Method and apparatus to manage multi-computer supply using a model based on an economic model of supply and cost of supply

• US 7,653,907 B2
• Filed: 10/23/2007
• Issued: 01/26/2010
• Est. Priority Date: 08/31/2001
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Wilkins, Thomas Wood, Moran, David Erin, Ditlow, Gary, Williams, Ralph James, Dooling, Daria Rose
• Primary Examiner(s): An; Meng-Ai
• Assistant Examiner(s): Tang; Kenneth
• Application Number:
  US11/976,281
• Publication Number:
  US 20080046887A1
• Time in Patent Office: 826 Days
• Field of Search: 718/100, 718/104, 709/201, 709217-218, 709223-224, 709/229, 705/1, 705/26, 705/37
• US Class Current: 718/104
• CPC Class Codes: G06F 2209/503 Resource availability
• CPC Class Codes: G06F 9/505 considering the load

Abstract

A method and structure for determining a listing of host processors on a network to perform a parallel application, including determining a listing of all possible hosts on the network for performing the parallel application, determining for each of the possible hosts a current capacity and a current utilization, calculating for each of the possible hosts a difference between the current capacity and the current utilization, and selecting from the listing of all possible hosts a listing of hosts based on sorting the calculated differences.

Claims

1. An apparatus, comprising:

a memory;

a network interface; and

a central processing unit (CPU), as interconnected by at least one bus, said CPU executing a set of instructions stored on said memory to select resources on a network connected to said apparatus via said network interface that are most suitable to execute a parallel processing application, said selection based upon an economic model of supply and cost of supply,

wherein said memory stores, for each resource on said network available to participate in said parallel processing application, a vector quantity representing;

a capacity; and

a current utilization, and

wherein said cost of supply is calculated by;

$\mathrm{Cost}<br><br>\left(t\right)=1-\underset{t_0}{\stackrel{T}{\int <br><br>}}<br><br>\mathrm{Supply}<br><br>\left(\mathrm{\tau <br><br>}\right)<br><br>d<br><br>\mathrm{\tau <br><br>},<br><br><br><br>\mathrm{where}<br><br><br><br>\mathrm{Supply}\left(\mathrm{\tau <br><br>}\right)=\frac{\mathrm{capacity}-\mathrm{utilization}\left(\mathrm{\tau <br><br>}\right)}{\mathrm{capacity}},$

t₀=an initial time,

T=a sampling period, and τ

=t/T.

2. The apparatus of claim 1, wherein said capacity vector comprises <

CPU, memory, temporary file, cache>

for each of an available computer of said network.

3. The apparatus of claim 1, said resource selection determined by ranking said cost of supply for each said resource on said network and selecting resources that are least costly.

4. The apparatus of claim 3, said resource selection being influenced by a history of said Supply (τ

) for each said resource, as measured by a median box plot.

5. The apparatus of claim 4, wherein said resource is not selected if its instantaneous value of Supply (τ

) is outside a median box of its median box plot.

6. The apparatus of claim 1, wherein said selection of resources is executed in real time concurrently with said parallel processing application.

7. A method of selecting resources for executing a parallel processing application, said method comprising:

executing, using a processor on a computer, a set of instructions that select resources on a network that are most suitable to execute a parallel processing application, said selection based upon an economic model of supply and cost of supply,

wherein, for each resource on said network available to participate in said parallel processing application, a vector quantity representing;

a capacity; and

a current utilization is calculated and stored in a memory, and

wherein said cost of supply is calculated by;

$\mathrm{Cost}<br><br>\left(t\right)=1-\underset{t_0}{\stackrel{T}{\int <br><br>}}<br><br>\mathrm{Supply}<br><br>\left(\mathrm{\tau <br><br>}\right)<br><br>d<br><br>\mathrm{\tau <br><br>},<br><br><br><br><br><br>\mathrm{where}<br><br><br><br>\mathrm{Supply}\left(\mathrm{\tau <br><br>}\right)=\frac{\mathrm{capacity}-\mathrm{utilization}\left(\mathrm{\tau <br><br>}\right)}{\mathrm{capacity}},$

t₀=an initial time,

T=a sampling period, and τ

=t/T.

8. The method of claim 7, wherein said capacity vector comprises <

CPU, memory, temporary file, cache>

for each of an available computer of said network.

9. The method of claim 7, said resource selection determined by ranking said cost of supply for each said resource on said network and selecting resources that are least costly.

10. The method of claim 9, said resource selection being influenced by a history of said Supply (τ

) for each said resource, as measured by a median box plot.

11. The method of claim 10, wherein said resource is not selected if its instantaneous value of Supply (τ

) is outside a median box of its median box plot.

12. The method of claim 7, wherein said selection of resources is executed in real time concurrently with said parallel processing application.

13. The method of claim 7, as embodied in a set of machine-readable instructions in a storage device.

14. A tangible machine-readable storage medium having tangibly encoded therein a set of machine-readable instructions to execute the method of claim 7.

15. The machine-readable storage medium of claim 14, as comprising one of:

a standalone diskette or other memory medium capable of being inserted into a computer drive;

a memory device on a computer that stores said instructions for later execution by said computer or for transmission of said instructions to another computer; and

a memory device on said computer storing said instructions as said method is being executed by said computer.