Energy aware processing load distribution system and method

Patent metadata

Field	Value
Publication number	US-12189449-B2
Application number	US-202318387052-A
Country	US
Kind code	B2
Filing date	Nov 5, 2023
Priority date	Jul 21, 2009
Publication date	Jan 7, 2025
Grant date	Jan 7, 2025

Abstract

A method for controlling a data center, comprising a plurality of server systems, each associated with a cooling system and a thermal constraint, comprising: a concurrent physical condition of a first server system; predicting a future physical condition based on a set of future states of the first server system; dynamically controlling the cooling system in response to at least the input and the predicted future physical condition, to selectively cool the first server system sufficient to meet the predetermined thermal constraint; and controlling an allocation of tasks between the plurality of server systems to selectively load the first server system within the predetermined thermal constraint and selectively idle a second server system, wherein the idle second server system can be recruited to accept tasks when allocated to it, and wherein the cooling system associated with the idle second server system is selectively operated in a low power consumption state.

First claim

What is claimed is:

1. A load distributor for a set of servers, the set of servers being associated with a cooling system, each server having respective energy consumption and performance characteristics, the load distributor comprising:
an input configured to receive a new task to be processed;
at least one processor, configured to analyze the new task, and to allocate the new task to a respective server of the set of servers to optimize an energy efficiency of the set of servers, dependent on at least:
the respective energy consumption and performance characteristics of each server of the set of servers, and
a thermal environment of each server of the set of servers; and

an output configured to communicate the allocation of the task to the respective server.

2. The load distributor according to claim 1 , further comprising a queue comprising a set of existing tasks pending processing, wherein the at least one processor is further configured to reallocate an existing task for processing by a first server to a second server selectively dependent on a risk of exceeding a thermal constraint by the first server.

3. The load distributor according to claim 1 , wherein the cooling system comprises an associated dynamically controllable cooling system for each server, and the at least one automated processor is further configured to alter the associated dynamically controllable cooling system in dependence on a predicted effect of the new task on the thermal environment of the respective server.

4. The load distributor according to claim 3 , wherein the at least one automated processor is further configured to control the associated dynamically controllable cooling system for the respective server in anticipation of at least a future workload of the respective server.

5. The load distributor according to claim 1 , wherein the at least one automated processor is further configured to activate and deactivate the respective server.

6. The load distributor according to claim 1 , wherein the at least one automated processor is further configured to allocate the new task to a respective server selectively in dependence on a predicted efficiency of the cooling system.

7. The load distributor according to claim 1 , wherein the at least one automated processor is further configured to allocate the new task to a respective server selectively in dependence on a predicted efficiency of the cooling system.

8. The load distributor according to claim 1 , further comprising a thermal and performance computational model of the respective server, wherein the at least one automated processor is further configured to predict a temperature of the respective server based on at least the thermal and performance computational model a sequence of tasks performed by the respective server.

9. The load distributor according to claim 1 , wherein the respective server has a dynamic voltage and frequency scaling setting, the set of servers comprise servers having a plurality of different dynamic voltage and frequency scaling settings, and at least one automated processor is further configured to characterize the new task with respect to a repetitive task type, and to allocate the characterized new task to the respective server selectively based on the dynamic voltage and frequency scaling setting.

10. The load distributor according to claim 1 , wherein the at least one automated processor is further configured to allocate the task to the respective server further dependent on a headroom of the respective server to accept additional tasks after allocation of the respective task to the respective server.

11. The load distributor according to claim 1 , wherein the energy efficiency of the set of servers comprises an energy consumption of a cooling system for the set of servers.

12. The load distributor according to claim 1 , wherein the at least one processor is further configured to concentrate allocation of tasks to a subset of an available portion of the set of servers.

13. A load distributor for a set of servers, each server having respective energy consumption and performance characteristics and an associated cooling system, the load distributor comprising:
an input configured to receive a new task to be processed, the task having a value of execution and an energy consumption of execution;
at least one processor, configured to analyze the new task, and to allocate the new task to a respective server of the set of servers to optimize a cost-benefit of the set of servers, dependent on at least:
the respective energy consumption and performance characteristics of each server of the set of servers,
an energy consumption of the associated cooling system for each server, and
a thermal environment of each server of the set of servers; and

an output configured to communicate the allocation of the task to the respective server.

14. The load distributor according to claim 13 , further comprising a queue comprising a set of existing tasks pending processing, wherein the at least one processor is further configured to reallocate an existing task for processing by a first server to a second server selectively dependent on a risk of exceeding a thermal constraint by the first server.

15. The load distributor according to claim 13 , wherein the at least one automated processor is further configured to dynamically alter the associated cooling system in dependence on a predicted effect of the new task on the thermal environment of the respective server.

16. The load distributor according to claim 13 , wherein the at least one automated processor is further configured to control a cooling system for the respective server in anticipation of at least a future workload of the respective server.

17. The load distributor according to claim 13 , further comprising a thermal and performance computational model of the respective server, wherein the at least one automated processor is further configured to predict a temperature of the respective server based on at least the thermal and performance computational model a sequence of tasks performed by the respective server.

18. The load distributor according to claim 13 , wherein the respective server has a dynamic voltage and frequency scaling setting, the set of servers comprise servers having a plurality of different dynamic voltage and frequency scaling settings, and at least one automated processor is further configured to characterize the new task with respect to a repetitive task type, and to allocate the characterized new task to the respective server selectively based on the dynamic voltage and frequency scaling setting.

19. A method of distributing computational tasks to a set of server systems, the set of server systems being associated with a cooling system, each server system having respective energy consumption and performance characteristics, the method comprising:
receiving a new task to be processed;
analyzing the new task;
allocating the new task to a respective server system of the set of servers to optimize an energy efficiency of the set of server systems, dependent on at least:
the respective energy consumption and performance characteristics of each server system of the set of server systems, and
a thermal environment of each server system of the set of server systems; and

communicating the allocation of the task to the respective server system.

20. The method according to claim 19 , wherein the energy efficiency of the set of server systems comprises an energy consumption of a computational comp

Assignees

• Univ New York State Res Found

Inventors

• Ghose Kanad

Classifications

• A61K36/185

  Magnoliopsida (dicotyledons) · CPC title

• Y02D10/00

  Energy efficient computing, e.g. low power processors, power management or thermal management · CPC title

• H05K7/20836

  Thermal management, e.g. server temperature control · CPC title

• G05D23/19

  characterised by the use of electric means {(G05D23/1393 takes precedence)} · CPC title

• G06F1/3209 Primary

  Monitoring remote activity, e.g. over telephone lines or network connections · CPC title