CN110445939A - The prediction technique and device of capacity resource - Google Patents
The prediction technique and device of capacity resource Download PDFInfo
- Publication number
- CN110445939A CN110445939A CN201910729638.1A CN201910729638A CN110445939A CN 110445939 A CN110445939 A CN 110445939A CN 201910729638 A CN201910729638 A CN 201910729638A CN 110445939 A CN110445939 A CN 110445939A
- Authority
- CN
- China
- Prior art keywords
- process number
- host
- prediction
- obtains
- capacity resource
- 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.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/147—Network analysis or design for predicting network behaviour
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M3/00—Automatic or semi-automatic exchanges
- H04M3/22—Arrangements for supervision, monitoring or testing
- H04M3/36—Statistical metering, e.g. recording occasions when traffic exceeds capacity of trunks
- H04M3/362—Traffic simulation
Abstract
The embodiment of the present invention provides the prediction technique and device of a kind of capacity resource, after the target service process number and target service parameter of the business for obtaining running on host according to predicting strategy, target service process number is first inputted to the prediction model for the reference capacity resource that training obtains in advance, obtain the benchmark treatment effeciency of host, then using benchmark treatment effeciency as a factor, the prediction model for the actual capacity resource that training obtains in advance is inputted together with target service parameter, obtain the actual treatment efficiency of host, and then it can be according to the actual treatment efficiency of host, predict capacity resource required in following preset time period, both production environment is considered, test environment has been taken into account again, the accuracy of capacity prediction is greatly improved, and it does not need manually to participate in, forecasting efficiency with higher.
Description
Technical field
The present invention relates to the prediction techniques and device of electronic technology field more particularly to a kind of capacity resource.
Background technique
With the continuous expansion of common carrier data scale, telecommunication service type is increasing, for handling telecommunications
The processing pressure of each business host of business is also increasing, and the processing upper limit to host, energy of meeting an urgent need are needed in production process
The capacity resources such as power, system stability, which have, explicitly to be estimated.
Tradition estimates mode to capacity resource at present, is mostly based on daily statistical data and empirical equation completion is added to estimate.
Generally estimating process includes: to be based on script and database in the case where artificial participate in, and utilizes simple average tendency analysis etc.
Method estimates out current or recent capacity resource situation according to artificial micro-judgment.
But the low efficiency of existing prediction mode and prediction result accuracy is low.
Summary of the invention
The embodiment of the present invention provides the prediction technique and device of a kind of capacity resource, with the effect of the prediction of hoist capacity resource
Rate.
First aspect of the embodiment of the present invention provides a kind of prediction technique of capacity resource, comprising:
According to the target service process number and target service parameter of the business that predicting strategy obtains running on host；
The prediction model that the target service process number is inputted to the reference capacity resource that training obtains in advance obtains described
The benchmark treatment effeciency of host；
The target service parameter and benchmark treatment effeciency input are trained into obtained actual capacity resource in advance
Prediction model obtains the actual treatment efficiency of the host；
According to the actual treatment efficiency of the host, capacity resource required in following preset time period is predicted.
Optionally, the target service parameter includes at least one of following parameter: targeted loads rate, target ticket
Amount, target user's number and targeted peripheral resource proportion；And/or
The capacity resource includes at least one of following resource: host number, the quantity of virtual machine, central processing unit
The processing speed and memory size of CPU.
Optionally, when the service parameter includes the target ticket amount and target user's number, according to prediction plan
Slightly obtain the target service parameter of the business run on host, comprising:
Using autoregressive moving average ARMA algorithm, operation is carried out to history ticket amount and historical user's number, is obtained described
Target ticket amount and target user's number.
Optionally, the target service process number of the business for obtaining running on host according to predicting strategy, comprising:
The prediction model that the initial value of business process number is inputted to the reference capacity resource, obtains the test of the host
Treatment effeciency；
The value of the business process number is successively increased according to preset step-length, and inputs the prediction of the reference capacity resource
Model obtains the test processes efficiency of the business process number corresponding host under different value conditions；
According to the test processes efficiency of the business process number corresponding host under different value conditions, institute is determined
The test processes efficiency for stating host when being maximum value the value of the corresponding business process number be the target service process number.
Optionally, the business process number includes at least one of following process number: average core business process number is put down
Important business process number and average total business process number.
Second aspect of the embodiment of the present invention provides a kind of prediction meanss of capacity resource, comprising:
Module is obtained, the target service process number and target industry of the business for obtaining running on host according to predicting strategy
Business parameter；
First input module trains obtained reference capacity resource for inputting the target service process number in advance
Prediction model obtains the benchmark treatment effeciency of the host；
Second input module, for obtaining the target service parameter and the preparatory training of benchmark treatment effeciency input
Actual capacity resource prediction model, obtain the actual treatment efficiency of the host；
Prediction module is predicted required in following preset time period for the actual treatment efficiency according to the host
Capacity resource.
Optionally, the target service parameter includes at least one of following parameter: targeted loads rate, target ticket
Amount, target user's number and targeted peripheral resource proportion；And/or
The capacity resource includes at least one of following resource: host number, the quantity of virtual machine, central processing unit
The processing speed and memory size of CPU.
Optionally, described to obtain mould when the service parameter includes the target ticket amount and target user's number
Block is also used to:
Using autoregressive moving average ARMA algorithm, operation is carried out to history ticket amount and historical user's number, is obtained described
Target ticket amount and target user's number.
Optionally, the module that obtains is also used to:
The prediction model that the initial value of business process number is inputted to the reference capacity resource, obtains the test of the host
Treatment effeciency；The value of the business process number is successively increased according to preset step-length, and inputs the pre- of the reference capacity resource
Model is surveyed, the test processes efficiency of the business process number corresponding host under different value conditions is obtained；According to institute
The test processes efficiency of business process number corresponding host under different value conditions is stated, at the test for determining the host
The value for managing corresponding business process number when efficiency is maximum value is the target service process number.
Optionally, the business process number includes at least one of following process number: average core business process number is put down
Important business process number and average total business process number.
The third aspect of the embodiment of the present invention provides a kind of electronic equipment, comprising: processor, memory and computer journey
Sequence；Wherein, the computer program is stored in the memory, and is configured as being executed by the processor, described
Computer program includes for executing the instruction such as the described in any item methods of aforementioned first aspect.
Fourth aspect of the embodiment of the present invention provides a kind of computer readable storage medium, the computer readable storage medium
It is stored with computer program, the computer program is performed, and realizes such as the described in any item methods of aforementioned first aspect.
The embodiment of the present invention compared with the existing technology the utility model has the advantages that
In the prediction technique and device of a kind of capacity resource provided in an embodiment of the present invention, it is contemplated that in model training usually
Need a large amount of training data, and production environment generally can not obtain a large amount of creation data, according only to the number of production environment
According to can not usually carry out model training or can not train to obtain accurate prediction model, therefore in the embodiment of the present invention, use
The prediction model of two-part, specifically, the prediction model of reference capacity resource can be the model of the training in test environment,
The test data in all kinds of situations can be collected in test environment to be trained, therefore, the prediction mould of the reference capacity resource
Type can export the benchmark treatment effeciency in more situation；When being predicted by the prediction model of actual capacity resource, by benchmark
Treatment effeciency and creation data are used as the input of the prediction model of capacity resource, then not only consider production environment, but also take into account
Environment is tested, the accuracy of capacity prediction is greatly improved, and does not need manually to participate in, prediction effect with higher
Rate.Specifically, in the embodiment of the present invention, according to the target service process number and mesh of the business that predicting strategy obtains running on host
After marking service parameter, target service process number is first inputted to the prediction model for the reference capacity resource that training obtains in advance, is obtained
The benchmark treatment effeciency of host inputs preparatory then using benchmark treatment effeciency as a factor together with target service parameter
The prediction model for the actual capacity resource that training obtains, obtains the actual treatment efficiency of host, and then can be according to the reality of host
Border treatment effeciency predicts capacity resource required in following preset time period, not only considers production environment, but also taken into account test
The accuracy of environment, capacity prediction is greatly improved, and does not need manually to participate in, forecasting efficiency with higher.
Detailed description of the invention
Fig. 1 is the flow diagram of the prediction technique of capacity resource provided in an embodiment of the present invention；
Fig. 2 is the first test result schematic diagram of the prediction technique of capacity resource provided in an embodiment of the present invention；
Fig. 3 is second of test result schematic diagram of the prediction technique of capacity resource provided in an embodiment of the present invention；
Fig. 4 is the third test result schematic diagram of the prediction technique of capacity resource provided in an embodiment of the present invention；
Fig. 5 is the 4th kind of test result schematic diagram of the prediction technique of capacity resource provided in an embodiment of the present invention；
Fig. 6 is that a kind of benchmark of the embodiment of the present invention handles EFFICIENCY PREDICTION effect diagram；
Fig. 7 is a kind of prediction model prediction effect schematic diagram of capacity resource of the embodiment of the present invention；
Fig. 8 is the model framework schematic diagram of the prediction technique of capacity resource provided in an embodiment of the present invention；
Fig. 9 is the structural schematic diagram of one embodiment of prediction meanss of capacity resource provided by the invention.
Specific embodiment
Below in conjunction with the attached drawing in the embodiment of the present application, technical solutions in the embodiments of the present application is described, and is shown
So, described embodiments are only a part of embodiments of the present application, instead of all the embodiments.Based on the reality in the application
Example is applied, every other embodiment obtained by those of ordinary skill in the art without making creative efforts all belongs to
In the range of the application protection.
Example embodiments are described in detail here, and the example is illustrated in the accompanying drawings.Following description is related to
When attached drawing, unless otherwise indicated, the same numbers in different drawings indicate the same or similar elements.Following exemplary embodiment
Described in embodiment do not represent all embodiments consistented with the present invention.On the contrary, they be only with it is such as appended
The example of device and method being described in detail in claims, some aspects of the invention are consistent.
It will be appreciated that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.Base
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts it is all its
Its embodiment, shall fall within the protection scope of the present invention.
The term used in embodiments of the present invention is only to be not intended to be limiting merely for for the purpose of describing particular embodiments
The present invention.In the embodiment of the present invention and the "an" of singular used in the attached claims, " described " and "the"
It is also intended to including most forms, unless the context clearly indicates other meaning.
It should be appreciated that term "and/or" used herein is only a kind of incidence relation for describing affiliated partner, indicate
There may be three kinds of relationships, for example, A and/or B, can indicate: individualism A, exist simultaneously A and B, individualism B these three
Situation.In addition, character "/" herein, typicallys represent the relationship that forward-backward correlation object is a kind of "or".
It should be appreciated that description of the invention and claims and term " first " in above-mentioned attached drawing, " second ", " the
Three ", the (if present)s such as " 4th " are to be used to distinguish similar objects, without for describing specific sequence or successive time
Sequence.
Depending on context, word as used in this " if ", " if " can be construed to " ... when " or
" when ... " or " in response to determination " or " in response to detection ".Similarly, context is depended on, phrase " if it is determined that " or " such as
Fruit detection (condition or event of statement) " can be construed to " when determining " or " in response to determination " or " when detection (statement
Condition or event) when " or " in response to detection (condition or event of statement) ".
It should also be noted that, the terms "include", "comprise" or its any other variant are intended to nonexcludability
Include, so that commodity or system including a series of elements not only include those elements, but also including not clear
The other element listed, or further include for this commodity or the intrinsic element of system.In the feelings not limited more
Under condition, the element that is limited by sentence "including a ...", it is not excluded that in the commodity or system for including the element also
There are other identical elements.
In the prediction technique and device of a kind of capacity resource provided in an embodiment of the present invention, it is contemplated that in model training usually
Need a large amount of training data, and production environment generally can not obtain a large amount of creation data, according only to the number of production environment
According to can not usually carry out model training or can not train to obtain accurate prediction model, therefore in the embodiment of the present invention, use
The prediction model of two-part, specifically, the prediction model of reference capacity resource can be the model of the training in test environment,
The test data in all kinds of situations can be collected in test environment to be trained, therefore, the prediction mould of the reference capacity resource
Type can export the benchmark treatment effeciency in more situation；When being predicted by the prediction model of actual capacity resource, by benchmark
Treatment effeciency and creation data are used as the input of the prediction model of capacity resource, then not only consider production environment, but also take into account
Environment is tested, the accuracy of capacity prediction is greatly improved, and does not need manually to participate in, prediction effect with higher
Rate.Specifically, in the embodiment of the present invention, according to the target service process number and mesh of the business that predicting strategy obtains running on host
After marking service parameter, target service process number is first inputted to the prediction model for the reference capacity resource that training obtains in advance, is obtained
The benchmark treatment effeciency of host inputs preparatory then using benchmark treatment effeciency as a factor together with target service parameter
The prediction model for the actual capacity resource that training obtains, obtains the actual treatment efficiency of host, and then can be according to the reality of host
Border treatment effeciency predicts capacity resource required in following preset time period, not only considers production environment, but also taken into account test
The accuracy of environment, capacity prediction is greatly improved, and does not need manually to participate in, forecasting efficiency with higher.
The embodiment of the present invention can be using in the terminal, and terminal may include: mobile phone, tablet computer, laptop, platform
Formula computer or server etc. can be with the electronic equipments of the prediction technique of working capacity resource.
Treatment effeciency described in the embodiment of the present invention is the portfolio handled in the host unit time, and portfolio specifically may be used
To include: ticket amount, request of data number and/or data package size etc..
The prediction model of reference capacity resource described in the embodiment of the present invention can be based on depth learning technology, and
The neural network model that test data training in test environment obtains.Specifically, depth learning technology is one kind by extensive
The machine learning algorithm that neuron is constituted, which, which mainly passes through, collects a large amount of basic data, studies the rule between wherein data
Rule, forms certain statistical model, constructs model of mind using regression analysis, time series, neural network scheduling algorithm.
The prediction model of actual capacity resource described in the embodiment of the present invention can be based on depth learning technology, and
The neural network model that creation data training in production environment obtains.
In concrete application, this field is typically easy to think in the prediction model of training capacity resource neural network based
To the test sample that is only based on be trained, or be based only upon production sample be trained.
But be based only upon in the model that test sample is trained, because being deposited between test sample and actual sample
In difference, therefore accurate capacity prediction effect generally can not be obtained；And it is based only upon the model that production sample is trained
In, due to being difficult to obtain a large amount of production sample, and then can not be according to a large amount of production sample training neural network model.Also just
It is the lower detections of efficiency such as artificial prediction in the application scenarios of capacity resource prediction, to be generallyd use based on the reason, and do not have
There is the technical solution that capacity prediction is carried out based on neural network model.
Based on this, in the embodiment of the present invention, using the prediction model of two-part, specifically, reference capacity resource
Prediction model can be the model of the training in test environment, and the test data in all kinds of situations can be collected in test environment
It is trained, therefore, the prediction model of the reference capacity resource can export the benchmark treatment effeciency in more situation；Passing through
When the prediction model prediction of actual capacity resource, benchmark treatment effeciency and creation data are used as to the prediction model of capacity resource
Input, then not only consider production environment, but also taken into account test environment, capacity prediction accuracy be greatly improved,
And it does not need manually to participate in, forecasting efficiency with higher.
As shown in FIG. 1, FIG. 1 is the flow diagrams of the prediction technique of capacity resource provided in an embodiment of the present invention.The party
Method can specifically include:
Step S101: the target service process number and target service for the business for obtaining running on host according to predicting strategy are joined
Number.
In the embodiment of the present invention, predicting strategy can be determined according to actual application scenarios, exemplary, can be according to history
Data predict the data of certain time, for example, according to the target service process number and mesh of the business of double ten a period of time hosts operations in this year
Service parameter is marked, can predict the target service process number and target service parameter of the business of double ten a period of time hosts operations in next year
Deng, it will be understood that predicting strategy can determine that the present invention is not especially limit this according to actual application scenarios.
Optionally, the business process number includes at least one of following process number: average core business process number is put down
Important business process number and average total business process number.
By analysis hosted environment in the embodiment of the present invention, discovery is centered on telecommunication service, by system resource with industry
Business dimension be split, by test environment continuous test collection test data, reapply regression analysis, principal component analysis,
After the data processing methods such as Data Dimensionality Reduction, by business complexity based on business process, production actual treatment bottleneck, host
In the factors such as central processor (central processing unit, CPU) memory accounting, it is only necessary to study host average traffic
Process number, therefore be extracted three main factors: average core business process number, average basal business process number,
Average total business process number.According to three Main Factors training, can be built in conjunction with the methods of missing values completion, data normalization
The prediction model of vertical reference capacity resource.
Therefore, in predicting strategy, obtained business process number includes: average core business process number, average important industry
When at least one of business process number and average total business process number, preferable prediction effect can be obtained.
Core business process number, average important service process number and average total industry it is appreciated that average core business is averaged
The particular content of business process number can be determined according to actual application scenarios, and the embodiment of the present invention does not limit this specifically
It is fixed.
Optionally, the target service process number of the business for obtaining running on host according to predicting strategy, comprising:
The prediction model that the initial value of business process number is inputted to the reference capacity resource, obtains the test of the host
Treatment effeciency；The value of the business process number is successively increased according to preset step-length, and inputs the pre- of the reference capacity resource
Model is surveyed, the test processes efficiency of the business process number corresponding host under different value conditions is obtained；According to institute
The test processes efficiency of business process number corresponding host under different value conditions is stated, at the test for determining the host
The value for managing corresponding business process number when efficiency is maximum value is the target service process number.
In the embodiment of the present invention, the initial value and preset step-length of business process number can be according to actual application scenarios
It determines, the present invention is not especially limit this.
In the embodiment of the present invention, imitated according to the test processes of business process number corresponding host under different value conditions
Rate, the test processes efficiency for determining host when being maximum value the value of corresponding business process number be target service process number, because
Test processes efficiency for host is that maximum value is desired treatment effeciency value, it is thus determined that the test processes efficiency of host is most
The value of corresponding business process number is target service process number when big value, can obtain desired benchmark treatment effeciency.
Optionally, the target service parameter includes at least one of following parameter: targeted loads rate, target ticket
Amount, target user's number and targeted peripheral resource proportion.
In the embodiment of the present application, it is contemplated that with the continuous expansion of current common carrier data scale, type of business
It is increasing, the processing pressure of each business host is also increasing in daily production, and the place to host is needed in production process
The capacity resources such as the reason upper limit, emergency capability, system stability, which have, explicitly to be estimated.It therefore, can be by the appearance of the embodiment of the present invention
The prediction technique of amount resource is applied in based on telecommunication system business and distributed production environment, and inter-trade data mining mark is based on
The process of quasi- process (cross-industry standard process for data mining, CRISP-DM) specification is come
It realizes.
In based on telecommunication system business and distributed production environment, load factor, ticket amount, number of users and peripheral resource are matched
Than being more commonly used and important service parameter, therefore, in embodiments of the present invention, the target service parameter includes following ginseng
At least one of number: targeted loads rate, target ticket amount, target user's number and targeted peripheral resource proportion.
Optionally, when the service parameter includes the target ticket amount and target user's number, according to prediction plan
Slightly obtain the target service parameter of the business run on host, comprising: use autoregressive moving average (Auto-Regressive
And Moving Average Model, ARMA) algorithm, operation is carried out to history ticket amount and historical user's number, is obtained described
Target ticket amount and target user's number.
Step S102: the target service process number is inputted to the prediction mould for the reference capacity resource that training obtains in advance
Type obtains the benchmark treatment effeciency of the host.
In the embodiment of the present invention, the prediction model of reference capacity resource can be by the methods of convolutional neural networks, first
Obtain the test data sample under each scene, it will be understood that because the flexibility ratio of test environment is higher, can will be all kinds of
The sample of prominent situation and various risks prediction scene is included, subsequent when emergency is predicted or faces complex scene, only needs
The factor input configuration for changing current scene, can obtain accurate prediction result, just so as to improve production environment
The ability of upper reply risk.
Specifically, the prediction model building process of reference capacity resource M1 can be with are as follows:
The prediction model of reference capacity resource M1: prediction benchmark treatment effeciency (Z1)=average core business process number (x1)
~important service process the number (x2) that is averaged~average total business process number (x3)
Functional expression indicates: Z1=M1 (x1, x2, x3)
In an experiment, the survey of following four kinds of methods is carried out for the prediction model for the reference capacity resource that training obtains in advance
Examination.
First method: as shown in Fig. 2, core business process number is incremented by by a certain percentage with total business process number, observation
The variation of core business efficiency.
Second method: as shown in figure 3, core business process number and benchmark business number are certain, core business process is observed
Several and relationship between efficiency.
The third method: as shown in figure 4, ignoring other factors, core business process number and relationship between efficiency are only observed.
Fourth method: as shown in figure 5, total business process number is certain, core business process number and relationship between efficiency are observed.
By experiment, as shown in fig. 6, the benchmark treatment effeciency accuracy rate that the embodiment of the present application obtains can reach 95% with
On.
Step S103: by the target service parameter and the benchmark treatment effeciency input practical appearance that training obtains in advance
The prediction model for measuring resource, obtains the actual treatment efficiency of the host.
In the embodiment of the present invention, the prediction model M2 of actual capacity resource be can be based on the training of production environment data.
Specifically, correlation analysis, Data Dimensionality Reduction can be used by acquisition production environment log, the in detail information such as single stream, host configuration
The methods of, extract main several factors in model: load factor (loading condition of description individual process), forecasting efficiency (M1
In prediction result, be integrated into M2 in a manner of the factor here), ticket amount, number of users, peripheral resource proportion (business process
Operation needs the peripheral resource used, and when resource type is consistent, the size of proportion influences whether actual efficiency), foundation
Saturation degree efficiency Model.Using the methods of completion, normalizing, final model is trained using feedforward neural network algorithm.
Exemplary, the model of M2 constitutes as follows:
Wholesale price efficiency (Z2)=saturation degree (Y1)~baseline efficiency (Z1)~ticket amount (Y3)~number of users (Y4)~periphery
Resource matches (Y5).
Functional expression indicates: Z2=M2 (Y1, Z1, Y3, Y4, Y5).
It integrates two models to obtain: Z2=M2 (Y1, M1 (x1, x2, x3), Y3, Y4, Y5).
It is appreciated that since production environment can not arbitrarily adjust configuration, if being added without the result factor of M1 model, M2
Model can only be according to available production real data, and flexibility, prediction scene, accuracy are all poor.The embodiment of the present invention
In, it is correlation factor by the volumetric efficiency under full load under M1 ideal conditions, is integrated into M2 model, in this way in the model of M2
The advantages of M1 model for sufficiently including, and creation data is tightly depended on, compensate for the deficiency of M1 model.
It is exemplary, in experiment, by the prediction model of actual capacity resource, obtain the effect of the actual treatment efficiency of host
As shown in Figure 7, it is seen that global error controls within 0.03.
Step S104: according to the actual treatment efficiency of the host, capacity required in following preset time period is predicted
Resource.
In the embodiment of the present invention, the actual treatment efficiency of host is that the portfolio handled in the host unit time therefore can
It is pre- to obtain the future divided by the actual treatment efficiency of host according to the total traffic that may be handled in the following preset time period
It is prepared in advance so as to reach if the capacity resource needed to configure in the period, guarantee that business is suitable in the future preset time period
The effect freely realized.
Optionally, the capacity resource includes at least one of following resource: host number, the quantity of virtual machine, in
The processing speed and memory size of central processor CPU.
In the embodiment of the present invention, it is contemplated that host number, the quantity of virtual machine, the processing speed of central processor CPU and
Memory size is that the principal element of influence host computing capacity therefore can be by host number, the quantity of virtual machine, centre
The processing speed and memory size for managing device CPU are as needing the capacity resource of Xu Ce to be predicted.
In conclusion in the prediction technique and device of a kind of capacity resource provided in an embodiment of the present invention, it is contemplated that model
A large amount of training data is usually required in training, and production environment generally can not obtain a large amount of creation data, according only to life
The data for producing environment can not usually carry out model training or can not train to obtain accurate prediction model, therefore the embodiment of the present invention
In, using the prediction model of two-part, specifically, the prediction model of reference capacity resource can be instructs in test environment
Experienced model can collect the test data in all kinds of situations in test environment and be trained, therefore, the reference capacity resource
Prediction model can export the benchmark treatment effeciency in more situation；It is predicted by the prediction model of actual capacity resource
When, benchmark treatment effeciency and creation data are used as to the input of the prediction model of capacity resource, then both consider production environment,
Test environment has been taken into account again, and the accuracy of capacity prediction is greatly improved, and does not need manually to participate in, with higher
Forecasting efficiency.Specifically, in the embodiment of the present invention, according to the target service process for the business that predicting strategy obtains running on host
After several and target service parameter, target service process number is first inputted to the prediction mould for the reference capacity resource that training obtains in advance
Type obtains the benchmark treatment effeciency of host, defeated together with target service parameter then using benchmark treatment effeciency as a factor
The prediction model for entering the actual capacity resource that training obtains in advance, obtains the actual treatment efficiency of host, and then can be according to master
The actual treatment efficiency of machine predicts capacity resource required in following preset time period, not only considers production environment, but also take into account
The accuracy of test environment, capacity prediction is greatly improved, and does not need manually to participate in, and prediction with higher is imitated
Rate.
Fig. 8 is the model framework schematic diagram of the prediction technique of capacity resource provided in an embodiment of the present invention.As shown in figure 8,
In the prediction technique of capacity resource provided by the invention:
Including two data models: the prediction model of the prediction model (M1) of reference capacity resource, actual capacity resource
(M2), M1 is obtained by core business process number, basic business process number, the training of total business process number, then by target service
It, can be using the benchmark treatment effeciency of M1 model prediction as the factor, in addition the factor collected by production environment after short range number inputs M1
Parameter: load factor, ticket amount, number of users, memory bank number, training forms model M 2 again, finally according to certain predicting strategy
Obtain prediction result.
On the one hand the embodiment of the present invention based on data, eliminates the interference of artificial experience, the stabilization of the prediction of guarantee
Property, so that the prediction result that can be completely dependent on model when production configuration be made to make the distribution of capacity resource, reduce unnecessary
The wasting of resources, on the other hand, the many-sided impact factor being added in a model allows to all kinds of prominent situations and all kinds of wind
Danger prediction scene is included, when emergency is predicted or faces complex scene, it is only necessary to which the factor input for changing current scene is matched
It sets, just can obtain accurate prediction result, improve the ability for coping with risk in production environment.
In conclusion in the prediction technique and device of a kind of capacity resource provided in an embodiment of the present invention, it is contemplated that model
A large amount of training data is usually required in training, and production environment generally can not obtain a large amount of creation data, according only to life
The data for producing environment can not usually carry out model training or can not train to obtain accurate prediction model, therefore the embodiment of the present invention
In, using the prediction model of two-part, specifically, the prediction model of reference capacity resource can be instructs in test environment
Experienced model can collect the test data in all kinds of situations in test environment and be trained, therefore, the reference capacity resource
Prediction model can export the benchmark treatment effeciency in more situation；It is predicted by the prediction model of actual capacity resource
When, benchmark treatment effeciency and creation data are used as to the input of the prediction model of capacity resource, then both consider production environment,
Test environment has been taken into account again, and the accuracy of capacity prediction is greatly improved, and does not need manually to participate in, with higher
Forecasting efficiency.Specifically, in the embodiment of the present invention, according to the target service process for the business that predicting strategy obtains running on host
After several and target service parameter, target service process number is first inputted to the prediction mould for the reference capacity resource that training obtains in advance
Type obtains the benchmark treatment effeciency of host, defeated together with target service parameter then using benchmark treatment effeciency as a factor
The prediction model for entering the actual capacity resource that training obtains in advance, obtains the actual treatment efficiency of host, and then can be according to master
The actual treatment efficiency of machine predicts capacity resource required in following preset time period, not only considers production environment, but also take into account
The accuracy of test environment, capacity prediction is greatly improved, and does not need manually to participate in, and prediction with higher is imitated
Rate.
Fig. 9 is the structural schematic diagram of one embodiment of prediction meanss of capacity resource provided by the invention.As shown in figure 9, this
The prediction meanss of capacity resource that embodiment provides include:
Module 210 is obtained, the target service process number and mesh of the business for obtaining running on host according to predicting strategy
Mark service parameter；
First input module 220 is provided for the target service process number to be inputted the reference capacity that training obtains in advance
The prediction model in source obtains the benchmark treatment effeciency of the host；
Second input module 230, for training the target service parameter and benchmark treatment effeciency input in advance
The prediction model of obtained actual capacity resource obtains the actual treatment efficiency of the host；
Prediction module 240, for the actual treatment efficiency according to the host, needed for predicting in following preset time period
Capacity resource.
Optionally, the target service parameter includes at least one of following parameter: targeted loads rate, target ticket
Amount, target user's number and targeted peripheral resource proportion；And/or
The capacity resource includes at least one of following resource: host number, the quantity of virtual machine, central processing unit
The processing speed and memory size of CPU.
Optionally, described to obtain mould when the service parameter includes the target ticket amount and target user's number
Block is also used to:
Using autoregressive moving average ARMA algorithm, operation is carried out to history ticket amount and historical user's number, is obtained described
Target ticket amount and target user's number.
Optionally, the module that obtains is also used to:
The prediction model that the initial value of business process number is inputted to the reference capacity resource, obtains the test of the host
Treatment effeciency；The value of the business process number is successively increased according to preset step-length, and inputs the pre- of the reference capacity resource
Model is surveyed, the test processes efficiency of the business process number corresponding host under different value conditions is obtained；According to institute
The test processes efficiency of business process number corresponding host under different value conditions is stated, at the test for determining the host
The value for managing corresponding business process number when efficiency is maximum value is the target service process number.
Optionally, the business process number includes at least one of following process number: average core business process number is put down
Important business process number and average total business process number.
The third aspect of the embodiment of the present invention provides a kind of electronic equipment, comprising: processor, memory and computer journey
Sequence；Wherein, the computer program is stored in the memory, and is configured as being executed by the processor, described
Computer program includes for executing the instruction such as the described in any item methods of aforementioned first aspect.
In conclusion in the prediction technique and device of a kind of capacity resource provided in an embodiment of the present invention, it is contemplated that model
A large amount of training data is usually required in training, and production environment generally can not obtain a large amount of creation data, according only to life
The data for producing environment can not usually carry out model training or can not train to obtain accurate prediction model, therefore the embodiment of the present invention
In, using the prediction model of two-part, specifically, the prediction model of reference capacity resource can be instructs in test environment
Experienced model can collect the test data in all kinds of situations in test environment and be trained, therefore, the reference capacity resource
Prediction model can export the benchmark treatment effeciency in more situation；It is predicted by the prediction model of actual capacity resource
When, benchmark treatment effeciency and creation data are used as to the input of the prediction model of capacity resource, then both consider production environment,
Test environment has been taken into account again, and the accuracy of capacity prediction is greatly improved, and does not need manually to participate in, with higher
Forecasting efficiency.Specifically, in the embodiment of the present invention, according to the target service process for the business that predicting strategy obtains running on host
After several and target service parameter, target service process number is first inputted to the prediction mould for the reference capacity resource that training obtains in advance
Type obtains the benchmark treatment effeciency of host, defeated together with target service parameter then using benchmark treatment effeciency as a factor
The prediction model for entering the actual capacity resource that training obtains in advance, obtains the actual treatment efficiency of host, and then can be according to master
The actual treatment efficiency of machine predicts capacity resource required in following preset time period, not only considers production environment, but also take into account
The accuracy of test environment, capacity prediction is greatly improved, and does not need manually to participate in, and prediction with higher is imitated
Rate.
The prediction meanss for the capacity resource that various embodiments of the present invention provide can be used for executing such as aforementioned each corresponding embodiment
Shown in method, implementation is identical as principle, repeats no more.
The embodiment of the present invention also provides a kind of electronic equipment, comprising: processor, memory and computer program；Wherein,
The computer program is stored in the memory, and is configured as being executed by the processor, the computer journey
Sequence includes the instruction for executing the method as described in any one of previous embodiment.
The embodiment of the present invention also provides a kind of computer readable storage medium, and the computer-readable recording medium storage has
Computer program, the computer program are performed, and realize the method as described in any one of previous embodiment.
Those of ordinary skill in the art will appreciate that: realize that all or part of the steps of above-mentioned each method embodiment can lead to
The relevant hardware of program instruction is crossed to complete.Program above-mentioned can be stored in a computer readable storage medium.The journey
When being executed, execution includes the steps that above-mentioned each method embodiment to sequence；And storage medium above-mentioned include: ROM, RAM, magnetic disk or
The various media that can store program code such as person's CD.
Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention., rather than its limitations；To the greatest extent
Pipe present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: its according to
So be possible to modify the technical solutions described in the foregoing embodiments, or to some or all of the technical features into
Row equivalent replacement；And these are modified or replaceed, various embodiments of the present invention technology that it does not separate the essence of the corresponding technical solution
The range of scheme.
Claims (10)
1. a kind of prediction technique of capacity resource, which is characterized in that the described method includes:
According to the target service process number and target service parameter of the business that predicting strategy obtains running on host；
The prediction model that the target service process number is inputted to the reference capacity resource that training obtains in advance, obtains the host
Benchmark treatment effeciency；
By the prediction of the target service parameter and the benchmark treatment effeciency input actual capacity resource that training obtains in advance
Model obtains the actual treatment efficiency of the host；
According to the actual treatment efficiency of the host, capacity resource required in following preset time period is predicted.
2. the method according to claim 1, wherein the target service parameter include in following parameter at least
One: targeted loads rate, target ticket amount, target user's number and targeted peripheral resource proportion；And/or
The capacity resource includes at least one of following resource: host number, the quantity of virtual machine, central processor CPU
Processing speed and memory size.
3. according to the method described in claim 2, it is characterized in that, when the service parameter includes the target ticket amount and institute
When stating target user's number, according to the target service parameter for the business that predicting strategy obtains running on host, comprising:
Using autoregressive moving average ARMA algorithm, operation is carried out to history ticket amount and historical user's number, obtains the target
Ticket amount and target user's number.
4. method according to claim 1-3, which is characterized in that described to obtain transporting on host according to predicting strategy
The target service process number of capable business, comprising:
The prediction model that the initial value of business process number is inputted to the reference capacity resource, obtains the test processes of the host
Efficiency；
The value of the business process number is successively increased according to preset step-length, and inputs the prediction mould of the reference capacity resource
Type obtains the test processes efficiency of the business process number corresponding host under different value conditions；
According to the test processes efficiency of the business process number corresponding host under different value conditions, the master is determined
The test processes efficiency of machine when being maximum value the value of the corresponding business process number be the target service process number.
5. according to the method described in claim 4, it is characterized in that, the business process number include in following process number at least
One: average core business process number, average important service process number and average total business process number.
6. a kind of prediction meanss of capacity resource characterized by comprising
Module is obtained, target service process number and the target service ginseng of the business for obtaining running on host according to predicting strategy
Number；
First input module, for the target service process number to be inputted to the prediction for the reference capacity resource that training obtains in advance
Model obtains the benchmark treatment effeciency of the host；
Second input module, for the target service parameter and benchmark treatment effeciency input to be trained obtained reality in advance
The prediction model of border capacity resource obtains the actual treatment efficiency of the host；
Prediction module predicts capacity required in following preset time period for the actual treatment efficiency according to the host
Resource.
7. device according to claim 6, which is characterized in that the target service parameter include in following parameter at least
One: targeted loads rate, target ticket amount, target user's number and targeted peripheral resource proportion；And/or
The capacity resource includes at least one of following resource: host number, the quantity of virtual machine, central processor CPU
Processing speed and memory size.
8. device according to claim 7, which is characterized in that when the service parameter includes the target ticket amount and institute
When stating target user's number, the module that obtains is also used to:
Using autoregressive moving average ARMA algorithm, operation is carried out to history ticket amount and historical user's number, obtains the target
Ticket amount and target user's number.
9. according to the described in any item devices of claim 6-8, which is characterized in that the module that obtains is also used to:
The prediction model that the initial value of business process number is inputted to the reference capacity resource, obtains the test processes of the host
Efficiency；The value of the business process number is successively increased according to preset step-length, and inputs the prediction mould of the reference capacity resource
Type obtains the test processes efficiency of the business process number corresponding host under different value conditions；According to the industry
The test processes efficiency of business process number corresponding host under different value conditions determines the test processes effect of the host
The value of the corresponding business process number is the target service process number when rate is maximum value.
10. device according to claim 9, which is characterized in that the business process number include in following process number extremely
It is one few: average core business process number, average important service process number and average total business process number.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910729638.1A CN110445939B (en) | 2019-08-08 | 2019-08-08 | Capacity resource prediction method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910729638.1A CN110445939B (en) | 2019-08-08 | 2019-08-08 | Capacity resource prediction method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110445939A true CN110445939A (en) | 2019-11-12 |
| CN110445939B CN110445939B (en) | 2021-03-30 |
Family
ID=68433984
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910729638.1A Active CN110445939B (en) | 2019-08-08 | 2019-08-08 | Capacity resource prediction method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110445939B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111275268A (en) * | 2020-02-27 | 2020-06-12 | 中国联合网络通信集团有限公司 | Pricing process efficiency prediction method, device, equipment and storage medium |
| CN111581070A (en) * | 2020-05-07 | 2020-08-25 | 拉扎斯网络科技（上海）有限公司 | Capacity determination method and device, electronic equipment and computer readable storage medium |
| CN113742187A (en) * | 2021-08-05 | 2021-12-03 | 中移（杭州）信息技术有限公司 | Capacity prediction method, device, equipment and storage medium of application system |
| CN114518988A (en) * | 2022-02-10 | 2022-05-20 | 中国光大银行股份有限公司 | Resource capacity system, method of controlling the same, and computer-readable storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101304590A (en) * | 2008-04-14 | 2008-11-12 | 中国网络通信集团公司 | Apparatus and method for determining wireless network capacitance of mobile communication network |
| CN103257896A (en) * | 2013-01-31 | 2013-08-21 | 南京理工大学连云港研究院 | Max-D job scheduling method under cloud environment |
| CN105471655A (en) * | 2015-12-09 | 2016-04-06 | 中国联合网络通信集团有限公司 | Method and device for determining power-on or power-off state of physical device in virtual cluster |
| US20170255240A1 (en) * | 2016-03-03 | 2017-09-07 | Lenovo Enterprise Solutions (Singapore) Pte. Ltd. | Energy efficient workload placement management based on observed server efficiency measurements |
| CN107294764A (en) * | 2017-04-26 | 2017-10-24 | 中国科学院信息工程研究所 | Intelligent supervision method and intelligent monitoring system |
| CN108932182A (en) * | 2018-07-06 | 2018-12-04 | 许继集团有限公司 | A kind of messaging bus performance test methods and system |
| CN109714395A (en) * | 2018-12-10 | 2019-05-03 | 平安科技（深圳）有限公司 | Cloud platform resource uses prediction technique and terminal device |
-
2019
- 2019-08-08 CN CN201910729638.1A patent/CN110445939B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101304590A (en) * | 2008-04-14 | 2008-11-12 | 中国网络通信集团公司 | Apparatus and method for determining wireless network capacitance of mobile communication network |
| CN103257896A (en) * | 2013-01-31 | 2013-08-21 | 南京理工大学连云港研究院 | Max-D job scheduling method under cloud environment |
| CN105471655A (en) * | 2015-12-09 | 2016-04-06 | 中国联合网络通信集团有限公司 | Method and device for determining power-on or power-off state of physical device in virtual cluster |
| US20170255240A1 (en) * | 2016-03-03 | 2017-09-07 | Lenovo Enterprise Solutions (Singapore) Pte. Ltd. | Energy efficient workload placement management based on observed server efficiency measurements |
| CN107294764A (en) * | 2017-04-26 | 2017-10-24 | 中国科学院信息工程研究所 | Intelligent supervision method and intelligent monitoring system |
| CN108932182A (en) * | 2018-07-06 | 2018-12-04 | 许继集团有限公司 | A kind of messaging bus performance test methods and system |
| CN109714395A (en) * | 2018-12-10 | 2019-05-03 | 平安科技（深圳）有限公司 | Cloud platform resource uses prediction technique and terminal device |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111275268A (en) * | 2020-02-27 | 2020-06-12 | 中国联合网络通信集团有限公司 | Pricing process efficiency prediction method, device, equipment and storage medium |
| CN111581070A (en) * | 2020-05-07 | 2020-08-25 | 拉扎斯网络科技（上海）有限公司 | Capacity determination method and device, electronic equipment and computer readable storage medium |
| CN111581070B (en) * | 2020-05-07 | 2023-08-18 | 拉扎斯网络科技（上海）有限公司 | Capacity determination method, device, electronic equipment and computer readable storage medium |
| CN113742187A (en) * | 2021-08-05 | 2021-12-03 | 中移（杭州）信息技术有限公司 | Capacity prediction method, device, equipment and storage medium of application system |
| CN114518988A (en) * | 2022-02-10 | 2022-05-20 | 中国光大银行股份有限公司 | Resource capacity system, method of controlling the same, and computer-readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110445939B (en) | 2021-03-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110445939A (en) | The prediction technique and device of capacity resource | |
| CN109214719A (en) | A kind of system and method for the marketing inspection analysis based on artificial intelligence | |
| CN105678398A (en) | Power load forecasting method based on big data technology, and research and application system based on method | |
| CN106716454A (en) | Utilizing machine learning to identify non-technical loss | |
| CN110162445A (en) | The host health assessment method and device of Intrusion Detection based on host log and performance indicator | |
| CN112465011A (en) | Project risk prediction method and system based on project research and development process | |
| WO2017071369A1 (en) | Method and device for predicting user unsubscription | |
| CN107506872A (en) | A kind of residential block part throttle characteristics and the Categorical research method of model prediction | |
| CN104680328A (en) | Power grid construction quality monitoring method based on client perception values | |
| Zhang et al. | Short-term passenger flow forecast of rail transit station based on MIC feature selection and ST-LightGBM considering transfer passenger flow | |
| CN113706291A (en) | Fraud risk prediction method, device, equipment and storage medium | |
| CN116050947A (en) | Method, device, computer equipment and storage medium for evaluating vehicle dispatching effect | |
| CN109446278A (en) | A kind of big data management platform system based on block chain | |
| CN110635474B (en) | Power grid dynamic trajectory trend prediction method based on long-term and short-term memory network | |
| CN114912854B (en) | Subway train operation adjusting method and device, electronic equipment and storage medium | |
| CN115792505B (en) | Rail transit power supply management method and system | |
| CN115600824B (en) | Carbon emission early warning method and device, storage medium and electronic equipment | |
| CN107093018A (en) | Communication engineering project information method for visualizing and device based on health model | |
| CN116128197A (en) | Intelligent airport management system and method | |
| CN115471215A (en) | Business process processing method and device | |
| CN109376898A (en) | A kind of coulometric analysis prediction technique based on inside and outside Index Influence | |
| CN114742412A (en) | Software technology service system and method | |
| CN106980675A (en) | A kind of efficient bridge structure health early warning system | |
| CN111105148A (en) | Method and device for evaluating job leaving probability and computer readable storage medium | |
| CN109063936A (en) | A kind of method and device for establishing wind power prediction model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |