CLOUD PROVISIONING ACCELERATOR

• US 20130139152A1
• Filed: 11/29/2011
• Published: 05/30/2013
• Est. Priority Date: 11/29/2011
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Chang, Rong N., Perng, Chang-Shing, Shaikh, Hidayatullah H., So, Edward C., Choi, Mihwa, Jiang, Yexi, Laker, Meir, Tao, Tao
• Granted Patent: US 8,826,277 B2
• Time in Patent Office: Days
• US Class Current: 718/1
• CPC Class Codes: G06F 2009/45562 Creating, deleting, cloning virtual machine instances
• CPC Class Codes: G06F 2009/4557 Distribution of virtual machine instances; Migration and load balancing
• CPC Class Codes: G06F 2209/5019 Workload prediction
• CPC Class Codes: G06F 40/00 Handling natural language data speech analysis or synthesis, speech recognition G10L
• CPC Class Codes: G06F 9/45545 Guest-host, i.e. hypervisor is an application program itself, e.g. VirtualBox
• CPC Class Codes: G06F 9/45558 Hypervisor-specific management and integration aspects
• CPC Class Codes: G06F 9/5072 Grid computing

Abstract

High level characteristics are extracted from a virtual machine provisioning request stream for a cloud computing environment. A plurality of prediction models are generated for a plurality of different types of virtual machines. A regression-correlation ensemble is applied to appropriate ones of the prediction models to predict future virtual machine demand for the cloud computing environment.

Claims

1. A method comprising:

extracting high level characteristics from a virtual machine provisioning request stream for a cloud computing environment;

generating a plurality of prediction models for a plurality of different types of virtual machines; and

applying a regression-correlation ensemble to appropriate ones of said prediction models to predict future virtual machine demand for said cloud computing environment.

2. The method of claim 1, further comprising pre-provisioning a plurality of virtual machines in said cloud computing requirement in accordance with said predicted future virtual machine demand.

3. The method of claim 2, wherein said extracting comprises filtering irrelevant features and records and transforming said high level characteristics into time series.

4. The method of claim 3, wherein said filtering filters out characteristics other than virtual machine type and request time.

5. The method of claim 3, wherein said generating step is carried out based on an asymmetric cost measurement that treats under-prediction and over-prediction costs differently.

6. The method of claim 5, wherein said generating step is carried out based on cloud prediction cost.

7. The method of claim 3, further comprising repeating said extracting and generating steps as said virtual machine provisioning request stream moves forward in time.

8. The method of claim 3, wherein said applying step comprises:

combining a plurality of said prediction models via a weighted linear combination to obtain an intermediate prediction; and

post-processing said intermediate prediction via correlation between positively correlated ones of said time series to obtain said predicted future virtual machine demand.

9. The method of claim 3, further comprising reserving idle ones of said pre-provisioned virtual machines in said cloud computing environment in accordance with said predicted future virtual machine demand.

10. The method of claim 1, further comprising providing a system, wherein the system comprises distinct software modules, each of the distinct software modules being embodied on a computer-readable storage medium, and wherein the distinct software modules comprise a raw data filtering and aggregating module, a model generation module, and a demand prediction module;

wherein;

said extracting is carried out by said raw data filtering and aggregating module executing on at least one hardware processor;

said generating is carried out by said model generation module executing on said at least one hardware processor; and

said applying is carried out by said demand prediction module executing on said at least one hardware processor.

11. A computer program product comprising a computer readable storage medium having computer readable program code embodied therewith, said computer readable program code comprising:

computer readable program code configured to extract high level characteristics from a virtual machine provisioning request stream for a cloud computing environment;

computer readable program code configured to generate a plurality of prediction models for a plurality of different types of virtual machines; and

computer readable program code configured to apply a regression-correlation ensemble to appropriate ones of said prediction models to predict future virtual machine demand for said cloud computing environment.

12. The computer program product of claim 11, further comprising computer readable program code configured to pre-provision a plurality of virtual machines in said cloud computing requirement in accordance with said predicted future virtual machine demand.

13. The computer program product of claim 12, wherein said computer readable program code configured to extract comprises computer readable program code configured to filter irrelevant features and records and transform said high level characteristics into time series.

14. The computer program product of claim 13, wherein said computer readable program code configured to filter filters out characteristics other than virtual machine type and request time.

15. An apparatus comprising:

a memory; and

at least one processor, coupled to said memory, and operative to;

extract high level characteristics from a virtual machine provisioning request stream for a cloud computing environment;

generate a plurality of prediction models for a plurality of different types of virtual machines; and

apply a regression-correlation ensemble to appropriate ones of said prediction models to predict future virtual machine demand for said cloud computing environment.

16. The apparatus of claim 15, wherein said at least one processor is further operative to pre-provision a plurality of virtual machines in said cloud computing requirement in accordance with said predicted future virtual machine demand.

17. The apparatus of claim 16, wherein said at least one processor is operative to extract said high level characteristics by filtering irrelevant features and records and transforming said high level characteristics into time series.

18. The apparatus of claim 17, wherein said at least one processor is operative to filter by filtering out characteristics other than virtual machine type and request time.

19. The apparatus of claim 17, wherein said at least one processor is operative to generate said plurality of prediction models based on an asymmetric cost measurement that treats under-prediction and over-prediction costs differently.

20. The apparatus of claim 19, wherein said at least one processor is operative to generate said plurality of prediction models based on cloud prediction cost.

21. The apparatus of claim 17, wherein said at least one processor is further operative to repeat said extracting and generating as said virtual machine provisioning request stream moves forward in time.

22. The apparatus of claim 17, wherein said at east one processor is operative to apply said regression correlation ensemble by:

combining a plurality of said prediction models via a weighted linear combination to obtain an intermediate prediction; and

post-processing said intermediate prediction via correlation between positively correlated ones of said time series to obtain said predicted future virtual machine demand.

23. The apparatus of claim 17, wherein said at least one processor is further operative to reserve idle ones of said pre-provisioned virtual machines in said cloud computing environment in accordance with said predicted future virtual machine demand.

24. The apparatus of claim 15, further comprising a plurality of distinct software modules, each of the distinct software modules being embodied on a computer-readable storage medium, and wherein the distinct software modules comprise a raw data filtering and aggregating module, a model generation module, and a demand prediction module;

wherein;

said at least one processor is operative to extract by executing said raw data filtering and aggregating module;

said at least one processor is operative to generate by executing said model generation module; and

said at least one processor is operative to apply by executing said demand prediction module.

25. An apparatus comprising:

means for extracting high level characteristics from a virtual machine provisioning request stream for a cloud computing environment;

means for generating a plurality of prediction models for a plurality of different types of virtual machines; and

means for applying a regression-correlation ensemble to appropriate ones of said prediction models to predict future virtual machine demand for said cloud computing environment.