CN105975385A - Fuzzy neural network-based virtual machine energy consumption prediction method and system - Google Patents
Fuzzy neural network-based virtual machine energy consumption prediction method and system Download PDFInfo
- Publication number
- CN105975385A CN105975385A CN201610279167.5A CN201610279167A CN105975385A CN 105975385 A CN105975385 A CN 105975385A CN 201610279167 A CN201610279167 A CN 201610279167A CN 105975385 A CN105975385 A CN 105975385A
- Authority
- CN
- China
- Prior art keywords
- virtual machine
- energy consumption
- neural network
- fuzzy
- training
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
- G06F11/3452—Performance evaluation by statistical analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/3003—Monitoring arrangements specially adapted to the computing system or computing system component being monitored
- G06F11/301—Monitoring arrangements specially adapted to the computing system or computing system component being monitored where the computing system is a virtual computing platform, e.g. logically partitioned systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/3058—Monitoring arrangements for monitoring environmental properties or parameters of the computing system or of the computing system component, e.g. monitoring of power, currents, temperature, humidity, position, vibrations
- G06F11/3062—Monitoring arrangements for monitoring environmental properties or parameters of the computing system or of the computing system component, e.g. monitoring of power, currents, temperature, humidity, position, vibrations where the monitored property is the power consumption
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N7/00—Computing arrangements based on specific mathematical models
- G06N7/02—Computing arrangements based on specific mathematical models using fuzzy logic
- G06N7/04—Physical realisation
- G06N7/046—Implementation by means of a neural network
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Quality & Reliability (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Automation & Control Theory (AREA)
- Biomedical Technology (AREA)
- Fuzzy Systems (AREA)
- General Health & Medical Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Algebra (AREA)
- Artificial Intelligence (AREA)
- Computational Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a fuzzy neural network-based virtual machine energy consumption prediction method and system. The method comprises the following steps: previously carrying out model training on historical running parameters of virtual machines and historical running parameters of physical machines matched with the virtual machines on the basis of a fuzzy neural network so as to obtain a virtual machine energy consumption prediction model; and inputting running parameters of a to-be-predicted virtual machine and running parameters of a physical machine matched with the to-be-predicted virtual machine into the virtual machine energy consumption prediction model so as to obtain an energy consumption prediction value of the to-be-predicted virtual machine. According to the fuzzy neural network-based virtual machine energy consumption prediction method and system, the prediction model is constructed on the basis of the fuzzy neural network, so that the errors caused by simulation between hard threshold partitions can be avoided; and meanwhile, the neural network has a favorable approximation ability for nonlinear functions, so that the nonlinear relationship between the virtual machine energy consumption and related parameters can be better simulated. Moreover, model training is carried out on the historical running parameters of the virtual machines and physical machines, so that the prediction model can be used for predicting the energy consumption more comprehensively and correctly.
Description
Technical field
The present invention relates to virtual machine technique field, particularly to a kind of virtual machine based on fuzzy neural network
Energy consumption Forecasting Methodology and system.
Background technology
Currently, along with the Internet and the development of cloud computing, virtual machine technique has obtained the most widely should
With.And the prediction to Virtual Machine Worker energy consumption, not only concern cost and the price of cloud computing, but also directly
Connect scheduling and the management that have impact on virtual machine.
Existing conventional energy consumption of virtual machine Forecasting Methodology can be divided into linear prediction and nonlinear prediction.Wherein,
Linear prediction method is based on Mind on statistics, uses multiple linear regression mould based on least square
Intend, be not suitable for the complex nonlinear prediction of multi-dummy machine, cause forecast error bigger.And existing
In Non-linear, it is only to carry out hard-threshold segmentation according to virtual machine cpu busy percentage simply
Simulation, relatively low to the energy consumption predictablity rate of virtual machine under the complex environment of multi-dummy machine.
In sum it can be seen that how to improve energy consumption of virtual machine prediction accuracy be have at present to be solved
Problem.
Summary of the invention
In view of this, it is an object of the invention to provide a kind of energy consumption of virtual machine based on fuzzy neural network
Forecasting Methodology and system, improve the accuracy of energy consumption of virtual machine prediction.Its concrete scheme is as follows:
A kind of energy consumption of virtual machine Forecasting Methodology based on fuzzy neural network, including:
It is in advance based on fuzzy neural network, to the historical operating parameter of virtual machine and joins with described virtual machine
The historical operating parameter of the physical machine of set carries out model training, obtains energy consumption of virtual machine forecast model；
By operational factor and the fortune of the physical machine supporting with described virtual machine to be predicted of virtual machine to be predicted
Line parameter inputs described energy consumption of virtual machine forecast model, obtains the energy consumption predictive value of described virtual machine to be predicted.
Preferably, described in be in advance based on fuzzy neural network, to the historical operating parameter of virtual machine and with
The historical operating parameter of the physical machine that described virtual machine is supporting carries out model training, obtains energy consumption of virtual machine pre-
Survey the process of model, including:
Obtain historical operating parameter and the history of the physical machine supporting with described virtual machine of described virtual machine
Operational factor, obtains initial training parameter set；
Described initial training parameter set is carried out choice of parameters, using the parameter that filters out as eigenvalue,
To corresponding feature set；
Utilize the membership function that training in advance is good, eigenvalue all of in described feature set is carried out obfuscation
Process, obtain corresponding fuzzy set；
Utilize described fuzzy set to carry out the Weight Training of BP neutral net, obtain corresponding Fuzzy BP nerve net
Network, and described fuzzy BP neural network is defined as described energy consumption of virtual machine forecast model.
Preferably, the historical operating parameter of described virtual machine and going through of the physical machine supporting with described virtual machine
History operational factor the most at least includes cpu busy percentage, memory usage, execution instruction number in the unit interval
With the cache number lost in the unit interval.
Preferably, described described initial training parameter set being carried out choice of parameters, the parameter that will filter out is made
It is characterized value, obtains the process of corresponding feature set, including:
Based on principal component analysis, described initial training parameter set is carried out dimension-reduction treatment, with to described initially
Training parameter collection carries out choice of parameters, and using the parameter that filters out as eigenvalue, is correspondingly made available described
Feature set.
Preferably, described membership function is Gaussian membership function based on cluster.
Preferably, the training process of described Gaussian membership function, including:
Respectively each eigenvalue in described feature set is clustered, and calculate each cluster correspondence
Average and variance, generate and obtain Gaussian membership function.
Preferably, the process of the described Weight Training utilizing described fuzzy set to carry out BP neutral net, including:
Step S71: the weights of BP neutral net are initialized；
Step S72: obtain current BP neutral net, and each fuzzy parameter in described fuzzy set is inputted
Current BP neutral net, correspondingly to export energy consumption of virtual machine estimated value；
Step S73: calculate the difference between described energy consumption of virtual machine estimated value and virtual machine actual consumption value；
Step S74: judge whether this difference is less than predetermined threshold value, if it is, terminate training, if it does not,
Then according to this difference, the weights of BP neutral net are revised accordingly, so that BP neutral net is carried out
Update, and enter step S72.
The invention also discloses a kind of energy consumption of virtual machine prognoses system based on fuzzy neural network, including:
Model training module, is used for being in advance based on fuzzy neural network, the historical operating parameter to virtual machine
And the historical operating parameter of the physical machine supporting with described virtual machine carries out model training, obtain virtual machine
Energy consumption forecast model；
Energy consumption prediction module, for by the operational factor of virtual machine to be predicted and to be predicted virtual with described
The operational factor of the physical machine that machine is supporting inputs described energy consumption of virtual machine forecast model, obtains described to be predicted
The energy consumption predictive value of virtual machine.
Preferably, described model training module includes:
Training parameter acquiring unit, for obtain described virtual machine historical operating parameter and with described void
The historical operating parameter of the physical machine that plan machine is supporting, obtains initial training parameter set；
Choice of parameters device, for described initial training parameter set is carried out choice of parameters, the ginseng that will filter out
Number, as eigenvalue, obtains corresponding feature set；
Fuzzy Processing unit, for utilizing the membership function that training in advance is good, to all in described feature set
Eigenvalue carry out Fuzzy processing, obtain corresponding fuzzy set；
Neural network weight training aids, for utilizing described fuzzy set to carry out the Weight Training of BP neutral net,
Obtain corresponding fuzzy BP neural network, and described fuzzy BP neural network is defined as described virtual function
Consumption forecast model.
Preferably, described neural network weight training aids includes:
Weight initialization unit, for initializing the weights of BP neutral net；
Estimation of energy consumption unit, for obtaining current BP neutral net, and by mould each in described fuzzy set
Stick with paste the BP neutral net that parameter input is current, correspondingly to export energy consumption of virtual machine estimated value；
Difference computational unit, be used for calculating described energy consumption of virtual machine estimated value and virtual machine actual consumption value it
Between difference；
Judging unit, is used for judging whether this difference is less than predetermined threshold value, if it is, terminate training,
If it is not, then according to this difference, the weights of BP neutral net are revised accordingly, with neural to BP
Network is updated, and the BP neutral net after updating sends to described Estimation of energy consumption unit.
In the present invention, energy consumption of virtual machine Forecasting Methodology includes: be in advance based on fuzzy neural network, to virtual
The historical operating parameter of machine and the historical operating parameter of the physical machine supporting with virtual machine carry out model instruction
Practice, obtain energy consumption of virtual machine forecast model；Then by the operational factor of virtual machine to be predicted and with treat pre-
Survey the operational factor input energy consumption of virtual machine forecast model of the supporting physical machine of virtual machine, obtain void to be predicted
The energy consumption predictive value of plan machine.Visible, the present invention is that to build energy consumption of virtual machine based on fuzzy neural network pre-
Survey model, owing to fuzzy neural network can carry out fuzzy representation to parameter, it is possible to avoid existing
There is simulation produced error in hard-threshold by stages in technology, be conducive to improving energy consumption forecasting accuracy；With
Time there is the good approximation ability to nonlinear function due to neutral net such that it is able to preferably simulate void
Intend the non-linear relation between function consumption and each relevant parameter, improve energy consumption forecasting accuracy.It addition,
The present invention is the historical operating parameter of virtual machine and physical machine to be carried out model training simultaneously, and this shows this
Invention considers not only the impact on energy consumption of virtual machine of the historical operating parameter of virtual machine, it is also contemplated that relevant
The impact on energy consumption of virtual machine of the historical operating parameter of physical machine, so that the virtual function finally given
The energy consumption of virtual machine can be predicted by consumption forecast model more all-sidedly and accurately.As fully visible, this
Bright disclosed technical scheme is effectively improved the accuracy of energy consumption of virtual machine prediction.
Accompanying drawing explanation
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to reality
Execute the required accompanying drawing used in example or description of the prior art to be briefly described, it should be apparent that below,
Accompanying drawing in description is only embodiments of the invention, for those of ordinary skill in the art, not
On the premise of paying creative work, it is also possible to obtain other accompanying drawing according to the accompanying drawing provided.
Fig. 1 is a kind of energy consumption of virtual machine prediction side based on fuzzy neural network disclosed in the embodiment of the present invention
Method flow chart；
Fig. 2 is that energy consumption of virtual machine forecast model disclosed in the embodiment of the present invention builds flow chart；
Fig. 3 is the Weight Training flow chart of BP neutral net disclosed in the embodiment of the present invention；
Fig. 4 is a kind of energy consumption of virtual machine based on fuzzy neural network prediction system disclosed in the embodiment of the present invention
System structural representation.
Detailed description of the invention
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out
Clearly and completely describe, it is clear that described embodiment is only a part of embodiment of the present invention, and
It is not all, of embodiment.Based on the embodiment in the present invention, those of ordinary skill in the art are not doing
Go out the every other embodiment obtained under creative work premise, broadly fall into the scope of protection of the invention.
The embodiment of the invention discloses a kind of energy consumption of virtual machine Forecasting Methodology based on fuzzy neural network, ginseng
As shown in Figure 1, the method includes:
Step S11: be in advance based on fuzzy neural network, to the historical operating parameter of virtual machine and with virtual
The historical operating parameter of the physical machine that machine is supporting carries out model training, obtains energy consumption of virtual machine forecast model；
Step S12: by operational factor and the physical machine supporting with virtual machine to be predicted of virtual machine to be predicted
Operational factor input energy consumption of virtual machine forecast model, obtain the energy consumption predictive value of virtual machine to be predicted.
It should be noted that the fuzzy neural network in the present embodiment can select fuzzy radially base nerve net
Network, it would however also be possible to employ fuzzy BP neural network (BP, i.e. Back Propagation, back propagation).
It addition, the present embodiment is when carrying out model training based on fuzzy neural network, it is not only to void
The historical operating parameter of plan machine carries out model training, the most also the historical operating parameter of physical machine is carried out mould
Type training, so makes the embodiment of the present invention consider not only the historical operating parameter of virtual machine to virtual machine
The impact of energy consumption, it is also contemplated that the historical operating parameter of the related physical machine impact on energy consumption of virtual machine, from
And make the energy consumption of virtual machine forecast model finally given can energy consumption to virtual machine more all-sidedly and accurately
It is predicted.
After utilizing above-mentioned model training method to train energy consumption of virtual machine forecast model, when needs are treated pre-
The energy consumption surveying virtual machine is predicted, then only need to by operational factor current for virtual machine to be predicted and with this
The operational factor input above-mentioned energy consumption of virtual machine prediction mould that the matching used physical machine of virtual machine to be predicted is current
In type, just can obtain the energy consumption predictive value of this virtual machine to be predicted.
Visible, the embodiment of the present invention builds energy consumption of virtual machine forecast model based on fuzzy neural network,
Owing to fuzzy neural network can carry out fuzzy representation to parameter, it is possible to avoid in prior art hard
The produced error of threshold value by stages simulation, is conducive to improving energy consumption forecasting accuracy；Simultaneously because it is neural
Network has the good approximation ability to nonlinear function such that it is able to preferably simulation energy consumption of virtual machine with
Non-linear relation between each relevant parameter, improves energy consumption forecasting accuracy.It addition, the present invention implements
Example is the historical operating parameter of virtual machine and physical machine to be carried out model training simultaneously, and this shows the present invention
Embodiment considers not only the historical operating parameter impact on energy consumption of virtual machine of virtual machine, it is also contemplated that phase
The impact on energy consumption of virtual machine of the historical operating parameter of pass physical machine, so that the virtual machine finally given
The energy consumption of virtual machine can be predicted by energy consumption forecast model more all-sidedly and accurately.As fully visible, originally
Technical scheme disclosed in inventive embodiments is effectively improved the accuracy of energy consumption of virtual machine prediction.
The embodiment of the invention discloses a kind of concrete energy consumption of virtual machine prediction side based on fuzzy neural network
Method, relative to a upper embodiment, technical scheme has been made further instruction and optimization by the present embodiment.Tool
Body:
Shown in Figure 2, in upper embodiment step S11, it is in advance based on fuzzy neural network, to virtual
The historical operating parameter of machine and the historical operating parameter of the physical machine supporting with virtual machine carry out model instruction
Practice, obtain the process of energy consumption of virtual machine forecast model, specifically may include that
Step S111: obtain the historical operating parameter of virtual machine and going through of the physical machine supporting with virtual machine
History operational factor, obtains initial training parameter set；
Step S112: initial training parameter set is carried out choice of parameters, using the parameter that filters out as feature
Value, obtains corresponding feature set；
Step S113: utilize the membership function that training in advance is good, is carried out eigenvalue all of in feature set
Fuzzy processing, obtains corresponding fuzzy set；
Step S114: utilize fuzzy set to carry out the Weight Training of BP neutral net, obtain corresponding Fuzzy BP
Neutral net, and fuzzy BP neural network is defined as energy consumption of virtual machine forecast model.
It is understood that the historical operating parameter of above-mentioned virtual machine and the physical machine supporting with virtual machine
Historical operating parameter the most at least includes cpu busy percentage, memory usage, performs instruction in the unit interval
The cache number (cache, i.e. cache memory) lost in number and unit interval.
It addition, in above-mentioned steps S112, initial training parameter set is carried out choice of parameters, by filter out
Parameter, as eigenvalue, obtains the process of corresponding feature set, specifically may include that and divides based on main constituent
Analysis, carries out dimension-reduction treatment to initial training parameter set, so that initial training parameter set is carried out choice of parameters,
And using the parameter that filters out as eigenvalue, it is correspondingly made available feature set.Certainly, the present embodiment can also
Use correlation analysis, above-mentioned initial training parameter is carried out choice of parameters.
Membership function in above-mentioned steps S113 can be Gaussian membership function based on cluster
Wherein, the training process of Gaussian membership function, including: respectively to each feature in feature set
Value clusters, and calculates average corresponding to each cluster and variance, and generation obtains Gaussian and is subordinate to letter
Number.Wherein, the expression formula of Gaussian membership function is:
Wherein, xiRepresent ith feature value, cijRepresent the jth cluster of ith feature value in feature set
Average,Represent the variance of the jth cluster of ith feature value in feature set.It should be noted that one
The cluster number that individual eigenvalue is corresponding can be 1, it is also possible to more than 1.
Certainly, the present embodiment is except using Gaussian membership function, it would however also be possible to employ be equally based on cluster
Triangular membership.
It addition, shown in Figure 3, in above-mentioned steps S114, utilize fuzzy set to carry out the power of BP neutral net
The process of value training, specifically includes:
Step S1141: the weights of BP neutral net are initialized；
Step S1142: obtain current BP neutral net, and each fuzzy parameter input in fuzzy set is worked as
Front BP neutral net, correspondingly to export energy consumption of virtual machine estimated value；
Step S1143: calculate the difference between energy consumption of virtual machine estimated value and virtual machine actual consumption value；
Step S1144: judge whether this difference is less than predetermined threshold value, if it is, terminate training, if
No, then according to this difference, the weights of BP neutral net are revised accordingly, with to BP neutral net
It is updated, and enters step S1142.
Visible, the present embodiment, when carrying out the Weight Training of BP neutral net, is by the most virtual
Difference between function consumption estimated value and virtual machine actual consumption value, if this difference is more than or equal to presetting
Threshold value, then revise accordingly to the weights of BP neutral net, otherwise it is assumed that the BP now training out
Neutral net coincidence loss requirement, deconditioning just may be used.
Accordingly, the embodiment of the invention discloses the prediction of a kind of energy consumption of virtual machine based on fuzzy neural network
System, shown in Figure 4, this system includes:
Model training module 41, is used for being in advance based on fuzzy neural network, joins the history run of virtual machine
The historical operating parameter of number and the physical machine supporting with virtual machine carries out model training, obtains virtual function
Consumption forecast model；
Energy consumption prediction module 42, for by the operational factor of virtual machine to be predicted and with virtual machine to be predicted
The operational factor input energy consumption of virtual machine forecast model of supporting physical machine, obtains the energy of virtual machine to be predicted
Consumption predictive value.
Concrete, above-mentioned model training module can include training parameter acquiring unit, choice of parameters device,
Fuzzy Processing unit and neural network weight training aids；Wherein,
Training parameter acquiring unit, for obtaining the historical operating parameter of virtual machine and supporting with virtual machine
The historical operating parameter of physical machine, obtain initial training parameter set；
Choice of parameters device, for initial training parameter set is carried out choice of parameters, makees the parameter filtered out
It is characterized value, obtains corresponding feature set；
Fuzzy Processing unit, for utilizing the membership function that training in advance is good, to spy all of in feature set
Value indicative carries out Fuzzy processing, obtains corresponding fuzzy set；
Neural network weight training aids, for utilizing fuzzy set to carry out the Weight Training of BP neutral net,
To corresponding fuzzy BP neural network, and fuzzy BP neural network is defined as energy consumption of virtual machine prediction mould
Type.
It addition, the neural network weight training aids in the present embodiment specifically can include weight initialization unit,
Estimation of energy consumption unit, difference computational unit and judging unit.Wherein,
Weight initialization unit, for initializing the weights of BP neutral net；
Estimation of energy consumption unit, for obtaining current BP neutral net, and by fuzzy ginseng each in fuzzy set
The BP neutral net that number input is current, correspondingly to export energy consumption of virtual machine estimated value；
Difference computational unit, for calculating between energy consumption of virtual machine estimated value and virtual machine actual consumption value
Difference；
Judging unit, is used for judging whether this difference is less than predetermined threshold value, if it is, terminate training,
If it is not, then according to this difference, the weights of BP neutral net are revised accordingly, with neural to BP
Network is updated, and will update after BP neutral net send to Estimation of energy consumption unit.
It is understood that the BP neutral net after above-mentioned judging unit will update sends to Estimation of energy consumption list
During unit, also it is notified that Estimation of energy consumption unit utilizes the BP neutral net after updating to carry out Estimation of energy consumption.
Visible, the embodiment of the present invention builds energy consumption of virtual machine forecast model based on fuzzy neural network,
Owing to fuzzy neural network can carry out fuzzy representation to parameter, it is possible to avoid in prior art hard
The produced error of threshold value by stages simulation, is conducive to improving energy consumption forecasting accuracy；Simultaneously because it is neural
Network has the good approximation ability to nonlinear function such that it is able to preferably simulation energy consumption of virtual machine with
Non-linear relation between each relevant parameter, improves energy consumption forecasting accuracy.It addition, the present invention implements
Example is the historical operating parameter of virtual machine and physical machine to be carried out model training simultaneously, and this shows the present invention
Embodiment considers not only the historical operating parameter impact on energy consumption of virtual machine of virtual machine, it is also contemplated that phase
The impact on energy consumption of virtual machine of the historical operating parameter of pass physical machine, so that the virtual machine finally given
The energy consumption of virtual machine can be predicted by energy consumption forecast model more all-sidedly and accurately.As fully visible, originally
Technical scheme disclosed in inventive embodiments is effectively improved the accuracy of energy consumption of virtual machine prediction.
Finally, in addition it is also necessary to explanation, in this article, term " include ", " comprising " or its
What his variant is intended to comprising of nonexcludability, so that include the process of a series of key element, side
Method, article or equipment not only include those key elements, but also include other key elements being not expressly set out,
Or also include the key element intrinsic for this process, method, article or equipment.The most more
In the case of restriction, statement " including ... " key element limited, it is not excluded that described in including
The process of key element, method, article or equipment there is also other identical element.
Above to a kind of energy consumption of virtual machine Forecasting Methodology based on fuzzy neural network provided by the present invention and
System is described in detail, and principle and the embodiment of the present invention are entered by specific case used herein
Having gone elaboration, the explanation of above example is only intended to help to understand method and the core concept thereof of the present invention；
Simultaneously for one of ordinary skill in the art, according to the thought of the present invention, in detailed description of the invention and
All will change in range of application, in sum, this specification content should not be construed as the present invention
Restriction.
Claims (10)
1. an energy consumption of virtual machine Forecasting Methodology based on fuzzy neural network, it is characterised in that including:
It is in advance based on fuzzy neural network, to the historical operating parameter of virtual machine and joins with described virtual machine
The historical operating parameter of the physical machine of set carries out model training, obtains energy consumption of virtual machine forecast model；
By operational factor and the fortune of the physical machine supporting with described virtual machine to be predicted of virtual machine to be predicted
Line parameter inputs described energy consumption of virtual machine forecast model, obtains the energy consumption predictive value of described virtual machine to be predicted.
Energy consumption of virtual machine Forecasting Methodology based on fuzzy neural network the most according to claim 1, its
Be characterised by, described in be in advance based on fuzzy neural network, to the historical operating parameter of virtual machine and with institute
The historical operating parameter stating the supporting physical machine of virtual machine carries out model training, obtains energy consumption of virtual machine prediction
The process of model, including:
Obtain historical operating parameter and the history of the physical machine supporting with described virtual machine of described virtual machine
Operational factor, obtains initial training parameter set；
Described initial training parameter set is carried out choice of parameters, using the parameter that filters out as eigenvalue,
To corresponding feature set；
Utilize the membership function that training in advance is good, eigenvalue all of in described feature set is carried out obfuscation
Process, obtain corresponding fuzzy set；
Utilize described fuzzy set to carry out the Weight Training of BP neutral net, obtain corresponding Fuzzy BP neural
Network, and described fuzzy BP neural network is defined as described energy consumption of virtual machine forecast model.
Energy consumption of virtual machine Forecasting Methodology based on fuzzy neural network the most according to claim 2, its
It is characterised by, the historical operating parameter of described virtual machine and the history of the physical machine supporting with described virtual machine
Operational factor the most at least include cpu busy percentage, memory usage, execution instruction number in the unit interval and
The cache number lost in unit interval.
Energy consumption of virtual machine Forecasting Methodology based on fuzzy neural network the most according to claim 2, its
Be characterised by, described described initial training parameter set carried out choice of parameters, using the parameter that filters out as
Eigenvalue, obtains the process of corresponding feature set, including:
Based on principal component analysis, described initial training parameter set is carried out dimension-reduction treatment, with to described initially
Training parameter collection carries out choice of parameters, and using the parameter that filters out as eigenvalue, is correspondingly made available described
Feature set.
Energy consumption of virtual machine Forecasting Methodology based on fuzzy neural network the most according to claim 2, its
Being characterised by, described membership function is Gaussian membership function based on cluster.
Energy consumption of virtual machine Forecasting Methodology based on fuzzy neural network the most according to claim 5, its
It is characterised by, the training process of described Gaussian membership function, including:
Respectively each eigenvalue in described feature set is clustered, and calculate each cluster correspondence
Average and variance, generate and obtain Gaussian membership function.
Energy consumption of virtual machine Forecasting Methodology based on fuzzy neural network the most according to claim 2, its
It is characterised by, the process of the described Weight Training utilizing described fuzzy set to carry out BP neutral net, including:
Step S71: the weights of BP neutral net are initialized；
Step S72: obtain current BP neutral net, and by defeated for fuzzy parameter each in described fuzzy set
Enter current BP neutral net, correspondingly to export energy consumption of virtual machine estimated value；
Step S73: calculate the difference between described energy consumption of virtual machine estimated value and virtual machine actual consumption value；
Step S74: judge whether this difference is less than predetermined threshold value, if it is, terminate training, if it does not,
Then according to this difference, the weights of BP neutral net are revised accordingly, so that BP neutral net is entered
Row updates, and enters step S72.
8. an energy consumption of virtual machine prognoses system based on fuzzy neural network, it is characterised in that including:
Model training module, is used for being in advance based on fuzzy neural network, the historical operating parameter to virtual machine
And the historical operating parameter of the physical machine supporting with described virtual machine carries out model training, obtain virtual machine
Energy consumption forecast model；
Energy consumption prediction module, for by the operational factor of virtual machine to be predicted and to be predicted virtual with described
The operational factor of the physical machine that machine is supporting inputs described energy consumption of virtual machine forecast model, obtains described to be predicted
The energy consumption predictive value of virtual machine.
Energy consumption of virtual machine prognoses system based on fuzzy neural network the most according to claim 8, its
Being characterised by, described model training module includes:
Training parameter acquiring unit, for obtain described virtual machine historical operating parameter and with described void
The historical operating parameter of the physical machine that plan machine is supporting, obtains initial training parameter set；
Choice of parameters device, for described initial training parameter set is carried out choice of parameters, the ginseng that will filter out
Number, as eigenvalue, obtains corresponding feature set；
Fuzzy Processing unit, for utilizing the membership function that training in advance is good, to all in described feature set
Eigenvalue carry out Fuzzy processing, obtain corresponding fuzzy set；
Neural network weight training aids, for utilizing described fuzzy set to carry out the weights instruction of BP neutral net
Practice, obtain corresponding fuzzy BP neural network, and described fuzzy BP neural network is defined as described void
Intend function consumption forecast model.
Energy consumption of virtual machine prognoses system based on fuzzy neural network the most according to claim 9,
It is characterized in that, described neural network weight training aids includes:
Weight initialization unit, for initializing the weights of BP neutral net；
Estimation of energy consumption unit, for obtaining current BP neutral net, and by mould each in described fuzzy set
Stick with paste the BP neutral net that parameter input is current, correspondingly to export energy consumption of virtual machine estimated value；
Difference computational unit, be used for calculating described energy consumption of virtual machine estimated value and virtual machine actual consumption value it
Between difference；
Judging unit, is used for judging whether this difference is less than predetermined threshold value, if it is, terminate training,
If it is not, then according to this difference, the weights of BP neutral net are revised accordingly, with to BP god
It is updated through network, and the BP neutral net after updating sends to described Estimation of energy consumption unit.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610279167.5A CN105975385A (en) | 2016-04-28 | 2016-04-28 | Fuzzy neural network-based virtual machine energy consumption prediction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610279167.5A CN105975385A (en) | 2016-04-28 | 2016-04-28 | Fuzzy neural network-based virtual machine energy consumption prediction method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105975385A true CN105975385A (en) | 2016-09-28 |
Family
ID=56994804
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610279167.5A Pending CN105975385A (en) | 2016-04-28 | 2016-04-28 | Fuzzy neural network-based virtual machine energy consumption prediction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105975385A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107038064A (en) * | 2017-04-18 | 2017-08-11 | 腾讯科技（深圳）有限公司 | Virtual machine management method and device, storage medium |
| CN107401816A (en) * | 2017-08-07 | 2017-11-28 | 珠海格力电器股份有限公司 | The determination method and apparatus of air conditioning energy consumption |
| CN108509324A (en) * | 2017-02-28 | 2018-09-07 | 通用汽车环球科技运作有限责任公司 | The system and method for selecting computing platform |
| CN109165443A (en) * | 2018-08-23 | 2019-01-08 | 珠海格力电器股份有限公司 | A kind of electric appliance Calculation Method of Energy Consumption and system |
| CN110955320A (en) * | 2018-09-27 | 2020-04-03 | 阿里巴巴集团控股有限公司 | Rack power consumption management equipment, system and method |
| CN111830350A (en) * | 2020-07-23 | 2020-10-27 | 珠海格力电器股份有限公司 | Energy consumption metering method and device and electric appliance |
| CN112150631A (en) * | 2020-09-23 | 2020-12-29 | 浙江大学 | Real-time energy consumption optimization drawing method and device based on neural network |
| CN112308734A (en) * | 2020-10-27 | 2021-02-02 | 中国科学院信息工程研究所 | Method for metering and allocating non-IT energy consumption of IT equipment and electronic device |
| CN113712525A (en) * | 2020-05-21 | 2021-11-30 | 深圳市理邦精密仪器股份有限公司 | Physiological parameter processing method and device and medical equipment |
| CN115202829A (en) * | 2022-08-15 | 2022-10-18 | 中国电信股份有限公司 | Power consumption prediction model training method, power consumption prediction method and device for virtual machine |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101907917A (en) * | 2010-07-21 | 2010-12-08 | 中国电信股份有限公司 | Method and system for measuring energy consumption of virtual machine |
| WO2011096859A1 (en) * | 2010-02-04 | 2011-08-11 | Telefonaktiebolaget L M Ericsson (Publ) | Network performance monitor for virtual machines |
| CN102854968A (en) * | 2012-05-04 | 2013-01-02 | 北京邮电大学 | Real-time energy consumption metering method of virtual machine |
| US20140191472A1 (en) * | 2003-08-19 | 2014-07-10 | Hasbro, Inc. | Action Figure Game Piece and Method of Playing Action Figure Game |
| CN104239982A (en) * | 2014-10-12 | 2014-12-24 | 刘岩 | Method for predicting energy consumption of buildings during holidays and festivals on basis of time series and neural networks |
| CN104298339A (en) * | 2014-10-11 | 2015-01-21 | 东北大学 | Server integration method oriented to minimum energy consumption |
| CN104636493A (en) * | 2015-03-04 | 2015-05-20 | 浪潮电子信息产业股份有限公司 | Dynamic data grading method based on multi-classifier fusion |
-
2016
- 2016-04-28 CN CN201610279167.5A patent/CN105975385A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140191472A1 (en) * | 2003-08-19 | 2014-07-10 | Hasbro, Inc. | Action Figure Game Piece and Method of Playing Action Figure Game |
| WO2011096859A1 (en) * | 2010-02-04 | 2011-08-11 | Telefonaktiebolaget L M Ericsson (Publ) | Network performance monitor for virtual machines |
| CN101907917A (en) * | 2010-07-21 | 2010-12-08 | 中国电信股份有限公司 | Method and system for measuring energy consumption of virtual machine |
| CN102854968A (en) * | 2012-05-04 | 2013-01-02 | 北京邮电大学 | Real-time energy consumption metering method of virtual machine |
| CN104298339A (en) * | 2014-10-11 | 2015-01-21 | 东北大学 | Server integration method oriented to minimum energy consumption |
| CN104239982A (en) * | 2014-10-12 | 2014-12-24 | 刘岩 | Method for predicting energy consumption of buildings during holidays and festivals on basis of time series and neural networks |
| CN104636493A (en) * | 2015-03-04 | 2015-05-20 | 浪潮电子信息产业股份有限公司 | Dynamic data grading method based on multi-classifier fusion |
Non-Patent Citations (2)
| Title |
|---|
| 徐浩: ""基于神经网络的虚拟机能耗预测模型研究"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 谭学奇等: ""基于模糊聚类和模糊神经网络的电价经济特性与预测研究"", 《现代企业文化》 * |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108509324A (en) * | 2017-02-28 | 2018-09-07 | 通用汽车环球科技运作有限责任公司 | The system and method for selecting computing platform |
| US11023351B2 (en) | 2017-02-28 | 2021-06-01 | GM Global Technology Operations LLC | System and method of selecting a computational platform |
| CN108509324B (en) * | 2017-02-28 | 2021-09-07 | 通用汽车环球科技运作有限责任公司 | System and method for selecting computing platform |
| CN107038064A (en) * | 2017-04-18 | 2017-08-11 | 腾讯科技（深圳）有限公司 | Virtual machine management method and device, storage medium |
| CN107401816A (en) * | 2017-08-07 | 2017-11-28 | 珠海格力电器股份有限公司 | The determination method and apparatus of air conditioning energy consumption |
| CN107401816B (en) * | 2017-08-07 | 2019-07-09 | 珠海格力电器股份有限公司 | The determination method and apparatus of air conditioning energy consumption |
| CN109165443A (en) * | 2018-08-23 | 2019-01-08 | 珠海格力电器股份有限公司 | A kind of electric appliance Calculation Method of Energy Consumption and system |
| CN109165443B (en) * | 2018-08-23 | 2021-09-14 | 珠海格力电器股份有限公司 | Electric appliance energy consumption calculation method and system |
| CN110955320A (en) * | 2018-09-27 | 2020-04-03 | 阿里巴巴集团控股有限公司 | Rack power consumption management equipment, system and method |
| CN110955320B (en) * | 2018-09-27 | 2023-04-07 | 阿里巴巴集团控股有限公司 | Rack power consumption management equipment, system and method |
| CN113712525A (en) * | 2020-05-21 | 2021-11-30 | 深圳市理邦精密仪器股份有限公司 | Physiological parameter processing method and device and medical equipment |
| CN111830350A (en) * | 2020-07-23 | 2020-10-27 | 珠海格力电器股份有限公司 | Energy consumption metering method and device and electric appliance |
| CN112150631B (en) * | 2020-09-23 | 2021-09-21 | 浙江大学 | Real-time energy consumption optimization drawing method and device based on neural network |
| CN112150631A (en) * | 2020-09-23 | 2020-12-29 | 浙江大学 | Real-time energy consumption optimization drawing method and device based on neural network |
| CN112308734A (en) * | 2020-10-27 | 2021-02-02 | 中国科学院信息工程研究所 | Method for metering and allocating non-IT energy consumption of IT equipment and electronic device |
| CN112308734B (en) * | 2020-10-27 | 2024-01-05 | 中国科学院信息工程研究所 | IT equipment non-IT energy consumption metering and cost sharing method and electronic device |
| CN115202829A (en) * | 2022-08-15 | 2022-10-18 | 中国电信股份有限公司 | Power consumption prediction model training method, power consumption prediction method and device for virtual machine |
| CN115202829B (en) * | 2022-08-15 | 2023-11-21 | 中国电信股份有限公司 | Power consumption prediction model training method and device of virtual machine and power consumption prediction method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105975385A (en) | Fuzzy neural network-based virtual machine energy consumption prediction method and system | |
| CN110943983B (en) | Network security prevention method based on security situation awareness and risk assessment | |
| CN113656879A (en) | Curtain wall engineering refinement construction method, system, terminal and medium based on BIM | |
| CN104182474B (en) | A kind of pre- recognition methods for being lost in user | |
| CN103365727B (en) | Host load forecasting method in cloud computing environment | |
| CN104408518B (en) | Based on the neural network learning optimization method of particle swarm optimization algorithm | |
| CN109472321A (en) | A kind of prediction towards time series type surface water quality big data and assessment models construction method | |
| CN109191312A (en) | A kind of anti-fraud air control method and device of Claims Resolution | |
| CN105676814A (en) | SFLA-SVM-based digital water island online agent adding control method | |
| CN104063550A (en) | Multi-physics field CAE system based on cloud computing platform | |
| CN102915510A (en) | Power project network post-evaluation system based on multilevel fuzzy integrative evaluation model | |
| CN106296434B (en) | Grain yield prediction method based on PSO-LSSVM algorithm | |
| CN103885867B (en) | Online evaluation method of performance of analog circuit | |
| Li et al. | Improved LSTM-based prediction method for highly variable workload and resources in clouds | |
| CN106447219A (en) | Cloud service provider assessment method for auto-control valve enterprise under cloud manufacturing mode | |
| Shuai et al. | Integrated parallel forecasting model based on modified fuzzy time series and SVM | |
| CN106503889A (en) | A kind of subjective and objective integrated evaluating method of land use intensively and device | |
| CN113822441A (en) | Decision model training method and device, terminal equipment and storage medium | |
| CN109165740A (en) | The fault time calculation method of product subsystem based on section step analysis | |
| CN115102195A (en) | Three-phase load unbalance treatment method and device based on rural power grid | |
| CN109038550A (en) | Electric system self-healing index calculating method based on voltage static stability | |
| CN113987261A (en) | Video recommendation method and system based on dynamic trust perception | |
| CN110262891B (en) | Automatic multifunctional resource recycling system across virtualization platforms | |
| CN117435308B (en) | Modelica model simulation method and system based on parallel computing algorithm | |
| Lian et al. | A discrete particle swarm optimization algorithm for job-shop scheduling problem to maximizing production |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20160928 |