CN106897113A - The method and device of a kind of virtualized host operation conditions prediction - Google Patents
The method and device of a kind of virtualized host operation conditions prediction Download PDFInfo
- Publication number
- CN106897113A CN106897113A CN201710100386.7A CN201710100386A CN106897113A CN 106897113 A CN106897113 A CN 106897113A CN 201710100386 A CN201710100386 A CN 201710100386A CN 106897113 A CN106897113 A CN 106897113A
- Authority
- CN
- China
- Prior art keywords
- value
- main frame
- data
- probable value
- probable
- 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
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
- G06F2009/45591—Monitoring or debugging support
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses the method and device of a kind of prediction of virtualized host operation conditions, by the real-time running data for obtaining main frame;According to above-mentioned real-time running data and predetermined threshold value, corresponding characteristic vector is drawn;Based on Bayesian statistical model and default Prior Probability, features described above vector is calculated, draw to characterize above-mentioned main frame and normally run the first probable value of probability and characterize above-mentioned main frame and break down the second probable value of probability;According to above-mentioned first probable value and above-mentioned second probable value, the operation conditions to above-mentioned main frame is predicted.The real-time running data of Intrusion Detection based on host generates corresponding characteristic vector, Bayesian statistical model is recycled to calculate characteristic vector, judge that the real-time characteristic vector of main frame is belonging to normally to run class and still fall within failure classes, calculate corresponding probable value, predict whether main frame will break down according to the probable value for calculating.It can be seen that, the application can predict whether virtualized host will break down.
Description
Technical field
The present invention relates to virtualize calculating field, more particularly to a kind of method of virtualized host operation conditions prediction with
And device.
Background technology
With the development of network technology and Intel Virtualization Technology, Intel Virtualization Technology is by large-scale application in all trades and professions.
Intel Virtualization Technology is a kind of resource management techniques, and it can be by the various actual resources of computer, such as server, net
Network, internal memory and storage etc., show after giving abstract, conversion, break the not cleavable obstacle between entity structure, make user
These resources can be applied.Intel Virtualization Technology can be created that different virtual machines, can create multiple virtual on same main frame
It is separate between machine, and multiple virtual machine, with play to greatest extent the resource utilization of main frame, availability, security with
And autgmentability.
The stable operation of main frame is the available primary condition of whole virtualized environment under virtualized environment, i.e. main frame carries
Virtual machine runs the demand of resource, and main frame breaks down can cause virtual machine crashes, and then may result in the industry in virtual machine
Business cannot normal process.Failure to virtualized host is predicted, and can well ensure the stable operation of virtualization system.
Therefore it is this area problem demanding prompt solution that failure predication how is carried out to virtualized host.
The content of the invention
It is an object of the invention to provide the method and device of a kind of prediction of virtualized host operation conditions, it is therefore intended that solution
The problem of failure predication cannot be certainly carried out to virtual machine in the prior art.
In order to solve the above technical problems, the present invention provides a kind of method of virtualized host operation conditions prediction, the method
Including:
Obtain the real-time running data of main frame;
According to the real-time running data and predetermined threshold value, corresponding characteristic vector is drawn;
Based on Bayesian statistical model and default Prior Probability, the characteristic vector is calculated, draw sign
The main frame normally runs the first probable value of probability and characterizes the main frame and breaks down the second probable value of probability;
According to first probable value and second probable value, the operation conditions to the main frame is predicted.
Alternatively, it is described according to the real-time running data and predetermined threshold value, show that corresponding characteristic vector includes:
According to the default characteristic item of characteristic vector, corresponding feature item data is chosen from the real-time running data;
Compare the size of each described feature item data and corresponding predetermined threshold value;
When the feature item data is more than the predetermined threshold value, the numerical value of individual features is set to the first present count
Value;
When the feature item data is less than the predetermined threshold value, the numerical value of individual features is set to the second present count
Value;
According to first default value and second default value, the characteristic vector is drawn.
Alternatively, it is described based on Bayesian statistical model and default Prior Probability, the characteristic vector is counted
Calculate, draw to characterize the main frame and normally run the first probable value of probability and characterize the main frame and break down the second general of probability
Rate value includes:
The the first default Prior Probability normally run using the Bayesian statistical model and the sign main frame, to institute
State characteristic vector to be calculated, draw first probable value;
Using the Bayesian statistical model and the second default Prior Probability of the sign hostdown, to the spy
Levy vector to be calculated, draw second probable value.
Alternatively, it is described according to first probable value and second probable value, to the operation conditions of the main frame
Be predicted including:
Compare the size of first probable value and second probable value;
When first probable value is more than second probable value, judge that the main frame is in normal operating condition;
When first probable value is less than second probable value, judge that the main frame will break down.
Alternatively, judge that the main frame will be sent out when first probable value is less than second probable value described
Also include after raw failure:
The information of the main frame is sent to specified receiving terminal, so that the specified receiving terminal is sent out according to described information
Go out early warning information.
Alternatively, also included before the real-time running data of the acquisition main frame:
Each history run daily record data of main frame is obtained, using the history run daily record data as number of training
According to;
The training sample data are pre-processed, the predetermined threshold value is drawn;
Based on the number of training according to this and the corresponding predetermined threshold value, draw corresponding training sample feature to
Amount;
Using the Bayesian statistical model and the sampling feature vectors, the default Prior Probability is calculated.
Alternatively, it is described that the training sample data are pre-processed, show that the predetermined threshold value includes:
The training sample data are divided into normal operation class data and the class of failure classes data two;
According to previously selected training characteristics, calculate respectively in the normal operation class data and the failure classes data
Training characteristics item data average value, using the average value as the predetermined threshold value.
Additionally, present invention also offers a kind of device of virtualized host operation conditions prediction, the device includes:
Acquisition module, the real-time running data for obtaining main frame;
Feature vector generation module, for according to the real-time running data and predetermined threshold value, draw corresponding feature to
Amount;
Probable value computing module, for based on Bayesian statistical model and default Prior Probability, to the feature to
Amount is calculated, and is drawn to characterize the main frame and normally run the first probable value of probability and characterize the main frame and is broken down probability
The second probable value;
Prediction module, for according to first probable value and second probable value, to the operation shape of the main frame
Condition is predicted.
Alternatively, the feature vector generation module includes:
Feature item data chooses unit, for the default characteristic item according to characteristic vector, from the real-time running data
Choose corresponding feature item data;
First comparing unit, the size for comparing each described feature item data and corresponding predetermined threshold value;
First setup unit, for when the feature item data is more than the predetermined threshold value, by the number of individual features
Value is set to the first default value;
Second setup unit, for when the feature item data is less than the predetermined threshold value, by the number of individual features
Value is set to the second default value;
Generation unit, for according to first default value and second default value, drawing the characteristic vector.
Alternatively, the prediction module includes:
Second comparing unit, the size for comparing first probable value and second probable value;
First judging unit, for when first probable value is more than second probable value, judging at the main frame
In normal operating condition;
Second judging unit, for when first probable value is less than second probable value, judging that the main frame will
Break down.
The method and device of a kind of virtualized host operation conditions prediction provided by the present invention, by obtaining main frame
Real-time running data;According to above-mentioned real-time running data and predetermined threshold value, corresponding characteristic vector is drawn;Based on Bayesian statistics
Model and default Prior Probability, calculate features described above vector, show that the above-mentioned main frame of sign normally runs probability
First probable value and characterize above-mentioned main frame and break down the second probable value of probability;According to above-mentioned first probable value and above-mentioned
Two probable values, the operation conditions to above-mentioned main frame is predicted.The real-time running data of Intrusion Detection based on host generate corresponding feature to
Amount, recycles Bayesian statistical model to calculate characteristic vector, judges that the real-time characteristic vector of main frame is belonging to normally run class
Failure classes are still fallen within, that is, calculates corresponding probable value, predict whether main frame will occur according to the probable value for calculating
Failure.It can be seen that, the application can predict whether virtualized host will break down.
Brief description of the drawings
For the clearer explanation embodiment of the present invention or the technical scheme of prior art, below will be to embodiment or existing
The accompanying drawing to be used needed for technology description is briefly described, it should be apparent that, drawings in the following description are only this hair
Some bright embodiments, for those of ordinary skill in the art, on the premise of not paying creative work, can be with root
Other accompanying drawings are obtained according to these accompanying drawings.
The stream that a kind of specific implementation of the virtualized host operation conditions prediction that Fig. 1 is provided by the embodiment of the present invention sends
Journey schematic diagram;
The structured flowchart of the virtualized host operation conditions prediction meanss that Fig. 2 is provided by the embodiment of the present invention.
Specific embodiment
In order that those skilled in the art more fully understand the present invention program, with reference to the accompanying drawings and detailed description
The present invention is described in further detail.Obviously, described embodiment is only a part of embodiment of the invention, rather than
Whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art are not making creative work premise
Lower obtained every other embodiment, belongs to the scope of protection of the invention.
Refer to Fig. 1, a kind of specific reality of the virtualized host operation conditions prediction that Fig. 1 is provided by the embodiment of the present invention
The schematic flow sheet of transmission is applied, the method is comprised the following steps:
Step 101:Obtain the real-time running data of main frame;
It should be noted that above-mentioned real-time running data can refer to the service data of the main frame being currently running.Should
Real-time running data can refer to host monitor data, i.e. the data can reflect the current operating conditions of main frame.
Above-mentioned main frame can be one or many, you can pre- to carry out failure to one or multiple host simultaneously
Survey.
It is understood that the real-time running data of main frame can be got by the agent services run on main frame.
When above-mentioned main frame there are many, agent services can be run on each main frame, to get required service data.When
So, the mode for obtaining the real-time running data of main frame can also be other, be not limited thereto.
Before the real-time running data of current hosts is obtained, can be entered according to substantial amounts of history run daily record data
Row sample training, predetermined threshold value and default Prior Probability needed for drawing subsequent step.
In some embodiments of the invention, can also include before the real-time running data of the acquisition main frame:Obtain
Each history run daily record data of main frame is taken, using the history run daily record data as training sample data;To the instruction
Practice sample data to be pre-processed, draw the predetermined threshold value;It is according to this and corresponding described default based on the number of training
Threshold value, draws corresponding training sample characteristic vector;Using the Bayesian statistical model and the sampling feature vectors, calculate
Draw the default Prior Probability.
It should be noted that above-mentioned predetermined threshold value can refer to for main frame service data binaryzation to be characterized into vector
Standard, it can be worth to by calculating the average of training sample data.
Above-mentioned default Prior Probability can refer to the prior probability in Bayesian statistics, and the prior probability can be by right
Corresponding training sample data carry out calculating and arrive.Prior probability can correspondingly have main frame according to the classification of training sample data
The corresponding prior probability of data of normal operation, and the corresponding prior probability of data that main frame breaks down.
Drawing for predetermined threshold value can draw by calculating the average value of corresponding data, therefore in some embodiments of the present invention
In, the calculating process of threshold value can be specially:The training sample data are divided into normal operation class data and failure classes data
Two classes;According to previously selected training characteristics, calculate respectively in the normal operation class data and the failure classes data
The average value of training characteristics item data, using the average value as the predetermined threshold value.
Above-mentioned normal operation class data can refer to monitoring data when main frame normally runs.Above-mentioned failure classes data can be with
It refer to monitoring data when main frame breaks down.
Above-mentioned training characteristics can refer to data different classes of in service data, for example, can be with service data
Include cpu, I/O and Time class data.The characteristic item of the characteristic vector of composition main frame can be preset, for example, can be with
By the composition category setting of the characteristic vector of main frame for cpu, mem, bandwidth, I/O and Time, i.e. main frame feature to
Measure and be<Cpu, mem, bandwidth, I/O, Time>, also will cpu, mem, bandwidth, I/O and Time data make
It is characterized data.Certainly, characteristic vector can also be made up of other characteristics, be not limited thereto.
Above-mentioned average value can refer to that the characteristic of the characteristic vector of main frame average drawn after operation
Numerical value, will all characteristics of each characteristic item first sue for peace, it is resulting corresponding then divided by the item number of this feature
Average value.For example, characteristic vector is<t1, t2, t3, t4, t5>, one has n bar history run daily record datas, then have n characteristic item
t1, by n characteristic item t1Data all add up, draw additive value, will add up value divided by n, then can obtain characteristic item t1
Threshold value.
It should be evident that composition characteristic vector each characteristic item correspond to its corresponding threshold value, for example, feature to
Measure and be<t1, t2, t3, t4, t5>, for characteristic item t1、t2、t3、t4And t5For, its corresponding threshold value T1、T2、T3、T4And
T5。
After the characteristic item of preselect composition characteristic vector, and after calculating the threshold value of each characteristic item, can be right
Characteristic vector carries out binaryzation, to facilitate the calculating of follow-up Prior Probability.
Each history log service data can be expressed as a characteristic vector first, that is, choose each history day
Corresponding feature item data in will service data;Each feature item data is compared with corresponding threshold value again, works as characteristic item
When data are more than respective threshold, the numerical value of this feature can be taken as 1 or 0, and when feature item data is less than respective threshold,
The numerical value of this feature can be correspondingly taken as 0 or 1.Characteristic vector binaryzation can so be represented using 0 and 1
Characteristic vector.
For example, when a certain bar service data is<t1, t2, t3, t4, t5>When, setting takes when feature item data is more than threshold value
It is 1, conversely, being then 0;Now, t1>T1, t1<T1, t1>T1, t1<T1, t1<T1, then this feature vector be<1,0,1,0,0>.
The training sample characteristic vector of each training sample data is obtained using above-mentioned binarization method, then can be utilized
Bayesian statistical model, the training sample characteristic vector to drawing is calculated.
Above-mentioned Bayesian statistical model can be specially
BxtWhat is represented is the weight for occurring, BxtAppearance, B are represented when=1xtRepresented when=0 and occurred without;dxWhat is represented is data sample;cj
What is represented is sample class;wtWhat is represented is characteristic item;P(wt|cj) what is represented is that data sample belongs to cjWhen class, feature
Item wtThe probability of appearance;Conditional probability is the product of the class conditional probability of all characteristic items, characteristic item wtOccur in the text,
The item for multiplying is P (wt|cj), if not then multiplying 1-P (wt|cj)。
Above-mentioned P (dx|cj) service data d can be representedxBelong to class cjProbability, class cjIt can be above-mentioned normal operation
Class and failure classes it is therein any one.
Using Bayesian formula, training sample data are calculated, it can be deduced that belong to the elder generation of normal operation class data
Test probable value, and the Prior Probability for belonging to failure classes data.It is general that the Prior Probability for calculating is above-mentioned default priori
Rate value.It should be evident that the technology that the calculating process of Prior Probability is well known to those skilled in the art, no longer goes to live in the household of one's in-laws on getting married herein
State.
Step 102:According to the real-time running data and predetermined threshold value, corresponding characteristic vector is drawn;
It should be noted that above-mentioned predetermined threshold value can refer to the phase calculated previously according to history run daily record data
The average value answered, its calculating process can be referring specifically to corresponding contents above.The quantity of predetermined threshold value can have multiple, each
Predetermined threshold value may correspond to a feature item data.
The generation of characteristic vector can preselect feature item data, and selected feature item data is being carried out into two-value
Change, can then draw corresponding characteristic vector.Therefore in some embodiments of the invention, its process can be specially:According to
The default characteristic item of characteristic vector, chooses corresponding feature item data from the real-time running data;Compare each spy
Levy the size of item data and corresponding predetermined threshold value;When the feature item data is more than the predetermined threshold value, by individual features
The numerical value of item is set to the first default value;When the feature item data is less than the predetermined threshold value, by the number of individual features
Value is set to the second default value;According to first default value and second default value, the characteristic vector is drawn.
It is understood that above-mentioned default characteristic item can refer to previously selected data category, its quantity and classification can
With optional, for example, default characteristic item can be set as into cpu, mem, bandwidth, I/O and Time.
Features described above item data can refer to the specific data corresponding to previously selected data category, for example, when default
When characteristic item is set as cpu, mem, bandwidth, I/O and Time, feature item data correspond to cpu data, mem data,
Bandwidth data, I/O data and Time data.
Above-mentioned first default value can refer to 0 or 1, and correspondingly, above-mentioned second default value can refer to 1 or 0, that is, work as
When first default value is 0, the second default value corresponding positions 1.
Feature item data is compared size with corresponding predetermined threshold value, if feature item data is more than predetermined threshold value,
The value of current signature is taken as 1 or 0, conversely, being then taken as 0 or 1.Feature item data binaryzation can so be drawn corresponding
Characteristic vector.
It should be noted that when the characteristic item of features described above vector has 5, it is concretely<t1, t2, t3, t4, t5>,
Wherein, t1、t2、t3、t4And t5Span be interval [0,1].Certainly, the feature item number that characteristic vector is included can be
Other, are not limited thereto.Specific statement as characteristic vector may refer to corresponding contents above, will not be repeated here.
Step 103:Based on Bayesian statistical model and default Prior Probability, the characteristic vector is calculated,
Draw to characterize the main frame and normally run the first probable value of probability and characterize the main frame and break down the second probability of probability
Value;
Above-mentioned Bayesian statistical model can be specially
Above-mentioned first probable value can characterize the possibility that main frame normally runs, and the second probable value can characterize main frame and incite somebody to action
The possibility to be broken down.
Default prior probability can be general including the first default priori calculated by above normal operation class data
Rate and the second default prior probability drawn by above failure classes numerical computations.Therefore in some embodiments of the present invention
In, the calculating process of its probable value can be specially:Normally run using the Bayesian statistical model and the sign main frame
The first default Prior Probability, the characteristic vector is calculated, draw first probable value;Using the Bayes
Second default Prior Probability of statistical model and the sign hostdown, calculates the characteristic vector, draws institute
State the second probable value.
It should be noted that known prior probability and characteristic vector, posterior probability is calculated using Bayesian model
It is technology well-known to those skilled in the art, will not be repeated here.
Step 104:According to first probable value and second probable value, the operation conditions to the main frame is carried out
Prediction.
It is understood that above-mentioned first probable value and the second probable value each mean that the posteriority in Bayesian statistical model is general
Rate.Can be belonging to the current signature vector for judging main frame by comparing the size of the first probable value and the second probable value
Whether failure classes still fall within normal operation class, you can normal with the operation conditions for predicting main frame.
In some embodiments of the invention, predict that the process of hostdown can be specific by comparing the size of probable value
For:Compare the size of first probable value and second probable value;When first probable value is more than second probability
During value, judge that the main frame is in normal operating condition;When first probable value is less than second probable value, institute is judged
Stating main frame will break down.
It should be noted that when the first probable value and equal the second probable value, it can be determined that main frame is in normal operation
State, it is also possible to judge that main frame will break down, its criterion can artificially set.
In order to the main frame in time to that will break down is processed, it can will be broken down main frame is judged
Afterwards, early warning is reported to operate accordingly.
Therefore in some embodiments of the invention, described when first probable value is less than second probable value,
Can also include after judging the main frame and will breaking down:The information of the main frame is sent to specified receiving terminal, with
The specified receiving terminal is set to send early warning information according to described information.
It should be noted that above-mentioned specified receiving terminal can be preassigned, it can be fixed terminal, it is also possible to
It is mobile terminal.
The information of above-mentioned main frame can refer to the specifying information of current hosts, and user can be true rapidly according to the information
It is fixed to being which platform main frame breaks down.
And above-mentioned early warning information can refer to the warning message that warning device is sent.Type of alarm can be by buzzing
Device sends alarm;Corresponding actuation of an alarm can also be carried out by alarm lamp;Can also be other alarm operations.
The method of the virtualized host operation conditions prediction that the embodiment of the present invention is provided, by the real-time fortune for obtaining main frame
Row data;According to above-mentioned real-time running data and predetermined threshold value, corresponding characteristic vector is drawn;Based on Bayesian statistical model with
And default Prior Probability, features described above vector is calculated, show that characterizing above-mentioned main frame normally runs the first general of probability
Rate value and characterize above-mentioned main frame and break down the second probable value of probability;According to above-mentioned first probable value and above-mentioned second probability
Value, the operation conditions to above-mentioned main frame is predicted.The real-time running data of Intrusion Detection based on host generates corresponding characteristic vector, then profit
Characteristic vector is calculated with Bayesian statistical model, is judged that the real-time characteristic vector of main frame is belonging to normally to run class and is still fallen within
Failure classes, that is, calculate corresponding probable value, predicts whether main frame will break down according to the probable value for calculating.Can
See, the method can predict whether virtualized host will break down.
Virtualized host operation conditions prediction meanss provided in an embodiment of the present invention are introduced below, it is described below
Virtualized host operation conditions prediction meanss can mutually corresponding ginseng with above-described virtualized host operation conditions Forecasting Methodology
According to.
The structured flowchart of the virtualized host operation conditions prediction meanss that Fig. 2 is provided by the embodiment of the present invention, reference picture 2
Virtualized host operation conditions prediction meanss can include:
Acquisition module 201, the real-time running data for obtaining main frame;
Feature vector generation module 202, for according to the real-time running data and predetermined threshold value, drawing corresponding feature
Vector;
Probable value computing module 203, for based on Bayesian statistical model and default Prior Probability, to the feature
Vector is calculated, and is shown that the sign main frame normally runs the first probable value of probability and characterizes the main frame and is broken down generally
Second probable value of rate;
Prediction module 204, for according to first probable value and second probable value, the operation to the main frame
Situation is predicted.
Alternatively, the feature vector generation module includes:
Feature item data chooses unit, for the default characteristic item according to characteristic vector, from the real-time running data
Choose corresponding feature item data;
First comparing unit, the size for comparing each described feature item data and corresponding predetermined threshold value;
First setup unit, for when the feature item data is more than the predetermined threshold value, by the number of individual features
Value is set to the first default value;
Second setup unit, for when the feature item data is less than the predetermined threshold value, by the number of individual features
Value is set to the second default value;
Generation unit, for according to first default value and second default value, drawing the characteristic vector.
Alternatively, the prediction module includes:
Second comparing unit, the size for comparing first probable value and second probable value;
First judging unit, for when first probable value is more than second probable value, judging at the main frame
In normal operating condition;
Second judging unit, for when first probable value is less than second probable value, judging that the main frame will
Break down.
The device of the virtualized host operation conditions prediction that the embodiment of the present invention is provided, the real time execution number of Intrusion Detection based on host
According to corresponding characteristic vector is generated, recycle Bayesian statistical model to calculate characteristic vector, judge the real-time characteristic of main frame to
Amount is belonging to normally to run class and still falls within failure classes, that is, calculate corresponding probable value, according to the probable value for calculating, comes pre-
Survey whether main frame will break down.It can be seen that, the device can predict whether virtualized host will break down.
Each embodiment is described by the way of progressive in this specification, and what each embodiment was stressed is and other
The difference of embodiment, between each embodiment same or similar part mutually referring to.For being filled disclosed in embodiment
For putting, because it is corresponded to the method disclosed in Example, so description is fairly simple, related part is referring to method part
Illustrate.
Professional further appreciates that, with reference to the unit of each example of the embodiments described herein description
And algorithm steps, can be realized with electronic hardware, computer software or the combination of the two, in order to clearly demonstrate hardware and
The interchangeability of software, generally describes the composition and step of each example according to function in the above description.These
Function is performed with hardware or software mode actually, depending on the application-specific and design constraint of technical scheme.Specialty
Technical staff can realize described function to each specific application using distinct methods, but this realization should not
Think beyond the scope of this invention.
The step of method or algorithm for being described with reference to the embodiments described herein, directly can be held with hardware, processor
Capable software module, or the two combination is implemented.Software module can be placed in random access memory (RAM), internal memory, read-only deposit
Reservoir (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technology
In field in known any other form of storage medium.
Virtualized host operation conditions Forecasting Methodology provided by the present invention and device are described in detail above.
Specific case used herein is set forth to principle of the invention and implementation method, and the explanation of above example is use
Understand the method for the present invention and its core concept in help.It should be pointed out that for those skilled in the art,
Under the premise without departing from the principles of the invention, some improvement and modification can also be carried out to the present invention, these improve and modify
Fall into the protection domain of the claims in the present invention.
Claims (10)
1. a kind of method that virtualized host operation conditions is predicted, it is characterised in that including:
Obtain the real-time running data of main frame;
According to the real-time running data and predetermined threshold value, corresponding characteristic vector is drawn;
Based on Bayesian statistical model and default Prior Probability, the characteristic vector is calculated, show that sign is described
Main frame normally runs the first probable value of probability and characterizes the main frame and breaks down the second probable value of probability;
According to first probable value and second probable value, the operation conditions to the main frame is predicted.
2. the method for claim 1, it is characterised in that described according to the real-time running data and predetermined threshold value, obtains
Going out corresponding characteristic vector includes:
According to the default characteristic item of characteristic vector, corresponding feature item data is chosen from the real-time running data;
Compare the size of each described feature item data and corresponding predetermined threshold value;
When the feature item data is more than the predetermined threshold value, the numerical value of individual features is set to the first default value;
When the feature item data is less than the predetermined threshold value, the numerical value of individual features is set to the second default value;
According to first default value and second default value, the characteristic vector is drawn.
3. the method for claim 1, it is characterised in that described based on Bayesian statistical model and default prior probability
Value, calculates the characteristic vector, show that characterizing the main frame normally runs the first probable value of probability and characterize described
Break down the second probable value of probability of main frame includes:
The the first default Prior Probability normally run using the Bayesian statistical model and the sign main frame, to the spy
Levy vector to be calculated, draw first probable value;
Using the Bayesian statistical model and characterize the second default Prior Probability of the hostdown, to the feature to
Amount is calculated, and draws second probable value.
4. the method for claim 1, it is characterised in that described according to first probable value and second probability
Value, the operation conditions of the main frame is predicted including:
Compare the size of first probable value and second probable value;
When first probable value is more than second probable value, judge that the main frame is in normal operating condition;
When first probable value is less than second probable value, judge that the main frame will break down.
5. method as claimed in claim 4, it is characterised in that described when first probable value is less than second probability
During value, judge that the main frame will also include after will breaking down:
The information of the main frame is sent to specified receiving terminal, so that the specified receiving terminal sends pre- according to described information
Alert information.
6. the method as described in any one of claim 1 to 5, it is characterised in that in the real-time running data of the acquisition main frame
Also include before:
Each history run daily record data of main frame is obtained, using the history run daily record data as training sample data;
The training sample data are pre-processed, the predetermined threshold value is drawn;
Based on the number of training according to this and the corresponding predetermined threshold value, corresponding training sample characteristic vector is drawn;
Using the Bayesian statistical model and the sampling feature vectors, the default Prior Probability is calculated.
7. method as claimed in claim 6, it is characterised in that described to be pre-processed to the training sample data, draws
The predetermined threshold value includes:
The training sample data are divided into normal operation class data and the class of failure classes data two;
According to previously selected training characteristics, the instruction in the normal operation class data and the failure classes data is calculated respectively
Practice the average value of feature item data, using the average value as the predetermined threshold value.
8. the device that a kind of virtualized host operation conditions is predicted, it is characterised in that including:
Acquisition module, the real-time running data for obtaining main frame;
Feature vector generation module, for according to the real-time running data and predetermined threshold value, drawing corresponding characteristic vector;
Probable value computing module, for based on Bayesian statistical model and default Prior Probability, entering to the characteristic vector
Row is calculated, and is drawn to characterize the main frame and normally run the first probable value of probability and characterize the main frame and is broken down the of probability
Two probable values;
Prediction module, for according to first probable value and second probable value, the operation conditions to the main frame to be entered
Row prediction.
9. device as claimed in claim 8, it is characterised in that the feature vector generation module includes:
Feature item data chooses unit, for the default characteristic item according to characteristic vector, is chosen from the real-time running data
Corresponding feature item data;
First comparing unit, the size for comparing each described feature item data and corresponding predetermined threshold value;
First setup unit, for when the feature item data is more than the predetermined threshold value, the numerical value of individual features being set
It is the first default value;
Second setup unit, for when the feature item data is less than the predetermined threshold value, the numerical value of individual features being set
It is the second default value;
Generation unit, for according to first default value and second default value, drawing the characteristic vector.
10. device as claimed in claim 8, it is characterised in that the prediction module includes:
Second comparing unit, the size for comparing first probable value and second probable value;
First judging unit, for when first probable value is more than second probable value, judging that the main frame is in just
Normal running status;
Second judging unit, for when first probable value is less than second probable value, judging that the main frame will be sent out
Raw failure.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710100386.7A CN106897113A (en) | 2017-02-23 | 2017-02-23 | The method and device of a kind of virtualized host operation conditions prediction |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710100386.7A CN106897113A (en) | 2017-02-23 | 2017-02-23 | The method and device of a kind of virtualized host operation conditions prediction |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106897113A true CN106897113A (en) | 2017-06-27 |
Family
ID=59185426
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710100386.7A Pending CN106897113A (en) | 2017-02-23 | 2017-02-23 | The method and device of a kind of virtualized host operation conditions prediction |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106897113A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108919776A (en) * | 2018-06-19 | 2018-11-30 | 深圳市元征科技股份有限公司 | A kind of assessment of failure method and terminal |
CN110769003A (en) * | 2019-11-05 | 2020-02-07 | 杭州安恒信息技术股份有限公司 | Network security early warning method, system, equipment and readable storage medium |
CN111460140A (en) * | 2020-03-05 | 2020-07-28 | 拉扎斯网络科技(上海)有限公司 | Data processing method and device, electronic equipment and computer readable storage medium |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103699698A (en) * | 2014-01-16 | 2014-04-02 | 北京泰乐德信息技术有限公司 | Method and system for track traffic failure recognition based on improved Bayesian algorithm |
CN104063747A (en) * | 2014-06-26 | 2014-09-24 | 上海交通大学 | Performance abnormality prediction method in distributed system and system |
CN104807639A (en) * | 2015-04-23 | 2015-07-29 | 广西大学 | Fault diagnosis method and device for rolling bearing of running gear of locomotive |
CN105700518A (en) * | 2016-03-10 | 2016-06-22 | 华中科技大学 | Fault diagnosis method during industrial process |
-
2017
- 2017-02-23 CN CN201710100386.7A patent/CN106897113A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103699698A (en) * | 2014-01-16 | 2014-04-02 | 北京泰乐德信息技术有限公司 | Method and system for track traffic failure recognition based on improved Bayesian algorithm |
CN104063747A (en) * | 2014-06-26 | 2014-09-24 | 上海交通大学 | Performance abnormality prediction method in distributed system and system |
CN104807639A (en) * | 2015-04-23 | 2015-07-29 | 广西大学 | Fault diagnosis method and device for rolling bearing of running gear of locomotive |
CN105700518A (en) * | 2016-03-10 | 2016-06-22 | 华中科技大学 | Fault diagnosis method during industrial process |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108919776A (en) * | 2018-06-19 | 2018-11-30 | 深圳市元征科技股份有限公司 | A kind of assessment of failure method and terminal |
CN108919776B (en) * | 2018-06-19 | 2021-04-09 | 深圳市元征科技股份有限公司 | Fault assessment method and terminal |
CN110769003A (en) * | 2019-11-05 | 2020-02-07 | 杭州安恒信息技术股份有限公司 | Network security early warning method, system, equipment and readable storage medium |
CN110769003B (en) * | 2019-11-05 | 2022-02-22 | 杭州安恒信息技术股份有限公司 | Network security early warning method, system, equipment and readable storage medium |
CN111460140A (en) * | 2020-03-05 | 2020-07-28 | 拉扎斯网络科技(上海)有限公司 | Data processing method and device, electronic equipment and computer readable storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP3570494B1 (en) | A framework for intelligent automated operations for network, service & customer experience management | |
US11586972B2 (en) | Tool-specific alerting rules based on abnormal and normal patterns obtained from history logs | |
CN110351150B (en) | Fault source determination method and device, electronic equipment and readable storage medium | |
US10599506B2 (en) | Methods and systems for identifying action for responding to anomaly in cloud computing system | |
US11805005B2 (en) | Systems and methods for predictive assurance | |
CN111539493B (en) | Alarm prediction method and device, electronic equipment and storage medium | |
CN109670690A (en) | Data information center monitoring and early warning method, system and equipment | |
KR20220114986A (en) | Apparatus for VNF Anomaly Detection based on Machine Learning for Virtual Network Management and a method thereof | |
CN110287316A (en) | A kind of Alarm Classification method, apparatus, electronic equipment and storage medium | |
CN106897113A (en) | The method and device of a kind of virtualized host operation conditions prediction | |
US11882095B2 (en) | Firewall insights processing and machine learning | |
US20230205516A1 (en) | Software change analysis and automated remediation | |
CN110730164A (en) | Safety early warning method, related equipment and computer readable storage medium | |
CN116418653A (en) | Fault positioning method and device based on multi-index root cause positioning algorithm | |
CN111897700A (en) | Application index monitoring method and device, electronic equipment and readable storage medium | |
CN110414591A (en) | A kind of data processing method and equipment | |
CN109800124A (en) | CPU usage monitoring method, device, electronic equipment and storage medium | |
US20220222350A1 (en) | Vulnerability dashboard and automated remediation | |
CN112737835A (en) | Cloud host business management method and related device | |
CN113835961B (en) | Alarm information monitoring method, device, server and storage medium | |
CN115686756A (en) | Virtual machine migration method and device, storage medium and electronic equipment | |
CN114385398A (en) | Request response state determination method, device, equipment and storage medium | |
CN104883273B (en) | The processing method and system of service impact model in virtualization services management platform | |
CN113114538A (en) | Heartbeat detection method and device | |
US11991037B2 (en) | Systems and methods for reducing a quantity of false positives associated with rule-based alarms |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
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: 20170627 |