Managing workload distribution among computing systems to optimize heat dissipation by computing systems

• US 9,176,508 B2
• Filed: 01/09/2012
• Issued: 11/03/2015
• Est. Priority Date: 01/09/2012
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Geissler, Andrew, Hollinger, Michael C.
• Primary Examiner(s): Ali, Mohammad
• Assistant Examiner(s): Pan, Yuhui R
• Application Number:
  US13/345,995
• Publication Number:
  US 20130178999A1
• Time in Patent Office: 1,394 Days
• Field of Search: 700/300
• US Class Current: 1/1
• CPC Class Codes: G05D 23/19 characterised by the use of electric means G05D23/1393 takes precedence
• CPC Class Codes: G05D 23/1934 each space being provided with one sensor acting on one or more control means
• CPC Class Codes: G06F 9/5094 where the allocation takes into account power or heat criteria power management in computers in general G06F1/3203; thermal management in computers in general G06F1/206
• CPC Class Codes: H05K 7/20836 Thermal management, e.g. server temperature control
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

A thermal workload distribution controller receives, for each of multiple thermal controlled areas, at least one current thermal measurement from at least one separate thermostat node. The thermal workload distribution controller selects a particular thermal controlled area from among the multiple thermal controlled areas that is most optimal to receive additional heat based on the at least one current thermal measurement received for each of the thermal controlled areas. The thermal workload distribution controller distributes at least one workload to a server node associated with the particular thermal controlled area, wherein the heat dissipated by the server from executing the workload affects a thermal environment of the particular thermal controlled area.

Claims

1. A system for managing distribution of workloads in a computing environment, comprising:

a thermal workload distribution controller, operative on at least one computing node comprising at least one processor and at least one memory, and operative to receive, for each of a plurality of thermal controlled areas, at least one current thermal measurement, at least one requested thermal temperature, and a thermal output from at least one separate thermostat node from among a plurality of thermostat nodes via a network;

the thermal workload distribution controller operative to select a particular thermal controlled area from among the plurality of thermal controlled areas that is most optimal to receive additional heat, wherein;

if at least one first selection of thermal controlled areas from among the plurality of thermal controlled areas comprises the thermal output set to heat and at least one differential between the at least one requested temperature and the at least one particular current temperature for the at least one first selection of thermal controlled areas, wherein the at least one current temperature is lower than the at least one requested temperature, the particular thermal controlled area is set to a first thermal controlled area, from among the at least one first selection of thermal controlled areas, with a largest differential between the at least one requested temperature and the at least one particular temperature for the at least one first selection of thermal controlled areas; and

if at least one second selection of thermal controlled areas from among the plurality of thermal controlled areas comprises the thermal output set to heat and the at least one request temperature equal to the at least one particular current temperature for the at least one second selection of thermal controlled areas, the particular thermal controlled area is set to a second thermal controlled area, from among the at least one second selection of thermal controlled areas, requiring the greatest amount of heating output as specified in the thermal output;

the thermal workload distribution controller operative to distribute at least one workload, from among a plurality of workloads distributed by the thermal workload distribution controller across a plurality of server nodes each associated with one from among the plurality of thermal controlled areas, to at least one server node associated with the particular thermal controlled area from among the plurality of server nodes, wherein the heat dissipated by the at least one server node from executing the workload affects a thermal environment of the particular thermal controlled area;

a separate thermal controller associated with at least one server node from among the plurality of server nodes;

each separate thermal controller operative to receive, for a separate thermal controlled area from among the plurality of thermal controlled areas at least one current thermal measurement, at least one requested thermal temperature, and a thermal output from the at least one separate thermostat node from among the plurality of thermostat nodes via a network,

each separate thermal controller operative to compare the at least one requested thermal temperature with the at least one current thermal measurement;

each separate thermal controller operative, in response to determining the at least one requested thermal temperature is greater than the at least one current thermal measurement, for voting to increase a processor frequency and voltage on the at least one server node to unthrottle memory on the at least one server node, and to reduce the fan speed for the at least one server node;

each separate thermal controller operative, in response to determining the at least one requested thermal temperature is equal to the at least one current thermal measurement;

if the thermal output is heat, voting to increase a processor frequency and voltage on the at least one server node, to unthrottle memory on the at least one server node, and to reduce the fan speed for the at least one server node; and

if the thermal output is cool, voting to increase the fan speed for the at least one server node; and

each separate thermal controller operative, in response to determining the at least one requested thermal temperature is less than the at least one current thermal measurement, for voting to reduce the processor frequency and voltage on the at least one server node and to increase the fan speed for the at least one server node.

2. The system according to claim 1, further comprising:

the thermal workload distribution controller operative to receive the at least one current thermal measurement, the at least one requested thermal temperature, and the thermal output by a separate thermal control system for providing a heating output from each at least one separate thermostat node.

3. The system according to claim 1, further comprising:

the thermal workload distribution controller operative to identify the at least one server node associated with the particular thermal controlled area, wherein the at least one server node operates in a first area outside of the particular thermal controlled area and the heat dissipated by the at least one server node is directed by a data thermal heater into the particular thermal controlled area to maintain the at least one requested temperature for the particular thermal controlled area, wherein the data thermal heater heats the particular thermal controlled area in part through the heat dissipated by the at least one server node, wherein the data thermal heater directs the heat dissipated by the at least one server node to an exterior area if the heat is not required in the particular thermal controlled area.

4. The system according to claim 1, further comprising:

the thermal workload distribution controller identifying the server node associated with the particular thermal controlled area and operating in the particular thermal controlled area.

5. The system according to claim 1, wherein a thermal workload distribution controller, operative on at least one computing node comprising at least one processor and at least one memory, and operative to receive, for each of a plurality of thermal controlled areas, at least one current thermal measurement, at least one requested thermal temperature, and a thermal output from at least one separate thermostat node via a network, further comprises:

the thermal workload distribution controller operative to receive the at least one current thermal measurement, the at least one requested thermal temperature, and the thermal output from the at least one separate thermostat node, wherein the at least one separate thermostat node detects the thermal output for reporting, wherein the at least one separate thermostat node receives at least one separate ambient temperature vote from the at least one server node, wherein the at least one separate ambient temperature vote specifies a requested ambient temperature for operation of the at least one server node, wherein the at least one server node controls at least one operational setting for at least one component of the at least one server node to control an amount of thermal output by the at least one server node, wherein the at least one separate thermostat node determines the at least one requested thermal temperature based on the at least one separate ambient temperature vote and at least one additional requested temperature for the particular thermal controlled area.

6. The system according to claim 1, wherein the thermal output comprises a type of thermal output of one of cool output and heat output controlled by a thermal control system for the particular thermal controlled area to adjust an ambient air temperature in the particular thermal controlled area.

7. The system according to claim 1, wherein the voting, by each separate thermal controller, comprises sending one or more actuation signals each separately sent to directly control the processor frequency of one or more processors, to directly control throttling of the memory, and to directly control the fan speed of one or more fans of the at least one server node.

8. The system according to claim 1, wherein each separate thermal controller is a separate controller from the thermal workload distribution controller.

9. A computer program product for managing distribution of workloads in a computing environment, comprising:

one or more non-transitory computer-readable, tangible storage devices;

program instructions, stored on at least one of the one or more storage devices, to receive, for each of a plurality of thermal controlled areas, at least one current thermal measurement, at least one requested thermal temperature, and a thermal output from at least one separate thermostat node from among a plurality of thermostat nodes;

program instructions, stored on at least one of the one or more storage devices, to select a particular thermal controlled area from among the plurality of thermal controlled areas that is most optimal to receive additional heat, wherein;

if at least one first selection of thermal controlled areas from among the plurality of thermal controlled areas comprises the thermal output set to heat and at least one differential between the at least one requested temperature and the at least one particular current temperature for the at least one first selection of thermal controlled areas, wherein the at least one current temperature is lower than the at least one requested temperature, the particular thermal controlled area is set to a first thermal controlled area, from among the at least one first selection of thermal controlled areas, with a largest differential between the at least one requested temperature and the at least one particular temperature for the at least one first selection of thermal controlled areas; and

if at least one second selection of thermal controlled areas from among the plurality of thermal controlled areas comprises the thermal output set to heat and the at least one request temperature equal to the at least one particular current temperature for the at least one second selection of thermal controlled areas, the particular thermal controlled area is set to a second thermal controlled area, from among the at least one second selection of thermal controlled areas, requiring the greatest amount of heating output as specified in the thermal output;

program instructions, stored on at least one of the one or more storage devices, to distribute at least one workload, from among a plurality of workloads distributed across a plurality of server nodes each associated with one from among the plurality of thermal controlled areas, to at least one server node associated with the particular thermal controlled area from among the plurality of server nodes, wherein the heat dissipated by the at least one server node from executing the workload affects a thermal environment of the particular thermal controlled area;

program instructions, stored on at least one of the one or more storage devices, to control a separate thermal controller associated with at least one server node from among the plurality of server nodes;

program instructions, stored on at least one of the one or more storage devices, to receive, by each separate thermal controller, for a separate thermal controlled area from among the plurality of thermal controlled areas, at least one current thermal measurement, at least one requested thermal temperature, and a thermal output from the at least one separate thermostat node from among the plurality of thermostat nodes via a network;

program instructions, stored on at least one of the one or more storage devices, to compare, by each separate thermal controller, the at least one requested thermal temperature with the at least one current thermal measurement;

program instructions, stored on at least one of the one or more storage devices, for each separate thermal controller operative, in response to determining the at least one requested thermal temperature is greater than the at least one current thermal measurement, to vote to increase a processor frequency and voltage on the at least one server node, to unthrottle memory on the at least one server node, and to reduce the fan speed for the at least one server node;

program instructions, stored on at least one of the one or more storage devices, for each separate thermal controller operative, in response to determining the at least one requested thermal temperature is equal to the at least one current thermal measurement;

if the thermal output is heat, to vote to increase a processor frequency and voltage on the at least one server node, to unthrottle memory on the at least one server node, and to reduce the fan speed for the at least one server node; and

if the thermal output is cool, to vote to increase the fan speed for the at least one server node; and

program instructions, stored on at least one of the one or more storage devices, for each separate thermal controller operative, in response to determining the at least one requested thermal temperature is less than the at least one current thermal measurement, to vote to reduce the processor frequency and voltage on the at least one server node and to increase the fan speed for the at least one server node.

10. The computer program product of claim 9, further comprising:

program instructions, stored on at least one of the one or more storage devices, to receive the at least one current thermal measurement by a separate thermal control system for providing a heating output.