Global throttling of computing nodes in a modular, rack-configured information handling system

Patent metadata

Field	Value
Publication number	US-9625974-B2
Application number	US-201314139739-A
Country	US
Kind code	B2
Filing date	Dec 23, 2013
Priority date	Dec 23, 2013
Publication date	Apr 18, 2017
Grant date	Apr 18, 2017

Abstract

A computer-implemented method enables global throttling of processing nodes in a rack-configured information handling system (RIHS). A rack-level management controller receives power-usage data and operating parameters associated with processing nodes within separately-controlled blocks of the RIHS. A power subsystem of the RIHS regulates an amount of power supplied to the processing nodes of the RIHS based on the power-usage data and operating parameters for the processing nodes and a total amount of available power for distribution within the RIHS. In response to detecting a condition that reduces the total amount of available power for distribution within the IHS, the management controller autonomously initiates global throttling of the processing nodes within the IHS to reduce power consumption by at least one of the processing nodes. The global throttling is completed via a signal transfer over a select Ethernet cable wire to connected block controllers that control the processing nodes.

First claim

What is claimed is:

1. A computer implemented method that implements rack-level power control of processing nodes in a rack-configured information handling system, the method comprising:
receiving, at a rack-level management controller, power-usage data, operating parameters, and settings information associated with one or more processing nodes within one or more separately-controlled blocks of the information handling system (IHS);
triggering a power subsystem of the IHS to regulate an amount of power supplied to one or more of the processing nodes of the information handling system based on the received information for the one or more processing nodes and a total amount of available power for distribution within the IHS; and
in response to detecting a condition that reduces the total amount of available power for distribution within the IHS, the rack-level management controller autonomously initiating global throttling of power allocation within the IHS to reduce power consumption by at least one of the processing nodes via a separate block controller, wherein the global throttling is triggered by asserting a global throttle wire that is a subset of available wires within an Ethernet cable interconnecting the rack-level management controller and one or more block-level controllers;
wherein the communication between the management controller and each of the one or more block controllers occurs over an Ethernet cable, which has a subset of individual signal wires assigned for general system and network communication between an infrastructure manager and the one or more block controllers, based in part on triggering commands received from the management controller;
wherein at least one signal wire in the Ethernet cable is assigned as a global throttle wire for a specific connected one of the one or more block controllers to provide communication of a GTPR signal from the infrastructure controller to the specific block controller; and
wherein at least one other signal wire in the Ethernet cable is assigned as a global reset wire for a specific connected one of the one or more block controllers to allow for expedient communication of a reset signal from the management controller via the infrastructure controller to the specific block controller.

2. The method of claim 1 , further comprising:
initializing the management controller during start up configuration of the IHS;
establishing communication between the management controller, the power subsystem, and one or more block controllers that each control block-level operations of processing nodes within a corresponding block;
retrieving power profile data for each block within the IHS, the power profile data comprising the power-usage data and settings information of the individual processing nodes within the particular block;
generating a power allocation for each block based on the received power profile data; and
triggering the power subsystem to supply power to each block and by extension to each of the nodes based on the power allocation for that block.

3. The method of claim 1 , wherein the condition is one of a power supply unit (PSU) failure and an AC input power failure, and the method further comprises:
determining which condition has occurred;
in response to determining that the PSU failure has occurred, transmitting a signal to at least one of the block controllers to reduce power to at least one of the nodes, wherein an overall power is reduced by a first amount correlated to a loss of a PSU; and
in response to determining that the AC input power failure has occurred, transmitting a signal to at least one of the block controllers to reduce power to at least one of the nodes, wherein an overall power is reduced by a second amount correlated to a loss of an AC input.

4. The method of claim 3 , further comprising:
determining via at least one block controller, which processing nodes of the corresponding block are to be throttled in order to reduce power consumption by the corresponding block; and
the at least one of the block controller subsequently throttling operations of the processing nodes that are to be throttled within the corresponding blocks.

5. The method of claim 1 , further comprising:
establishing a maximum power capacity for the information handing system;
monitoring a total power usage and demand across the IHS; and
in response to determining that a current power demand across the IHS is greater than the maximum power capacity, transmitting a signal to at least one of the block controllers to reduce power to at least one of the nodes.

6. The method of claim 1 , wherein initiating global throttling of power allocation within the IHS to reduce power consumption by at least one of the processing nodes via an associated block controller comprises generating a signal on a global throttle wire allocated within each Ethernet cable directly connecting each block controller within the IHS with the management controller, wherein the block controllers within the IHS are pre-programmed to respond to an assertion of a signal on the global throttle wire by the management controller by immediately throttling operations of one or more processing nodes within a respective block being controlled by the block controller.

7. The method of claim 1 , further comprising:
in response to detecting that the overall thermal readings of the IHS are above a system thermal threshold, identifying at least one block drawing large amounts of power and contributing to higher than normal thermal threshold readings; and
transmitting a signal to at least each block controller corresponding to the identified at least one block to reduce power consumption by the at least one node.

8. The method of claim 1 , further comprising:
determining whether the condition remains present after reducing power to at least one of the nodes; and
in response to the condition remaining present, triggering at least one of the block controllers to shut down at least one of the nodes.

9. The method of claim 1 , further comprising:
determining whether at least one of an infrastructure manager or the block controllers are to be reset;
in response to determining that at least one of the infrastructure manager or the block controllers are to be reset, generating a reset signal on a reset wire allocated within each Ethernet cable directly connecting the management controller with the infrastructure manager and the block controllers, wherein the reset signal triggers at least one of the infrastructure manager or the block controllers to reset.

10. A rack-level power control system comprising:
a rack-level management controller having a processor and a memory coupled to the processor via a system interconnect;
a power subsystem communicatively coupled to the rack-level management controller;
a cooling subsystem communicatively coupled to the rack-level management controller;
one or more blocks communicatively coupled to the rack-level management controller, the blocks having at least one block controller communicatively coupled to the rack-level management controller and each of the blocks having one or more processing nodes; and
an Ethernet cable connected between the rack-level management controller and at least one of the block controllers, wherein: at least one of the wires in the Ethernet cable allocated as a global throttle wire to provide communication of a global throttle power reduction signal from the rack-level management controller to at least one of the block controllers and at least one of the nodes allowing an immediate reduction in power usage by the processing nodes upon receipt of the global throttle power reduction signal; a subset of individual signal wires wi

Assignees

• Dell Inc
• Dell Products Lp

Inventors

• Bailey Edmond
• Stuewe John R
• Vancil Paul W

Classifications

• G06F1/206

  comprising thermal management · CPC title

• G06F1/3234 Primary

  Power saving characterised by the action undertaken · CPC title

• G06F1/28

  Supervision thereof, e.g. detecting power-supply failure by out of limits supervision · CPC title

• G06F1/3206

  Monitoring of events, devices or parameters that trigger a change in power modality · CPC title

• G06F1/3296

  by lowering the supply or operating voltage · CPC title