups of packet processing components, a machine learning model to identify time periods at which errors on hardware components are likely to occur,

wherein the machine learning model has been trained on historical data that identifies trends of when certain types of hardware equipment are to age to a point of causing errors;

provide hardware information for the network device as input to the machine learning model to cause the machine learning model to output information identifying a projected time period at which an error on a hardware component of the packet processing component is likely to occur;

determine whether to modify a power state of a packet processing component, of the one or more groups of packet processing components, based on whether a power modification condition is satisfied,

wherein the power state of the packet processing component is determined to be modified prior to the projected time period; and

modify the power state of the packet processing component based on determining that the power modification condition is satisfied,

wherein the power state is capable of being modified to;

a full power state,

a low power state, or

a no power state,

wherein the power state of the packet processing component is able to be modified to a particular power state based on the configuration of resources that is to support the one or more groups of packet processing components, and

wherein the power state of the packet processing component is modified to the no power state or to the low power state prior to the projected time period.

2. The network device of claim 1, wherein the one or more processors, when modifying the power state of the packet processing component, are to:

modify the power state of the packet processing component without modifying power states of other packet processing components of the one or more groups of packet processing components,

wherein the power state is modified from the full power state to;

the low power state, or

the no power state.

3. The network device of claim 1, wherein the one or more processors, when determining whether to modify the power state of the packet processing component, are to:

receive an instruction to power on the packet processing component,

determine to modify the power state of the packet processing component based on receiving the instruction; and

wherein the one or more processors, when modifying the power state of the packet processing component, are to;

modify the power state of the packet processing component from the no power state to the full power state, and

modify power states for any other packet processing components, of the one or more groups of packet processing components, that share power resources or line card resources with the packet processing component,

wherein modifying the power state for any other packet processing components includes modifying the power state from the no power state to the full power state, and subsequently modifying the power state from the full power state to the low power state.

4. The network device of claim 1, wherein the one or more processors, when determining whether to modify the power state of the packet processing component, are to:

monitor capacity levels for one or more additional packet processing components, of the one or more groups of packet processing components, that are powered on and receiving traffic,

wherein the packet processing component, of the one or more groups of packet processing components, is in the no power state or the low power state and is not receiving traffic,

determine that a capacity level for a particular packet processing component, of the one or more additional packet processing components, satisfies a threshold capacity level, and

determine to modify the power state of the packet processing component based on determining that the capacity level for the particular packet processing component satisfies the threshold capacity level; and

wherein the one or more processors, when modifying the power state of the packet processing component, are to;

modify the power state to the full power state to enable the packet processing component to receive traffic and to reduce the capacity level of the particular packet processing component.

5. The network device of claim 1, wherein the one or more processors, when determining whether to modify the power state of the packet processing component, are to:

monitor one or more hardware components of the packet processing component while the packet processing component is powered on and receiving traffic,

detect an error on a hardware component, of the one or more hardware components, based on monitoring the one or more hardware components, and

determine to modify the power state of the packet processing component based on detecting the error on the hardware component; and

wherein the one or more processors, when modifying the power state of the packet processing component, are to;

modify the power state to the no power state or to the low power state based on the power supply configuration and the resource distribution configuration for a line card used to support the one or more groups of packet processing components,

wherein the line card is one of the one or more line cards.

6. The network device of claim 1, wherein the one or more processors, when providing the hardware information for the network device as input to the machine learning model, are further to provide one or more of:

information associated with an age of the hardware equipment as input to the machine learning model, or

information associated with a type of hardware equipment as input to the machine learning model.

7. The network device of claim 1, wherein the one or more processors, when modifying the power state of the packet processing component, are to:

modify the power state from the full power state to the low power state based on identifying that the packet processing component;

shares line card resources with at least one other packet processing component in the one or more groups of packet processing components, or

shares power resources with at least one other packet processing component in the one or more groups of packet processing components, or

modify the power state from the full power state to the no power state based on identifying that the packet processing component does not share line card resources with the at least one other packet processing component and based on identifying that the packet processing component does not share power resources with the at least one other packet processing component.

8. A non-transitory computer-readable medium storing instructions, the instructions comprising:

one or more instructions that, when executed by one or more processors, cause the one or more processors to;

execute a scanning technique to identify a configuration of resources that are to support one or more line cards that are attachable to a device associated with the one or more processors,

wherein the configuration includes at least one of;

a power supply configuration that is used to provide power to one or more groups of packet processing components that are supported by the one or more line cards, or

a resource distribution configuration indicating whether resources in the one or more line cards are shared between packet processing components included in the one or more groups of packet processing components;

train, before identifying the configuration of resources that are to support the one or more groups of packet processing components, a machine learning model to identify time periods at which errors on hardware components are likely to occur,

wherein the machine learning model has been trained on historical data that identifies trends of when certain types of hardware equipment are to age to a point of causing errors;

provide hardware information for the device as input to the machine learning model to cause the machine learning model to output information identifying a projected time period at which an error on a hardware component of the packet processing component is likely to occur;

determine whether to modify a power state of a packet processing component, of the one or more groups of packet processing components, based on whether a power modification condition is satisfied

wherein the power state of the packet processing component is determined to be modified prior to the projected time period; and

modify the power state of the packet processing component based on determining that the power modification condition is satisfied,

wherein the power state is capable of being modified from a first power state to;

a second power state, or

a third power state,

wherein the power state of the packet processing component is able to be modified to the second power state or to the third power state based on the configuration of resources that support the one or more line cards, and

wherein the power state of the packet processing component is modified to the third power state or to the second power state prior to the projected time period.

9. The non-transitory computer-readable medium of claim 8, wherein the one or more instructions, that cause the one or more processors to modify the power state, cause the one or more processors to:

modify the power state of the packet processing component from the first power state to the second power state or to the third power state,

wherein the first power state is a full power state, the second power state is a low power state, and the third power state is a no power state, and

wherein the power state is able to be modified without modifying power states of other packet processing components of the one or more groups of packet processing components.

10. The non-transitory computer-readable medium of claim 8, wherein the one or more instructions, that cause the one or more processors to execute the scanning technique to identify the configuration of resources, cause the one or more processors to:

execute the scanning technique to identify the configuration of resources,

wherein the scanning technique identifies a number of power supplies that are used to support the one or more groups of packet processing components, and

wherein the scanning technique identifies, for a line card of the one or more line cards, whether packet processing components that are supported by the line card share line card resources or have dedicated line card resources.

11. The non-transitory computer-readable medium of claim 8, wherein the one or more instructions, that cause the one or more processor to determine whether to modify the power state of the packet processing component, cause the one or more processors to:

monitor one or more elements of the packet processing component while the packet processing component is powered on and receiving traffic,

detect an error associated with an element, of the one or more elements, based on monitoring the one or more elements, and

determine to modify the power state of the packet processing component based on detecting the error associated with the element; and

wherein the one or more instructions, that cause the one or more processors to modify the power state of the packet processing component, cause the one or more processors to;

modify the power state from the first power state to the second power state or to the third power state based on the power supply configuration and the resource distribution configuration for a line card used to support the packet processing component.

12. The non-transitory computer-readable medium of claim 8, wherein the one or more instructions, when executed by the one or more processors, further cause the one or more processors to:

provide client data as input to the machine learning model to cause the machine learning model to output information identifying the projected time period at which an increase in traffic is likely to occur;

and

wherein the one or more instructions, that cause the one or more processors to modify the power state of the packet processing component, cause the one or more processors to;

modify the power state of the packet processing component from the first power state to the second power state or to the third power state prior to the project time period,

wherein the first power state is a full power state, the second power state is a low power state, and the third power state is a no power state.

13. The non-transitory computer-readable medium of claim 8, wherein the one or more instructions, that cause the one or more processor to determine whether to modify the power state, cause the one or more processors to:

monitor capacity levels for one or more additional packet processing components, of the one or more groups of packet processing components, that are powered on and receiving traffic,

wherein the packet processing component is in the first power state or the second power state and is not receiving traffic, and

wherein the first power state is a no power state and the second power state is a low power state,

determine that a capacity level for a particular packet processing component, of the one or more additional packet processing components, satisfies a threshold capacity level, and

determine to modify the power state of the packet processing component based on determining that the capacity level for the particular packet processing component satisfies the threshold capacity level; and

wherein the one or more instructions, that cause the one or more processors to modify the power state of the packet processing component, cause the one or more processors to;

modify the power state to the third power state to enable the packet processing component to receive traffic and to reduce the capacity level of the particular packet processing component,

wherein the third power state is a full power state.

14. The non-transitory computer-readable medium of claim 8, wherein the one or more instructions, that cause the one or more processors to modify the power state of the packet processing component, cause the one or more processors to:

modify the power state from the first power state to the second power state based on identifying that the packet processing component;

shares a line card resource with at least one other packet processing component of the one or more groups of packet processing components, and/or

shares a power resource with at least one other packet processing component of the one or more groups of packet processing components,

wherein the first power state is a full power state and the second power state is a low power state, or

modify the power state from the first power state to the third power state based on identifying that the packet processing component does not share a line card resource with the at least one other packet processing component and based on identifying that the packet processing component does not share a power resource with the at least one other packet processing component,

wherein the third power state is a no power state.

15. A method, comprising:

identifying, by a network device, a configuration of resources that are to support one or more line cards that are attachable to the network device,

wherein the configuration includes;

a power supply configuration that is used to provide power to one or more groups of packet processing components that are supported by the one or more line cards, and

a resource distribution configuration indicating whether resources in the one or more line cards are shared between packet processing components included in the one or more groups of packet processing components;

training, by the network device and before identifying the configuration of resources that are to support the one or more groups of packet processing components, a machine learning model to identify time periods at which errors on hardware components are likely to occur,

wherein the machine learning model has been trained on historical data that identifies trends of when certain types of hardware equipment are to age to a point of causing errors;

providing, by the network device, hardware information for the network device as input to the machine learning model to cause the machine learning model to output information identifying a projected time period at which an error on a hardware component of the packet processing component is likely to occur;

determining, by the network device, whether to modify a power state of a packet processing component, of the one or more groups of packet processing components, based on whether one or more power modification conditions are satisfied,

wherein the power state of the packet processing component is determined to be modified prior to the projected time period; and

modifying, by the network device, the power state of the packet processing component based on determining that the power modification condition is satisfied,

wherein the power state of the packet processing component is able to be modified to a particular power state based on the configuration of resources supporting the one or more groups of packet processing components, and

wherein the power state of the packet processing component is modified to the particular power state prior to the projected time period.

16. The method of claim 15, wherein the power state of the packet processing component is able to be modified without modifying power states of other packet processing components of the one or more groups of packet processing components.

17. The method of claim 15, wherein modifying the power state of the packet processing component comprises:

selectively modifying the power state of the packet processing component from a full power state to a low power state or from the full power state to a no power state based on the identified configuration of resources.

18. The method of claim 15, wherein identifying the configuration of resources comprises:

executing a scanning technique to identify the configuration of resources,

wherein executing the scanning technique causes one or more processors of the network device to identify the configuration of resources by analyzing one or more data structures that store the power supply configuration and the resource distribution configuration.

19. The method of claim 15, wherein determining whether to modify the power state of the packet processing component comprises:

determining that a capacity level for another packet processing component, of the one or more groups of packet processing components, satisfies a threshold capacity level,

wherein the packet processing component, of the one or more groups of packet processing components, is in a no power state and is not receiving traffic, and

determining to modify the power state of the packet processing component based on determining that the capacity level for the other packet processing component satisfies the threshold capacity level; and

wherein modifying the power state of the packet processing component comprises;

modifying the power state from the no power state to a full power state to enable the packet processing component to receive traffic and to reduce the capacity level of the other packet processing component.

20. The method of claim 15, further comprising:

providing the identified configuration of resources and additional network device information as input to the machine learning model to cause the machine learning model to output information identifying the projected time period at which an error is likely to occur;

and

wherein modifying the power state of the packet processing component comprises;

modifying the power state of the packet processing component from an on state to a no power state or to a low power state prior to the projected time period.