CN109039831A - A kind of load detection method and device - Google Patents
A kind of load detection method and device Download PDFInfo
- Publication number
- CN109039831A CN109039831A CN201811108240.8A CN201811108240A CN109039831A CN 109039831 A CN109039831 A CN 109039831A CN 201811108240 A CN201811108240 A CN 201811108240A CN 109039831 A CN109039831 A CN 109039831A
- Authority
- CN
- China
- Prior art keywords
- load
- data
- detected
- load data
- memory
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
- H04L43/0805—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters by checking availability
- H04L43/0817—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters by checking availability by checking functioning
Landscapes
- Engineering & Computer Science (AREA)
- Environmental & Geological Engineering (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Debugging And Monitoring (AREA)
Abstract
This application discloses a kind of load detection method and devices, it include: the load data for obtaining object to be detected, the object to be detected is data center, server, any one in terminal, and the load data of object to be detected includes the cpu load data of object to be detected, memory load data, memory I/O load data, at least two data in communication I/O load data;The load data of obtained object to be detected is input in the load detecting model pre-established, in order to which load detecting model can export the load value of object to be detected, which characterizes the loading level of object to be detected.It can be seen that, due to when treating test object progress load detecting, combining influences the various because usually being detected of data center's load, obtained load detecting result more can accurately reflect out the real load situation of object to be detected, to realize the accurate detection for treating test object real load situation.
Description
Technical field
This application involves load detecting technical fields, more particularly to a kind of load detection method and device.
Background technique
With the development of cloud computing technology, cloud data center operating system gradually forms and is committed to practice, raw in society
Increasingly important role is played in production and sphere of life.And while data center provides high-quality service, data
The load at center is also to need issues that need special attention.Reasonable, the efficient detection to data central loading situation are realized, for behaviour
The whole high availability and validity for making system play very crucial effect.
In the existing load detecting scheme to data center, usually there is no the various influences loads on data center
Factor is adequately analyzed, so that the true loading level of the loading level of the data center detected and data center exists
Large error there are problems that data center's real load situation can not be accurately reflected.
Summary of the invention
The embodiment of the present application provides a kind of load detection method and device, accurately detects such as data center to realize
The real load situation of object to be detected.
In a first aspect, the embodiment of the present application provides a kind of load detection method, which comprises
The load data of object to be detected is obtained, the object to be detected is for terminal, server and including server
Any one in data center, the load data of the object to be detected include the object to be detected cpu load data,
Memory load data, communicates at least two data in I/O load data at memory I/O load data;
The load data of the object to be detected is input to the load for being directed to the object to be detected pre-established
In detection model, in order to which the load detecting model exports the load value of the object to be detected, the load value is characterized
The loading level of the object to be detected;
Wherein, the load detecting model is in advance using the historic load of the object to be detected as input, standard
Load value is trained to obtain as output, and the historic load of the object to be detected includes going through for the object to be detected
History cpu load data, history memory load data, history memory I/O load data, at least two in historical communication I/O load data
Kind historical data.
In some possible embodiments, the cpu load data are specially cpu load ratio, the memory load
Data are specially memory load percentage, and the memory I/O load data are specially memory I/O load ratio, the communication I/O load
Data are specially to communicate I/O load ratio.
In some possible embodiments, if the object to be detected is data center, the acquisition is to be detected right
The load data of elephant, comprising:
Obtain the load data of each server of the data center;
According to being in advance the weight of server each in data center configuration, the load of the data center is calculated
Data;
Wherein, the sum of corresponding weight of Servers-all is 1 in the data center.
In some possible embodiments, the load value of the object to be detected, it is described to be detected right to be specifically as follows
The loading level grade of elephant or the numerical value for characterizing loading level.
In some possible embodiments, the method also includes:
The load value of the object to be detected is presented.
In some possible embodiments, the load detecting model is self organizing neural network model.
Second aspect, the embodiment of the present application also provides a kind of Weight detector, described device includes:
Obtain module, for obtaining the load data of object to be detected, the object to be detected be terminal, server and
Any one in data center including server, the load data of the object to be detected includes the object to be detected
Cpu load data, memory I/O load data, communicate at least two data in I/O load data at memory load data;
Input module, it is described to be checked for the load data of the object to be detected to be input to being directed to of pre-establishing
It surveys in the load detecting model of object, it is described in order to which the load detecting model exports the load value of the object to be detected
Load value characterizes the loading level of the object to be detected;
Wherein, the load detecting model is in advance using the historic load of the object to be detected as input, standard
Load value is trained to obtain as output, and the historic load of the object to be detected includes going through for the object to be detected
History cpu load data, history memory load data, history memory I/O load data, at least two in historical communication I/O load data
Kind historical data.
In some possible embodiments, if the object to be detected is data center, the acquisition module, packet
It includes:
Acquiring unit, for obtaining the load data of each server of the data center;
Computing unit, for calculating described according to being in advance the weight of server configuration each in the data center
The load data of data center;
Wherein, the sum of corresponding weight of Servers-all is 1 in the data center.
In some possible embodiments, described device further include:
Module is presented, for rendering the load value of the object to be detected.
In some possible embodiments, the load detecting model is self organizing neural network model.
In some possible embodiments, the cpu load data are specially cpu load ratio, the memory load
Data are specially memory load percentage, and the memory I/O load data are specially memory I/O load ratio, the communication I/O load
Data are specially to communicate I/O load ratio.
In some possible embodiments, the load value of the object to be detected, it is described to be detected right to be specifically as follows
The loading level grade of elephant or the numerical value for characterizing loading level.
In the above-mentioned implementation of the embodiment of the present application, influenced by combining such as data central loading waiting test object
Multiple factors detect the actual loading of object to be detected so that realizing accurately detect that the true of object to be detected is born
Load situation.Specifically, obtaining the load data of object to be detected, which not only can be data center, can also be with
It is server or terminal etc., the load data of object to be detected includes the factor that many aspects influence load, specifically be can wrap
It includes the cpu load data of object to be detected, memory load data, memory I/O load data, communicate at least two in I/O load data
Kind data;It is then possible to the load data of obtained object to be detected is input in the load detecting model pre-established, with
The load value of object to be detected can be exported convenient for load detecting model, which characterizes the load journey of object to be detected
Degree, wherein the load detecting model in advance using the historic load of object to be detected as input, using standard termination value as
Output is trained to obtain, and the historic load of the object to be detected includes the history cpu load data of object to be detected, goes through
History memory load data, history memory I/O load data, at least two historical datas in historical communication I/O load data.As it can be seen that
Since when treating test object progress load detecting, combining influences the various because usually being examined of data center's load
Survey, compared to it is existing be based only upon it is one-sided because of the usually technical solution of progress load detecting for, obtained load detecting
As a result it more can accurately reflect out the real load situation of object to be detected, treat test object real load situation to realize
Accurate detection.
Detailed description of the invention
In order to more clearly explain the technical solutions in the embodiments of the present application, make required in being described below to embodiment
Attached drawing is briefly described, it should be apparent that, the accompanying drawings in the following description is only some implementations as described in this application
Example, for those of ordinary skill in the art, is also possible to obtain other drawings based on these drawings.
Fig. 1 is an exemplary application schematic diagram of a scenario in the embodiment of the present application;
Fig. 2 is a kind of flow diagram of load detection method in the embodiment of the present application;
Fig. 3 is a kind of structural schematic diagram of Weight detector in the embodiment of the present application.
Specific embodiment
Inventor it has been investigated that, it is existing to data center carry out load detecting scheme in, usually only considered shadow
This one-sided factor of the cpu load situation of data center's load is rung, and in practical application scene, there is also other factors also can
Influence data center's load.For example, for the communication IO in data center, if having aggravated to communicate in data center suddenly
The load of IO then can also aggravate the actual loading of data center.Therefore, it is based only upon the one-sided factor for influencing data center's load
Come the load state at detection data center, the loading level for the data center that obtained testing result is characterized and data center
Actual loading level difference is larger, so that data center's real load situation can not be accurately reflected.
Based on this, the embodiment of the present application provides a kind of load detection method, influences such as data central loading by combining
The Multiple factors of test object are waited to detect the actual loading of object to be detected so that realize accurately detect it is to be detected
The real load situation of object.Specifically, obtaining the load data of object to be detected, which not only can be data
Center is also possible to server or terminal etc., and the load data of object to be detected includes the factor that many aspects influence load,
It can specifically include cpu load data, memory load data, the memory I/O load data, communication I/O load number of object to be detected
At least two data in;It is then possible to which the load data of obtained object to be detected to be input to the load inspection pre-established
It surveys in model, in order to which load detecting model can export the load value of object to be detected, which characterizes to be detected right
The loading level of elephant, wherein the load detecting model is in advance using the historic load of object to be detected as input, with standard
Load value is trained to obtain as output, and the historic load of the object to be detected includes the history CPU of object to be detected
Load data, history memory load data, history memory I/O load data, at least two history in historical communication I/O load data
Data.As it can be seen that due to treat test object carry out load detecting when, combine influence data center load it is various because
Usually detected, compared to it is existing be based only upon it is one-sided because of the usually technical solution of progress load detecting for, it is acquired
Load detecting result more can accurately reflect out the real load situation of object to be detected, to realize that treat test object true
The accurate detection of load state.
For example, the embodiment of the present application can be applied to exemplary application scene as shown in Figure 1.In the application scenarios
In, the cpu load data at combined data center, memory load data, memory I/O load data, communication I/O load data these four
Load effect factor detects the load of data center, specifically, user can be by terminal 101 to including multiple services
The data center 102 of device sends the load detecting request for being directed to data center 102, and data center 102 receives load inspection
Request is surveyed, obtains the load data in each server, namely obtain the load data of data center 102, the load data packet
Cpu load data, memory load data, memory I/O load data, the communication I/O load data for including each server then will
The load data of data center 102 is input in the load detecting model pre-established, in order to which load detecting model can be defeated
The load value of data center 102 out, the load value are capable of the loading level at characterize data center 102, wherein are directed in data
The load detecting model of the heart 102 in advance using the historic load of data center as input, standard termination value as export into
Row training obtains, and the historic load of the data center 102 includes the history CPU of each server in data center 102 negative
Carry data, history memory load data, history memory I/O load data, historical communication I/O load data;Data center 102 is obtaining
After the load value for obtaining the output of load detecting model, which can be sent to terminal 101, by terminal 101 in the aobvious of terminal
It is presented to the user in display screen, in order to which user can know according to the load value real load situation of data center.
It is understood that above-mentioned scene is only a Sample Scenario provided by the embodiments of the present application, the embodiment of the present application
It is not limited to this scene.For example, being also possible to carry out load detecting etc. to certain particular server in data center.
In order to make the above objects, features, and advantages of the present application more apparent, below in conjunction with attached drawing to this Shen
Please the various non-limiting implementations in embodiment illustrate.Obviously, described embodiment is the application one
Section Example, instead of all the embodiments.Based on the embodiment in the application, those of ordinary skill in the art are not doing
All other embodiment obtained under the premise of creative work out, shall fall in the protection scope of this application.
Referring to Fig.2, Fig. 2 shows the flow diagram of load detection method a kind of in the embodiment of the present application, this method tool
Body may include:
S201: obtaining the load data of object to be detected, and the object to be detected is for terminal, server and including server
Data center in any one, the load data of the object to be detected includes the cpu load data of object to be detected, memory
Load data, communicates at least two data in I/O load data at memory I/O load data.
In order to more accurately reflect the real load situation of object to be detected, in the present embodiment, shadow has been fully considered
The factor for ringing many aspects of object load to be detected, specifically analyzes the load data of the factor of many aspects.Cause
This, acquired load data can include a variety of load datas, and in a kind of example, the load data of object to be detected can be with
It simultaneously include cpu load data, memory load data, memory I/O load data, the communication I/O load data four of object to be detected
Kind data.Wherein, cpu load data are the data of the load state of CPU in operating status in characterization object to be detected, interior
Depositing load data is the how many data of the memory used for characterizing object to be detected, memory I/O load data be characterize it is to be detected
The data of the service condition of the memory IO of object, communication I/O load data are the service condition of the communication IO of characterization object to be detected
Data.
In a kind of illustrative embodiment, the load data of object to be detected specifically can be characterization load state
Load percentage can use 30%, the numerical value of 2/3 this load percentage is as load data by taking memory load data as an example.
Then, the load data of object to be detected, as cpu load ratio, memory load percentage, memory I/O load ratio and communication IO
At least two ratio datas in load percentage.
It is noted that object to be detected is not limited to data center in the present embodiment, it is also possible to other objects,
That is the technical solution of the present embodiment be not only can to data center carry out load detecting, can also be directed to some server or
Some terminal of person carries out load detecting.When object to be detected is some server or terminal, acquired load data is only
Data including a server or a terminal, and when object to be detected is data center, since data center is usual
It can include multiple servers, therefore, as an example, when obtaining the load data of data center, can first obtain data
The load data of each server in center calculates then according to being in advance the weight of server configuration each in data center
The load data of data center out, certainly, the sum of weight corresponding to Servers-all is 1 in data center.
For the cpu load data instance in load data to obtain data center, it is assumed that include 10 in data center
A server, then can first obtain the load percentage of this 10 servers, respectively 30%, 50%, 15%, 47%, 28%,
63%, 58%, 43%, 47%, 65%, and be in advance this 10 respectively arranged weight proportions of server be 0.1,0.15,
0.11,0.08,0.1,0.1,0.08,0.08,0.09,0.11, then calculated data center load percentage be 44.47%
(i.e. 30%*0.1+50%*0.15+15%*0.11+47%*0.08+28%*0.1+63%*0.1+58%*0 .08+43%*
It 0.08+47%*0.09+65%*0.11), then can be by the 44.47% cpu load data as data center.
In practical application, monitoring module can be set in object to be detected, which can be used for monitoring to be checked
It surveys the cpu load data of object, memory load data, memory I/O load data, communicate at least two data in I/O load data,
And store the load data that monitoring obtains, in this way, can be rung when the load detecting of test object is treated in user's triggering
Required load data directly should be obtained from the load data of the object to be detected of storage in the trigger action of user.
S202: the load data of object to be detected is input to the load detecting for being directed to object to be detected pre-established
In model, in order to which the load detecting model can export the load value of object to be detected, which characterizes to be detected right
The loading level of elephant, wherein for the load detecting model in advance using the historic load of object to be detected as input, standard is negative
Load value is trained to obtain as output, and the historic load of object to be detected includes the history cpu load of object to be detected
Data, history memory load data, history memory I/O load data, at least two history numbers in historical communication I/O load data
According to.
It is to treat test object using load detecting model to carry out load detecting in the present embodiment.Specifically, obtaining
After the load data of object to be detected, which can be input to and be directed to the negative of model foundation to be detected in advance
It carries in detection model, in this way, load detecting model can be exported to be checked based on the load data after operation load plus surveying model
The load value of object is surveyed, the load value exported can characterize the real load degree of object to be detected.
As an example it is assumed that the load data of object to be detected includes cpu load data, memory load data, memory IO
Load data, communication I/O load data, and load data is showed in the form of load percentage, then can measure load inspection
It surveys in model and inputs four dimensional vector P={ c, m, s, n }, wherein c is cpu load ratio, and m is memory load percentage, and s is memory
I/O load ratio, n are communication I/O load ratio.Four dimensional vector P is input to the load inspection of the object to be detected pre-established
After surveying model, which can export the load state that can characterize object to be detected based on the four dimensional vectors P
Load value.
In some possible embodiments, the load value of output described in load detecting model can be characterization load journey
The specific value or loading level grade of degree.For example, load value can show as 50%, 70% this specific value,
Can also showing as this loading level grade such as " health load ", " general load ", " dangerous load ", (loading level successively adds
It is deep).
It should be noted that treat test object carry out load detecting when, used load detecting model be based on
The model that the object to be detected is trained.In a kind of example, load detecting model specifically can be self-organizing feature map
Network model, correspondingly, deep learning algorithm can be advanced with to train to obtain the load detecting model.It is to be checked to this in advance
It is using the historic load of the object to be detected as input, standard termination value is come as output when survey object is trained
It is trained.Wherein, historic load, the as previous load data of the object to be detected, including this is to be detected
History cpu load data, history memory load data, the history memory I/O load data, historical communication I/O load data of object
In at least two historical datas;And standard termination value, it is the value based on determined by historic load as input, it can be pre-
First set by technical staff.
In practical application, after the load value for obtaining characterizing object load situation to be detected, can also be in by the load value
User (such as technical staff) now is given, in this way, the load state that user can treat test object based on the load value is commented
Sentence.As an example, when load value is presented, it can be and directly presented, which can also be converted into accordingly
Animation or graphic form presented, to obtain that effect preferably is presented.
It is noted that being treated in the illustrative embodiments that test object carries out load detecting a kind of, Ke Yichi
Continuous test object for the treatment of carries out load detecting, i.e., persistently obtains the load data of object to be detected, and be input to corresponding
In load detecting model, to obtain object to be detected continuous load value whithin a period of time.When specific implementation, can be with
One lesser time cycle periodically obtains load data and is input in load detecting model.In this way, being based on the load
Value, can embody the load situation of change of the object to be detected whithin a period of time.It, can be with for example, in application scenes
The load situation of change of the data center of different time sections (such as day and night) in detection one day.In this way, if based in data
The heart determines in the period in night that the load of data center is lower, then can close in intraday load situation of change
Closing part sub-server is to save energy consumption etc..
And another kind treat test object carry out load detecting illustrative embodiments in, be also possible to by user Lai
The load detecting of test object is treated in triggering, i.e., ought detect the presence of user and perform and treat the triggering that test object is detected
When operation, the trigger action can be responded, generates the detection instruction of starting load detection, to start to hold based on the detection instruction
Row treats the process (method flow i.e. shown in Fig. 2) that test object carries out load detecting.
In the present embodiment, by conjunction with influence as data central loading wait test object Multiple factors come detect to
The actual loading of test object, so that realizing the real load situation for accurately detecting object to be detected.Specifically, obtain to
The load data of test object, the object to be detected not only can be data center, be also possible to server or terminal etc., to
The load data of test object includes the factor that many aspects influence load, can specifically include the cpu load of object to be detected
Data, memory I/O load data, communicate at least two data in I/O load data at memory load data;It is then possible to will obtain
The load data of object to be detected be input in the load detecting model pre-established, in order to which load detecting model can be defeated
The load value of object to be detected out, the load value characterize the loading level of object to be detected, wherein the load detecting model is pre-
It first is trained to obtain as input, using standard termination value as output using the historic load of object to be detected, this is to be checked
The historic load for surveying object includes the history cpu load data of object to be detected, history memory load data, history memory
At least two historical datas in I/O load data, historical communication I/O load data.As it can be seen that due to being born treating test object
When carrying detection, combining influences the various because usually being detected of data center's load, is based only upon list compared to existing
For technical solution of the aspect because usually carrying out load detecting, obtained load detecting result more can accurately reflect out to be detected
The real load situation of object, to realize the accurate detection for treating test object real load situation.
In addition, the embodiment of the present application also provides a kind of Weight detectors.The application is shown referring to Fig. 3, Fig. 3 to implement
A kind of structural schematic diagram of Weight detector in example, described device 300 include:
Module 301 is obtained, for obtaining the load data of object to be detected, the object to be detected is terminal, server
And any one in the data center including server, the load data of the object to be detected include described to be detected right
The cpu load data of elephant, memory I/O load data, communicate at least two data in I/O load data at memory load data;
Input module 302, it is described for the load data of the object to be detected to be input to being directed to of pre-establishing
In the load detecting model of object to be detected, in order to which the load detecting model exports the load value of the object to be detected,
The load value characterizes the loading level of the object to be detected;
Wherein, the load detecting model is in advance using the historic load of the object to be detected as input, standard
Load value is trained to obtain as output, and the historic load of the object to be detected includes going through for the object to be detected
History cpu load data, history memory load data, history memory I/O load data, at least two in historical communication I/O load data
Kind historical data.
In some possible embodiments, if the object to be detected be data center, the acquisition module 301,
Include:
Acquiring unit, for obtaining the load data of each server of the data center;
Computing unit, for calculating described according to being in advance the weight of server configuration each in the data center
The load data of data center;
Wherein, the sum of corresponding weight of Servers-all is 1 in the data center.
In some possible embodiments, described device 300 further include:
Module is presented, for rendering the load value of the object to be detected.
In some possible embodiments, the load detecting model is self organizing neural network model.
In some possible embodiments, the cpu load data are specially cpu load ratio, the memory load
Data are specially memory load percentage, and the memory I/O load data are specially memory I/O load ratio, the communication I/O load
Data are specially to communicate I/O load ratio.
In some possible embodiments, the load value of the object to be detected, it is described to be detected right to be specifically as follows
The loading level grade of elephant or the numerical value for characterizing loading level.
In the present embodiment, it is seen then that since when treating test object progress load detecting, combining, which influences data center, is born
What is carried is various because usually being detected, and is based only upon the one-sided technical side because usually carrying out load detecting compared to existing
For case, obtained load detecting result more can accurately reflect out the real load situation of object to be detected, thus realization pair
The accurate detection of object real load situation to be detected.
As seen through the above description of the embodiments, those skilled in the art can be understood that above-mentioned implementation
All or part of the steps in example method can add the mode of general hardware platform to realize by software.Based on this understanding,
The technical solution of the application can be embodied in the form of software products, which can store is situated between in storage
In matter, such as read-only memory (English: read-only memory, ROM)/RAM, magnetic disk, CD etc., including some instructions to
So that a computer equipment (can be the network communication equipments such as personal computer, server, or router) executes
Method described in certain parts of each embodiment of the application or embodiment.
All the embodiments in this specification are described in a progressive manner, same and similar portion between each embodiment
Dividing may refer to each other, and each embodiment focuses on the differences from other embodiments.Especially for method reality
For applying example and apparatus embodiments, since it is substantially similar to system embodiment, so describe fairly simple, related place ginseng
See the part explanation of system embodiment.Equipment and system embodiment described above is only schematical, wherein making
It may or may not be physically separated for the module of separate part description, the component shown as module can be
Or it may not be physical module, it can it is in one place, or may be distributed over multiple network units.It can be with
Some or all of the modules therein is selected to achieve the purpose of the solution of this embodiment according to the actual needs.The common skill in this field
Art personnel can understand and implement without creative efforts.
The above is only the illustrative embodiment of the application, is not intended to limit the protection scope of the application.
Claims (10)
1. a kind of load detection method, which is characterized in that the described method includes:
The load data of object to be detected is obtained, the object to be detected is terminal, server and the data including server
Any one in center, the load data of the object to be detected includes the cpu load data of the object to be detected, memory
Load data, communicates at least two data in I/O load data at memory I/O load data;
The load data of the object to be detected is input to the load detecting for being directed to the object to be detected pre-established
In model, in order to which the load detecting model exports the load value of the object to be detected, the load value characterizes described
The loading level of object to be detected;
Wherein, the load detecting model is in advance using the historic load of the object to be detected as input, standard termination
Value is trained to obtain as output, and the historic load of the object to be detected includes the history of the object to be detected
Cpu load data, history memory load data, history memory I/O load data, at least two in historical communication I/O load data
Historical data.
2. the method according to claim 1, wherein the cpu load data are specially cpu load ratio, institute
Stating memory load data is specially memory load percentage, and the memory I/O load data are specially memory I/O load ratio, described
Communicating I/O load data is specially to communicate I/O load ratio.
3. described to obtain according to the method described in claim 2, it is characterized in that, if the object to be detected is data center
Take the load data of object to be detected, comprising:
Obtain the load data of each server of the data center;
According to being in advance the weight of server each in data center configuration, the load number of the data center is calculated
According to;
Wherein, the sum of corresponding weight of Servers-all is 1 in the data center.
4. the method according to claim 1, wherein the load value of the object to be detected, is specifically as follows institute
It states the loading level grade of object to be detected or characterizes the numerical value of loading level.
5. the method according to claim 1, wherein the method also includes:
The load value of the object to be detected is presented.
6. the method according to claim 1, wherein the load detecting model is self organizing neural network mould
Type.
7. a kind of Weight detector, which is characterized in that described device includes:
Obtain module, for obtaining the load data of object to be detected, the object to be detected be terminal, server and including
Any one in the data center of server, the load data of the object to be detected includes the CPU of the object to be detected
Load data, memory I/O load data, communicates at least two data in I/O load data at memory load data;
Input module, it is described to be detected right for the load data of the object to be detected to be input to being directed to of pre-establishing
In the load detecting model of elephant, in order to which the load detecting model exports the load value of the object to be detected, the load
Value characterizes the loading level of the object to be detected;
Wherein, the load detecting model is in advance using the historic load of the object to be detected as input, standard termination
Value is trained to obtain as output, and the historic load of the object to be detected includes the history of the object to be detected
Cpu load data, history memory load data, history memory I/O load data, at least two in historical communication I/O load data
Historical data.
8. device according to claim 7, which is characterized in that described to obtain if the object to be detected is data center
Modulus block, comprising:
Acquiring unit, for obtaining the load data of each server of the data center;
Computing unit, for calculating the data according to being in advance the weight of server configuration each in the data center
The load data at center;
Wherein, the sum of corresponding weight of Servers-all is 1 in the data center.
9. device according to claim 7, which is characterized in that described device further include:
Module is presented, for rendering the load value of the object to be detected.
10. device according to claim 7, which is characterized in that the load detecting model is self organizing neural network mould
Type.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811108240.8A CN109039831A (en) | 2018-09-21 | 2018-09-21 | A kind of load detection method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811108240.8A CN109039831A (en) | 2018-09-21 | 2018-09-21 | A kind of load detection method and device |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109039831A true CN109039831A (en) | 2018-12-18 |
Family
ID=64617936
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811108240.8A Pending CN109039831A (en) | 2018-09-21 | 2018-09-21 | A kind of load detection method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109039831A (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104503847A (en) * | 2015-01-22 | 2015-04-08 | 浪潮(北京)电子信息产业有限公司 | Data center energy saving method and device |
CN105260794A (en) * | 2015-10-12 | 2016-01-20 | 上海交通大学 | Load predicting method of cloud data center |
CN106293944A (en) * | 2016-08-11 | 2017-01-04 | 上海交通大学 | System and optimization method is accessed based on nonuniformity I/O under virtualization multi-core environment |
CN107196869A (en) * | 2017-07-14 | 2017-09-22 | 北京知道创宇信息技术有限公司 | The adaptive load balancing method, apparatus and system of Intrusion Detection based on host actual loading |
CN108021441A (en) * | 2016-11-03 | 2018-05-11 | 中国移动通信集团广东有限公司 | A kind of resources of virtual machine collocation method and device based on cloud computing |
-
2018
- 2018-09-21 CN CN201811108240.8A patent/CN109039831A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104503847A (en) * | 2015-01-22 | 2015-04-08 | 浪潮(北京)电子信息产业有限公司 | Data center energy saving method and device |
CN105260794A (en) * | 2015-10-12 | 2016-01-20 | 上海交通大学 | Load predicting method of cloud data center |
CN106293944A (en) * | 2016-08-11 | 2017-01-04 | 上海交通大学 | System and optimization method is accessed based on nonuniformity I/O under virtualization multi-core environment |
CN108021441A (en) * | 2016-11-03 | 2018-05-11 | 中国移动通信集团广东有限公司 | A kind of resources of virtual machine collocation method and device based on cloud computing |
CN107196869A (en) * | 2017-07-14 | 2017-09-22 | 北京知道创宇信息技术有限公司 | The adaptive load balancing method, apparatus and system of Intrusion Detection based on host actual loading |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108563548A (en) | Method for detecting abnormality and device | |
CN108318129A (en) | The true and false discriminating method of bridge structure modal parameter and terminal device | |
Chen et al. | Fault diagnosis method for machinery in unsteady operating condition by instantaneous power spectrum and genetic programming | |
CN106780121A (en) | A kind of multiplexing electric abnormality recognition methods based on power load pattern analysis | |
CN108926338A (en) | Heart rate prediction technique and device based on deep learning | |
CN106502889B (en) | The method and apparatus for predicting cloud software performance | |
CN110147745A (en) | A kind of key frame of video detection method and device | |
CN107491811A (en) | Method and system and neural network processor for accelerans network processing unit | |
CN108898261A (en) | A kind of air quality monitoring method and system based on environmentally friendly big data | |
CN110378346A (en) | Establish the method, apparatus, equipment and computer storage medium of Text region model | |
CN106445754A (en) | Method and system for inspecting cluster health status and cluster server | |
CN108877783A (en) | The method and apparatus for determining the audio types of audio data | |
CN108427941A (en) | Method, method for detecting human face and device for generating Face datection model | |
CN106973039A (en) | A kind of network security situation awareness model training method and device based on information fusion technology | |
CN109740508A (en) | A kind of image processing method and nerve network system based on nerve network system | |
CN107421763A (en) | A kind of equipment fault detection method and device | |
CN113655348A (en) | Power equipment partial discharge fault diagnosis method based on deep twin network, system terminal and readable storage medium | |
CN110070076A (en) | Method and apparatus for choosing trained sample | |
CN110445939A (en) | The prediction technique and device of capacity resource | |
CN109815789A (en) | Real-time multiple dimensioned method for detecting human face and system and relevant device on CPU | |
CN107545904A (en) | A kind of audio-frequency detection and device | |
CN110503181A (en) | Method and apparatus for generating multilayer neural network | |
CN107066364A (en) | Method for testing performance and device | |
CN109039831A (en) | A kind of load detection method and device | |
CN112966547A (en) | Neural network-based gas field abnormal behavior recognition early warning method, system, terminal and storage medium |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181218 |