Thermal relationships based workload planning

• US 8,949,647 B2
• Filed: 04/30/2012
• Issued: 02/03/2015
• Est. Priority Date: 08/18/2010
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Nastacio, Denilson
• Primary Examiner(s): Lee, Thomas
• Assistant Examiner(s): Chan, Danny
• Application Number:
  US13/459,377
• Publication Number:
  US 20120216065A1
• Time in Patent Office: 1,009 Days
• Field of Search: 713/340, 713/300, 713/320
• US Class Current: 713/340
• CPC Class Codes: G06F 1/206 comprising thermal management
• CPC Class Codes: G06F 9/505 considering the load
• CPC Class Codes: H05K 7/20836 Thermal management, e.g. server temperature control

Abstract

A method for thermal relationships based workload planning in a data processing environment is provided in the illustrative embodiments. For a data processing system having a cooling correspondence with a cooling unit in the data processing environment, a determination is made of a fraction of an additional heat generation potential of the data processing system that the cooling unit will have to dissipate. Using the fraction and a cooling capacity utilization of the cooling unit at a given time, an expected cooling capacity utilization is computed. A workload on the data processing system is adjusted in response to the expected cooling capacity utilization exceeding a threshold.

Claims

1. A computer implemented method for thermal relationships based workload planning in a data processing environment, the computer implemented method comprising:

determining, for a data processing system having a cooling correspondence with a cooling unit in the data processing environment, a fraction of an additional heat generation potential of the data processing system that the cooling unit will have to dissipate, wherein the additional heat generation potential of the data processing system is an amount of heat, in addition to a present amount of heat generated by the data processing system, that the data processing system will generate when the data processing system is operated at a predetermined capacity level, and wherein the cooling correspondence is a thermal relationship between the data processing system and the cooling unit representing a fraction of a total cooling used by the data processing system that is provided from the cooling unit;

computing, using the fraction and a cooling capacity utilization of the cooling unit at a given time, an expected cooling capacity utilization; and

adjusting a workload on the data processing system, responsive to the expected cooling capacity utilization exceeding a threshold.

2. The computer implemented method of claim 1, further comprising:

identifying a set of factors the contribute to heat generation in the data processing system;

computing a utilization of each factor in the set at the given time;

estimating a maximum power consumption of the data processing system; and

computing the additional heat generation potential of the data processing system using a power consumption of the data processing system at the given time, the utilization of each factor, and the estimated maximum power consumption.

3. The computer implemented method of claim 1, further comprising:

repeating the determining for a plurality of data processing systems, each data processing system in the plurality having a corresponding cooling correspondence with the cooling unit, wherein the expected cooling capacity utilization is computed using a fraction of additional heat generation potential determined for each data processing system in the plurality.

4. The computer implemented method of claim 1, wherein the adjusting further comprises:

selecting a first workload to be scheduled on the data processing system, the first workload contributing a first amount to the additional heat generation potential of the data processing system that the cooling unit has to dissipate;

selecting a second data processing system such that scheduling the first workload on the second data processing system causes the first workload to contribute a second amount to the additional heat generation potential of the second data processing system that the cooling unit has to dissipate, the second data processing system having a second cooling correspondence with the cooling unit; and

scheduling the first workload on the second data processing system responsive to the second amount being less than the first amount.

5. The computer implemented method of claim 1, wherein the adjusting further comprises:

selecting a first workload to be scheduled on the data processing system at the given time, the first workload contributing a first amount to the additional heat generation potential of the data processing system that the cooling unit has to dissipate;

selecting a second workload executing at the given time such that the second workload contributes a second amount to heat output of the data processing system at the given time, the second amount being substantially similar to the first amount; and

scheduling the first workload on the data processing system after the second workload has completed execution.

6. The computer implemented method of claim 1, further comprising:

allowing the workload to be scheduled on the data processing system at the given time, responsive to the expected cooling capacity utilization not exceeding a threshold.

7. The computer implemented method of claim 1, wherein the threshold is a maximum allowable cooling capacity utilization of the cooling unit.