CN109918205A - A kind of edge device dispatching method, system, device and computer storage medium - Google Patents
A kind of edge device dispatching method, system, device and computer storage medium Download PDFInfo
- Publication number
- CN109918205A CN109918205A CN201910229314.1A CN201910229314A CN109918205A CN 109918205 A CN109918205 A CN 109918205A CN 201910229314 A CN201910229314 A CN 201910229314A CN 109918205 A CN109918205 A CN 109918205A
- Authority
- CN
- China
- Prior art keywords
- edge device
- characteristic information
- training
- edge
- matching score
- 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
- 238000000034 method Methods 0.000 title claims abstract description 58
- 238000003860 storage Methods 0.000 title claims abstract description 32
- 238000012549 training Methods 0.000 claims description 116
- 238000012360 testing method Methods 0.000 claims description 91
- 230000006870 function Effects 0.000 claims description 14
- 238000005259 measurement Methods 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 8
- 230000008707 rearrangement Effects 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 4
- 238000004422 calculation algorithm Methods 0.000 claims description 4
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 230000008859 change Effects 0.000 claims description 2
- 230000008569 process Effects 0.000 description 10
- 238000010586 diagram Methods 0.000 description 8
- 241000208340 Araliaceae Species 0.000 description 3
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 3
- 235000003140 Panax quinquefolius Nutrition 0.000 description 3
- 238000004891 communication Methods 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 235000008434 ginseng Nutrition 0.000 description 3
- 238000007637 random forest analysis Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000004590 computer program Methods 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000013136 deep learning model Methods 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 238000011144 upstream manufacturing Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000000802 evaporation-induced self-assembly Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 229910052754 neon Inorganic materials 0.000 description 1
- GKAOGPIIYCISHV-UHFFFAOYSA-N neon atom Chemical compound [Ne] GKAOGPIIYCISHV-UHFFFAOYSA-N 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
This application discloses a kind of edge device dispatching method, system, device and computer storage mediums, this method comprises: receiving the characteristic information of goal task;Receive the characteristic information of each edge device in edge device cluster;Based on scheduling model parameter trained in advance, the matching score of the characteristic information of goal task and the characteristic information of each edge device is calculated;Edge device corresponding to the matching score of the maximum preset quantity of output numerical value is candidate edge equipment, to select target edge device performance objective task in candidate edge equipment according to selection rule.A kind of edge device dispatching method provided by the present application, matching score is calculated using scheduling model, not only improves computational efficiency, but also can comprehensively consider the characteristic information of each edge device, applicability is good.A kind of edge device scheduling system, device and computer readable storage medium provided by the present application also solve the problems, such as relevant art.
Description
Technical field
This application involves scheduling of resource technical field, more specifically to a kind of edge device dispatching method, system,
Device and computer storage medium.
Background technique
Edge device is that the equipment of entrance is provided to enterprise or service provider's core network, including various Metropolitan Area Network (MAN)s and wide
Domain net access device, such as family and some mobile devices of individual etc..The computing capability of single edge device is limited, but side
The cluster of edge equipment possesses powerful computing capability, and rational management facilitates appointing for huge calculation amount using the cluster of edge device
Business is completed within the limited time limit.
Existing edge device scheduling strategy mainly has First-Fit (adapting to for the first time) and Best-Fit (optimal adaptation).
In First-Fit strategy, scheduler can traverse all edge devices, and assign the task to first and can receive task
Edge device.In Best-Fit strategy, scheduler selects specific characteristic in all edge devices that can receive task
The edge device that combination is best suitable for requirement is allocated task.
However, First-Fit strategy does not take into account that the specific state feature of each edge device, task completion may cause
Need to wait for long period, low efficiency.In Best-Fit strategy, scheduler only considers the feature combination that edge device is selected, and adjusts
The effect of degree device depends on the feature combination of selection to a certain extent, it is difficult to consider edge device cluster itself and local environment
All feature combinations, limitation are big.
In conclusion how to provide a kind of edge device dispatching method that applicability is high be current those skilled in the art urgently
Problem to be solved.
Summary of the invention
The purpose of the application is to provide a kind of edge device dispatching method, can solve how to provide one to a certain extent
The technical issues of kind of applicability high edge device dispatching method.Present invention also provides a kind of edge device scheduling systems, dress
It sets and computer readable storage medium.
To achieve the goals above, a kind of edge device dispatching method provided by the present application, comprising:
Receive the characteristic information of goal task;
Receive the characteristic information of each edge device in edge device cluster;
Based on scheduling model parameter trained in advance, the characteristic information and each edge for calculating the goal task are set
The matching score of standby characteristic information;
Edge device corresponding to the matching score of the maximum preset quantity of output numerical value be candidate edge equipment, with according to
Selection rule selects target edge device in the candidate edge equipment and executes the goal task.
Preferably, described based on scheduling model parameter trained in advance, calculate the characteristic information of the goal task and every
Before the matching score of the characteristic information of a edge device, further includes:
Obtain the characteristic information of training mission and the characteristic information of test assignment;
It obtains the characteristic information of training edge device and tests the characteristic information of edge device;
The theory obtained between the training mission and the trained edge device matches score and the test assignment and institute
State the theoretical matching score between test edge device;
Using the characteristic information of the training mission, the trained edge device characteristic information as input, by the instruction
Practice task and matches score with the theory between the trained edge device as exporting, according to the training pattern and training frame of setting
Frame, the training scheduling model parameter;
Using the characteristic information of the test assignment, the characteristic information for testing edge device as input, according to training
Good scheduling model parameter calculates the actual match score between the test assignment and the test edge device;
Score and the test assignment and institute are matched based on the theory between the test assignment and the test edge device
The actual match score between test edge device is stated, according to the good tune of the loss function and model measurement standard adjusting training of setting
Model parameter is spent, until obtaining final scheduling model parameter.
Preferably, the training pattern includes the wide&deep model in deep width learning model.
Preferably, the training pattern includes first kind learning model or the second class learning model;
The first kind learning model includes Logic Regression Models or FM model or SVM model or naive Bayesian model
Or Random Forest model or GBDT model;
The second class learning model includes DeepFM model or XdeepFM model or deep&cross model.
Preferably, the trained frame includes TensorFlow training frame.
Preferably, the neural network parameter good according to the loss function and model measurement standard adjusting training of setting,
Include:
According to the loss function and model measurement standard regularized learning algorithm rate parameter of setting, batch size parameter, optimizer ginseng
Number, the number of iterations parameter, activation primitive parameter.
Preferably, edge device corresponding to the matching score of the maximum preset quantity of the output numerical value is candidate edge
After equipment, further includes:
Historical data is received, the historical data includes the characteristic information of history goal task, history target edge device
Characteristic information, the matching score between the history goal task and the history target edge device;
The scheduling model parameter is adjusted based on the historical data.
Preferably, it is described obtain training mission characteristic information and test assignment characteristic information before, further includes:
Initial data is acquired, the initial data includes the feature letter of the characteristic information of ancestral task, original edge equipment
The matching score of breath, the ancestral task and the original edge equipment room;
The initial data is resequenced;
Discretization or normalized are carried out to the initial data after rearrangement;
By treated, initial data cutting is training data and test data;
Wherein, the training data includes the characteristic information of training mission, the characteristic information of training edge device, the instruction
Practice task and matches score with the theory between the trained edge device;The test data include test assignment characteristic information,
It tests the characteristic information of edge device, the test assignment and matches score with the theory between the test edge device.
Preferably, after the acquisition initial data, it is described the initial data is resequenced before, also wrap
It includes:
According to the proportionate relationship between preset Different matching score, the initial data is adjusted.
It is preferably, described that by treated, initial data cutting is training data and test data, comprising:
Judgement treated initial data whether missing data will if so, initial data is filled to treated
Filled initial data cutting is the training data and the test data;
Wherein, to treated initial data is filled including average value filling, minimum value filling, maximum value filling,
Similarity number filling.
To achieve the above object, the application further provides for a kind of edge device scheduling system, comprising:
First receiving module, for receiving the characteristic information of goal task;
Second receiving module, for receiving the characteristic information of each edge device in edge device cluster;
First computing module, for calculating the feature letter of the goal task based on scheduling model parameter trained in advance
The matching score of breath and the characteristic information of each edge device;
First output module, edge device corresponding to the matching score for the maximum preset quantity of output numerical value are to wait
Edge device is selected, executes the target times to select target edge device in the candidate edge equipment according to selection rule
Business.
To achieve the above object, the application further provides for a kind of edge device dispatching device, and described device includes storage
Device and processor are stored with the edge device scheduler program that can be run on the processor, the edge on the memory
As above any method is realized when equipment scheduling program is executed by the processor.
Preferably, described device is the node for forming CDN network or block chain network.
To achieve the above object, the application further provides for a kind of computer readable storage medium, described computer-readable
Edge device scheduler program is stored on storage medium, the edge device scheduler program can be held by one or more processor
Row, to realize as above any edge device dispatching method.
To achieve the above object, the application further provides for a kind of edge device dispatching method, is applied to edge device, packet
It includes:
The characteristic information of itself is sent, the characteristic information is used to calculate the matching point between the characteristic information of goal task
Number, the matching score is for assessing whether the edge device executes the goal task;
Judge whether to receive the goal task, if so, executing the goal task.
A kind of edge device dispatching method provided by the present application, receives the characteristic information of goal task;Receive edge device
The characteristic information of each edge device in cluster;Based on scheduling model parameter trained in advance, the feature letter of goal task is calculated
The matching score of breath and the characteristic information of each edge device;Corresponding to the matching score of the maximum preset quantity of output numerical value
Edge device is candidate edge equipment, executes mesh to select target edge device in candidate edge equipment according to selection rule
Mark task.A kind of edge device dispatching method provided by the present application, based on scheduling model parameter trained in advance, to calculate target
Matching score between the characteristic information of task and the characteristic information of edge device, namely matching point is calculated using scheduling model
Number, not only improves computational efficiency, but also can comprehensively consider the characteristic information of each edge device, applicability is good.The application provides
A kind of edge device scheduling system, device and computer readable storage medium also solve the problems, such as relevant art.
Detailed description of the invention
In order to illustrate the technical solutions in the embodiments of the present application or in the prior art more clearly, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of application for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
Fig. 1 is the flow diagram of the application first embodiment;
Fig. 2 is the training schematic diagram of scheduling model parameter;
Fig. 3 is the flow chart of edge device dispatching method provided by the embodiments of the present application;
Fig. 4 is the structural schematic diagram that the edge device that one embodiment of the application discloses dispatches system;
Fig. 5 is the schematic diagram of internal structure for the edge device dispatching device that one embodiment of the application discloses.
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 carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of embodiments of the present application, instead of all the embodiments.It is based on
Embodiment in the application, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall in the protection scope of this application.
The description and claims of this application and term " first ", " second ", " third ", " in above-mentioned attached drawing
The (if present)s such as four " are to be used to distinguish similar objects, without being used to describe a particular order or precedence order.It should manage
The data that solution uses in this way are interchangeable under appropriate circumstances, so that the embodiments described herein can be in addition to illustrating herein
Or the sequence other than the content of description is implemented.In addition, term " includes " and " having " and their any deformation, it is intended that
Cover it is non-exclusive include, for example, containing the process, method, system, product or equipment of a series of steps or units need not limit
In step or unit those of is clearly listed, but may include be not clearly listed or for these process, methods, produce
The other step or units of product or equipment inherently.
It should be noted that the description for being related to " first ", " second " etc. in the present invention is used for description purposes only, and cannot
It is interpreted as its relative importance of indication or suggestion or implicitly indicates the quantity of indicated technical characteristic.Define as a result, " the
One ", the feature of " second " can explicitly or implicitly include at least one of the features.In addition, the skill between each embodiment
Art scheme can be combined with each other, but must be based on can be realized by those of ordinary skill in the art, when technical solution
Will be understood that the combination of this technical solution is not present in conjunction with there is conflicting or cannot achieve when, also not the present invention claims
Protection scope within.
The application provides a kind of edge device dispatching method applied to scheduling model.
Referring to Fig. 1, Fig. 1 is the flow diagram of the application first embodiment.
In the first embodiment, this method may include:
S101 receives the characteristic information of goal task.
In practical application, the characteristic information of goal task can be first received, the characteristic information of goal task may include appointing
The ID of business, IP address, nat type, affiliated province, network operator, stream information, version number and connection attribute etc..It should refer to
Out, edge device packet involved in the application but be not limited to the PC of Windows, Linux, Mac system, Android, Saipan,
The electronic equipment of the Intelligent mobile equipments such as mobile phone, the tablet computer of IOS system and various household NAS types.
S102 receives the characteristic information of each edge device in edge device cluster.
In practical application, after the characteristic information for receiving goal task, each edge in edge device cluster can receive
The characteristic information of equipment.The characteristic information of edge device may include the ID of edge device, IP address, version number, nat type,
Port numbers, current state, geographical attribute, operator's attribute, current upstream and downstream bandwidth, history maximum upstream and downstream bandwidth are current to take
Number of being engaged in etc. indicates the information etc. of edge device current state.
S103 calculates the characteristic information and each edge device of goal task based on scheduling model parameter trained in advance
Characteristic information matching score.
In practical application, the feature letter of received goal task based on scheduling model parameter trained in advance, can be calculated
Matching score between breath and the characteristic information of each edge device, matching score are used to indicate of edge device and goal task
With degree, matching score is higher, and edge device is more suitably executed goal task.By the description for the scheduling model parameter trained in advance
It can be appreciated that this application involves scheduling model is arrived, with independence, high accuracy for examination.
S104: edge device corresponding to the matching score of the maximum preset quantity of output numerical value is candidate edge equipment,
Target edge device performance objective task is selected in candidate edge equipment according to selection rule.
In practical application, after the matching score being calculated between goal task and each edge device, it can incite somebody to action
Edge device corresponding to the matching score of the maximum preset quantity of numerical value is exported as candidate edge equipment, the tool of preset quantity
Body numerical value can be determined according to concrete application scene.Specifically, can be selected in candidate edge equipment according to selection rule
Target edge device performance objective task, for example selected in candidate edge equipment according to the high selection rule of load balancing is chosen
Target edge device out selects target edge device in candidate edge equipment according to the close selection rule of communication distance is chosen
Deng.It should be pointed out that a kind of edge device dispatching method provided by the present application can be applied to the server connecting with edge device
In, it can also be to set with edge that the server connecting with edge device, which can be the local server connecting with edge device,
The cloud server etc. of standby connection.
A kind of edge device dispatching method provided by the present application, receives the characteristic information of goal task;Receive edge device
The characteristic information of each edge device in cluster;Based on scheduling model parameter trained in advance, the feature letter of goal task is calculated
The matching score of breath and the characteristic information of each edge device;Corresponding to the matching score of the maximum preset quantity of output numerical value
Edge device is candidate edge equipment, executes mesh to select target edge device in candidate edge equipment according to selection rule
Mark task.A kind of edge device dispatching method provided by the present application, based on scheduling model parameter trained in advance, to calculate target
Matching score between the characteristic information of task and the characteristic information of edge device, namely matching point is calculated using scheduling model
Number, not only improves computational efficiency, but also can comprehensively consider the characteristic information of each edge device, applicability is good.
Referring to Fig. 2, Fig. 2 is the training schematic diagram of scheduling model parameter.
In the first embodiment, in order to be quickly trained to scheduling model parameter, and guarantee extraneous to scheduling model ginseng
Several debugging calculates the characteristic information and each edge device of goal task based on scheduling model parameter trained in advance
Before the matching score of characteristic information, can with comprising steps of
S201 obtains the characteristic information of training mission and the characteristic information of test assignment.
Training mission refers to that the task for training scheduling model parameter, test assignment refer to that test dispatching model is joined
Several tasks.The characteristic information of training mission and test assignment can determine according to actual needs, and type can appoint with target
The type of business is identical.
S202 obtains the characteristic information of training edge device and tests the characteristic information of edge device.
Training edge device refers to that the edge device for training scheduling model, test edge device are referred to for surveying
Try the edge device of scheduling model.Training edge device and the characteristic information for testing edge device can according to actual needs really
Fixed, type can be identical as the training type of edge collection.
S203 obtains the theory between training mission and training edge device and matches score and test assignment and test edge
The theoretical matching score of equipment room.
Theory between training mission and training edge device matches score and refers to for training the matching of scheduler module to divide
Number, the theory between test assignment and test edge device match score and refer to matching score for test dispatching model.
S204, using the characteristic information of training mission, training edge device characteristic information as input, by training mission and
Theoretical matching score between training edge device is as output, according to the training pattern and training frame of setting, training scheduling mould
Shape parameter.
S205, using the characteristic information of test assignment, test edge device characteristic information as input, according to trained
Scheduling model parameter calculates test assignment and tests the actual match score between edge device.
S206 matches score and test assignment and test edge based on the theory between test assignment and test edge device
The actual match score of equipment room is joined according to the good scheduling model of the loss function and model measurement standard adjusting training of setting
Number, until obtaining final scheduling model parameter.
In practical application, loss function can be cross entropy loss, minimum mean error etc.;Model measurement standard can be
Accuracy rate, recall rate, F (F-Measure, comprehensive evaluation index) 1, ROC curve (receiver operating
Characteristic curve, Receiver operating curve) and AUC (Area Under Curve, below ROC curve
Size) area etc..
In the first embodiment, in order to guarantee the training effect and training effectiveness of scheduling model parameter, training pattern can be with
Including the wide&deep model in deep width learning model.
In the first embodiment, in order to meet the extraneous diversity requirement to scheduling model parameter, training pattern be can wrap
Include first kind learning model or the second class learning model;First kind learning model namely conventional machines learning model may include
Logic Regression Models or FM model or SVM model or naive Bayesian model or Random Forest model or GBDT model;Second class
Learning model namely deep learning model may include DeepFM model or XdeepFM model or deep&cross model.It is real
It, can be according to the suitable training pattern of the selections such as the running environment of scheduling model parameter and performance requirement in the application of border.
In the first embodiment, for the ease of the description of scheduling model parameter, training frame may include TensorFlow
Training frame.
In practical application, the training frame of scheduling model parameter can also for Caffe, Keras, CNTK, Torch7,
MXNet, Leaf, Theano, DeepLearning4, Lasagne, Neon etc..Specifically, can be according to scheduling model parameter
The parameters such as data volume, running environment determine training frame applied by its;For the ease of choosing training frame, the application is to each
The characteristic of class training frame is described, and please refers to table 1, table 1 is the performance number of all kinds of trained frames, correspondingly, choosing
When training frame, suitable training frame can be chosen in table 1 directly according to the relevant parameter of scheduling model parameter.
The performance parameter of all kinds of trained frames of table 1
In the first embodiment, it in order to accelerate the training of scheduling model parameter, weighs according to the loss function and model of setting
The process of the good neural network parameter of amount standard adjusting training can be with specifically: measures mark according to the loss function and model of setting
Quasi- regularized learning algorithm rate parameter, batch size parameter, optimizer parameter, the number of iterations parameter, activation primitive parameter.
It in the first embodiment, can be using historical data to scheduling mould in order to guarantee the accuracy rate of scheduling model parameter
Type is updated, then edge device corresponding to the matching score of the maximum preset quantity of output numerical value be candidate edge equipment it
Afterwards, historical data can also be received, historical data includes the spy of the characteristic information of history goal task, history target edge device
Matching score between reference breath, history goal task and history target edge device;Based on historical data adjustment scheduling model ginseng
Number.
In practical application, scheduling model parameter can be updated by historical data in real time, it can also be according between preset time
Every by historical data update scheduling model parameter.
Referring to Fig. 3, Fig. 3 is the flow chart of edge device dispatching method provided by the embodiments of the present application.
In edge device dispatching method provided by the embodiments of the present application, the characteristic information and test assignment of training mission are obtained
Characteristic information before, can also include:
S301, acquires initial data, and initial data includes the feature letter of the characteristic information of ancestral task, original edge equipment
The matching score of breath, ancestral task and original edge equipment room.
S302 resequences initial data.
In practical application, since the initial data of acquisition has timeliness, for example there is sequencing in time, in order to keep away
Exempt from influence of the timeliness to scheduling model, initial data can be resequenced.
S303 carries out discretization or normalized to the initial data after rearrangement.
In practical application, the format of the Various types of data in the initial data of acquisition may be inconsistent, can be thus scheduling mould
The training band of type is next difficult, in order to avoid such situation, can carry out discretization or normalizing to the initial data after rearrangement
Change processing, so that the uniform format of initial data is a kind of format.
S304, by treated, initial data cutting is training data and test data, wherein training data includes training
The characteristic information of task, the characteristic information of training edge device, training mission match score with the theory between training edge device;
Test data includes the characteristic information of test assignment, the characteristic information for testing edge device, test assignment and test edge device
Between theoretical matching score.
It can be training data and test number by treated initial data cutting according to preset ratio in practical application
According to, for example according to the ratio of 3:1 by treated initial data cutting be training data and test data etc..
In a second embodiment, the ratio data of Different matching score is different in the initial data of acquisition, such as original number
70 points are respectively less than according to the matching score of middle data, then final of the scheduling model parameter obtained using initial data training
It is difficult to be higher than 70 points with score, so that the accuracy of scheduling model parameter is poor, in order to improve the accuracy of scheduling model parameter,
It, can also be according between preset Different matching score before initial data is resequenced after acquiring initial data
Proportionate relationship, adjust initial data.Namely make the preset proportionate relationship of matching fractional part of initial data, so that all kinds of
Matching score can participate in the training of scheduling model parameter, to improve the accuracy of scheduling model parameter.
In practical application, during acquiring initial data, it is understood that there may be the case where partial data is lost, it thus can shadow
Ring the training effect of scheduling model parameter, in order to avoid such situation, will treated initial data cutting is training data and
The process of test data can be with specifically: judgement treated initial data whether missing data, if so, former to treated
Beginning data are filled, and are training data and test data by filled initial data cutting;Wherein, original to treated
Data are filled including average value filling, minimum value filling, maximum value filling, similarity number filling.That is, being lacked in initial data
, can be using the average value of corresponding data in the initial data of acquisition as the data of missing when losing data, it can be by the original of acquisition
Data of the minimum value of corresponding data as missing in beginning data, can be by the maximum value of corresponding data in the initial data of acquisition
It, can be using the similarity number of corresponding data in the initial data of acquisition as data of missing etc. as the data of missing.
On the other hand, the application provides a kind of edge device scheduling system.
Referring to Fig. 4, Fig. 4 is the structural schematic diagram that the edge device that one embodiment of the application discloses dispatches system.
A kind of edge device provided by the present application dispatches system, may include:
First receiving module 401, for receiving the characteristic information of goal task;
Second receiving module 402, for receiving the characteristic information of each edge device in edge device cluster;
First computing module 403, for calculating the characteristic information of goal task based on scheduling model parameter trained in advance
With the matching score of the characteristic information of each edge device;
First output module 404, edge device corresponding to the matching score for the maximum preset quantity of output numerical value
For candidate edge equipment, target edge device performance objective task is selected in candidate edge equipment according to selection rule.
A kind of edge device provided by the present application dispatches system, can also include:
First obtains module, for the first computing module based on scheduling model parameter trained in advance, calculates goal task
Characteristic information and each edge device characteristic information matching score before, obtain characteristic information and the test of training mission
The characteristic information of task;
Second obtains module, for obtaining the characteristic information of trained edge device and testing the characteristic information of edge device;
Third obtains module, matches score for obtaining the theory between training mission and training edge device, and test is appointed
The theory being engaged between test edge device matches score;
First training module, for using the characteristic information of training mission, training edge device characteristic information as input,
Theory between training mission and training edge device is matched to score as output, according to the training pattern and training frame of setting
Frame, training scheduling model parameter;
Second computing module, for using the characteristic information of test assignment, test edge device characteristic information as input,
Test assignment is calculated according to trained scheduling model parameter and tests the actual match score between edge device;
The first adjustment module, for matching score, and test times based on the theory between test assignment and test edge device
Actual match score between business and test edge device is good according to the loss function and model measurement standard adjusting training of setting
Scheduling model parameter, until obtaining final scheduling model parameter.
A kind of edge device provided by the present application dispatches system, and training pattern includes the wide& in deep width learning model
Deep model.
A kind of edge device provided by the present application dispatches system, and training pattern includes conventional machines learning model, depth
Practise model;
Conventional machines learning model includes Logic Regression Models, FM model, SVM model, naive Bayesian model, random
Forest model, GBDT model;
Deep learning model includes DeepFM model, XdeepFM model, deep&cross model.
A kind of edge device provided by the present application dispatches system, and training frame includes TensorFlow training frame.
A kind of edge device provided by the present application dispatches system, and the first adjustment module may include:
The first adjustment unit, for the loss function and model measurement standard regularized learning algorithm rate parameter, batch according to setting
Size parameter, optimizer parameter, the number of iterations parameter, activation primitive parameter.
A kind of edge device provided by the present application dispatches system, can also include:
Third receiving module, corresponding to the matching score for the first maximum preset quantity of output module output numerical value
To receive historical data after candidate edge equipment, historical data includes the characteristic information of history goal task, goes through edge device
Matching score between the characteristic information of history target edge device, history goal task and history target edge device;
Second adjustment module, for adjusting scheduling model parameter based on historical data.
A kind of edge device provided by the present application dispatches system, can also include:
First acquisition module obtains module for first and obtains the characteristic information of training mission and the feature letter of test assignment
Before breath, initial data is acquired, initial data includes the characteristic information of ancestral task, the characteristic information of original edge equipment, original
The matching score of beginning task and original edge equipment room;
First sorting module, for the initial data to be resequenced;
First processing module, for carrying out discretization or normalized to the initial data after rearrangement;
First cutting module, for initial data cutting to be training data and test data by treated;
Wherein, the training data includes the characteristic information of training mission, the characteristic information of training edge device, the instruction
Practice task and matches score with the theory between the trained edge device;The test data include test assignment characteristic information,
It tests the characteristic information of edge device, the test assignment and matches score with the theory between the test edge device.
A kind of edge device provided by the present application dispatches system, can also include:
The first adjustment module will described in the first sorting module after acquisition initial data described in the first acquisition module
Before the initial data is resequenced, according to the proportionate relationship between preset Different matching score, adjust described original
Data.
A kind of edge device provided by the present application dispatches system, and the first cutting module may include:
First judging unit, for judge treated initial data whether missing data, if so, former to treated
Beginning data are filled, and are the training data and the test data by filled initial data cutting;
Wherein, to treated initial data is filled including average value filling, minimum value filling, maximum value filling,
Similarity number filling.
Present invention also provides a kind of edge device dispatching methods, are applied to edge device, comprising the following steps:
The characteristic information of itself is sent, characteristic information is used to calculate the matching score between the characteristic information of goal task,
Matching score for assess edge device whether performance objective task;
Judge whether to receive goal task, if so, performance objective task.
Associated description about the edge device dispatching method provided by the embodiments of the present application applied to edge device please join
Above-described embodiment is read, details are not described herein.
On the other hand, the application provides a kind of edge device dispatching device.
It is the schematic diagram of internal structure for the edge device dispatching device that one embodiment of the application discloses referring to Fig. 5, Fig. 5.
In the present embodiment, edge device dispatching device 1 can be PC (Personal Computer, PC), can also
To be that smart phone, tablet computer, palm PC, portable computer, intelligent router, mine machine, network storage equipment terminal are set
It is standby.
The edge device dispatching device 1 can be the node of composition CDN network or block chain network.
The edge device dispatching device 1 may include memory 11, processor 12 and bus 13.
Wherein, memory 11 includes at least a type of readable storage medium storing program for executing, and readable storage medium storing program for executing includes flash memory, hard
Disk, multimedia card, card-type memory (for example, SD or DX memory etc.), magnetic storage, disk, CD etc..Memory 11 exists
It can be the internal storage unit of edge device dispatching device 1 in some embodiments, such as the edge device dispatching device 1 is hard
Disk.Memory 11 is also possible to the External memory equipment of edge device dispatching device 1 in further embodiments, such as edge is set
The plug-in type hard disk being equipped on standby dispatching device 1, intelligent memory card (Smart Media Card, SMC), secure digital
(Secure Digital, SD) card, flash card (Flash Card) etc..Further, memory 11 can also both include edge
The internal storage unit of equipment scheduling device 1 also includes External memory equipment.Memory 11 can be not only used for storage and be installed on
Application software and Various types of data, such as the code of edge device scheduler program 01 of edge device dispatching device 1 etc. can also be used
In temporarily storing the data that has exported or will export.
Processor 12 can be in some embodiments a central processing unit (Central Processing Unit,
CPU), controller, microcontroller, microprocessor or other data processing chips, the program for being stored in run memory 11
Code or processing data, such as execute edge device scheduler program 01 etc..
The bus 13 can be Peripheral Component Interconnect standard (peripheral component interconnect, abbreviation
PCI) bus or expanding the industrial standard structure (extended industry standard architecture, abbreviation EISA)
Bus etc..The bus can be divided into address bus, data/address bus, control bus etc..For convenient for indicating, in Fig. 5 only with one slightly
Line indicates, it is not intended that an only bus or a type of bus.
Further, edge device dispatching device can also include network interface 14, and network interface 14 optionally can wrap
Wireline interface and/or wireless interface (such as WI-FI interface, blue tooth interface) are included, commonly used in setting in the device 1 with other electronics
Communication connection is established between standby.
Optionally, which can also include user interface, and user interface may include display
(Display), input unit such as keyboard (Keyboard), optional user interface can also include standard wireline interface,
Wireless interface.Optionally, in some embodiments, it is aobvious to can be light-emitting diode display, liquid crystal display, touch control type LCD for display
Show that device and OLED (Organic Light-Emitting Diode, Organic Light Emitting Diode) touch device etc..Wherein, display
Can also it is appropriate be known as display screen or display unit, for be shown in the information handled in edge device dispatching device 1 and
For showing visual user interface.
Fig. 5 illustrates only the edge device dispatching device 1 with component 11-14 and edge device scheduler program 01, this
Field technical staff, can be with it is understood that the structure shown in Fig. 5 does not constitute the restriction to edge equipment scheduling device 1
Including perhaps combining certain components or different component layouts than illustrating less perhaps more components.
A kind of computer readable storage medium provided by the present application is stored with edge device on computer readable storage medium
Scheduler program, edge device scheduler program can be executed by one or more processor, to realize that any embodiment as above is retouched
The edge device dispatching method stated.
Computer readable storage medium referred to herein includes random access memory (RAM), memory, read-only memory
(ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technical field
Any other form of storage medium well known to interior.
In the above-described embodiments, can come wholly or partly by software, hardware, firmware or any combination thereof real
It is existing.When implemented in software, it can entirely or partly realize in the form of a computer program product.
The computer program product includes one or more computer instructions.Load and execute on computers the meter
When calculation machine program instruction, entirely or partly generate according to process or function described in the embodiment of the present invention.The computer can
To be general purpose computer, special purpose computer, computer network or other programmable devices.The computer instruction can be deposited
Storage in a computer-readable storage medium, or from a computer readable storage medium to another computer readable storage medium
Transmission, for example, the computer instruction can pass through wired (example from a web-site, computer, server or data center
Such as coaxial cable, optical fiber, Digital Subscriber Line (DSL)) or wireless (such as infrared, wireless, microwave) mode to another website
Website, computer, server or data center are transmitted.The computer readable storage medium can be computer and can deposit
Any usable medium of storage either includes that the data storages such as one or more usable mediums integrated server, data center are set
It is standby.The usable medium can be magnetic medium, (for example, floppy disk, hard disk, tape), optical medium (for example, DVD) or partly lead
Body medium (such as solid state hard disk Solid State Disk (SSD)) etc..
It is apparent to those skilled in the art that for convenience and simplicity of description, the system of foregoing description,
The specific work process of device and unit, can refer to corresponding processes in the foregoing method embodiment, and details are not described herein.
In several embodiments provided herein, it should be understood that disclosed system, device and method can be with
It realizes by another way.For example, the apparatus embodiments described above are merely exemplary, for example, the unit
It divides, only a kind of logical function partition, there may be another division manner in actual implementation, such as multiple units or components
It can be combined or can be integrated into another system, or some features can be ignored or not executed.Another point, it is shown or
The mutual coupling, direct-coupling or communication connection discussed can be through some interfaces, the indirect coupling of device or unit
It closes or communicates to connect, can be electrical property, mechanical or other forms.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.It can select some or all of unit therein according to the actual needs to realize the mesh of this embodiment scheme
's.
It, can also be in addition, each functional unit in each embodiment of the application can integrate in one processing unit
It is that each unit physically exists alone, can also be integrated in one unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also realize in the form of software functional units.
If the integrated unit is realized in the form of SFU software functional unit and sells or use as independent product
When, it can store in a computer readable storage medium.Based on this understanding, the technical solution of the application is substantially
The all or part of the part that contributes to existing technology or the technical solution can be in the form of software products in other words
It embodies, which is stored in a storage medium, including some instructions are used so that a computer
Equipment (can be personal computer, server or the network equipment etc.) executes the complete of each embodiment the method for the application
Portion or part steps.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only
Memory), random access memory (RAM, Random Access Memory), magnetic or disk etc. are various can store journey
The medium of sequence code.
It should be noted that the serial number of the above embodiments of the invention is only for description, do not represent the advantages or disadvantages of the embodiments.And
The terms "include", "comprise" herein or any other variant thereof is intended to cover non-exclusive inclusion, so that packet
Process, device, article or the method for including a series of elements not only include those elements, but also including being not explicitly listed
Other element, or further include for this process, device, article or the intrinsic element of method.Do not limiting more
In the case where, the element that is limited by sentence "including a ...", it is not excluded that including process, device, the article of the element
Or there is also other identical elements in method.
The foregoing description of the disclosed embodiments makes those skilled in the art can be realized or use the application.To this
A variety of modifications of a little embodiments will be apparent for a person skilled in the art, and the general principles defined herein can
Without departing from the spirit or scope of the application, to realize in other embodiments.Therefore, the application will not be limited
It is formed on the embodiments shown herein, and is to fit to consistent with the principles and novel features disclosed in this article widest
Range.
Claims (15)
1. a kind of edge device dispatching method characterized by comprising
Receive the characteristic information of goal task;
Receive the characteristic information of each edge device in edge device cluster;
Based on scheduling model parameter trained in advance, the characteristic information and each edge device of the goal task are calculated
The matching score of characteristic information;
Edge device corresponding to the matching score of the maximum preset quantity of output numerical value is candidate edge equipment, according to selection
Rule selects target edge device in the candidate edge equipment and executes the goal task.
2. the method according to claim 1, wherein the scheduling model parameter trained in advance based on itself,
Before the matching score for calculating the characteristic information of the goal task and the characteristic information of each edge device, further includes:
Obtain the characteristic information of training mission and the characteristic information of test assignment;
It obtains the characteristic information of training edge device and tests the characteristic information of edge device;
The theory obtained between the training mission and the trained edge device matches score and the test assignment and the survey
Try the theoretical matching score between edge device;
Using the characteristic information of the training mission, the trained edge device characteristic information as input, by the training appoint
It is engaged in matching score with the theory between the trained edge device as exporting, according to the training pattern and training frame of setting, instruction
Practice the scheduling model parameter;
Using the characteristic information of the test assignment, the characteristic information for testing edge device as input, according to trained
Scheduling model parameter calculates the actual match score between the test assignment and the test edge device;
Score and the test assignment and the survey are matched based on the theory between the test assignment and the test edge device
The actual match score between edge device is tried, according to the good scheduling mould of the loss function and model measurement standard adjusting training of setting
Shape parameter, until obtaining final scheduling model parameter.
3. according to the method described in claim 2, it is characterized in that, the training pattern includes in deep width learning model
Wide&deep model.
4. according to the method described in claim 2, it is characterized in that, the training pattern includes first kind learning model or second
Class learning model;
The first kind learning model include Logic Regression Models or FM model or SVM model or naive Bayesian model or with
Machine forest model or GBDT model;
The second class learning model includes DeepFM model or XdeepFM model or deep&cross model.
5. according to the method described in claim 2, it is characterized in that, the trained frame includes TensorFlow training frame.
6. according to the method described in claim 2, it is characterized in that, the loss function and model measurement standard according to setting
The good neural network parameter of adjusting training, comprising:
According to the loss function and model measurement standard regularized learning algorithm rate parameter of setting, batch size parameter, optimizer parameter, change
For count parameter, activation primitive parameter.
7. according to the described in any item methods of claim 2 to 6, which is characterized in that the maximum preset quantity of output numerical value
Matching score corresponding to edge device be candidate edge equipment after, further includes:
Historical data is received, the historical data includes the spy of the characteristic information of history goal task, history target edge device
Matching score between reference breath, the history goal task and the history target edge device;
The scheduling model parameter is adjusted based on the historical data.
8. according to the method described in claim 2, it is characterized in that, the characteristic information and test assignment for obtaining training mission
Characteristic information before, further includes:
Initial data is acquired, the initial data includes the characteristic information of ancestral task, the characteristic information of original edge equipment, institute
State the matching score of ancestral task Yu the original edge equipment room;
The initial data is resequenced;
Discretization or normalized are carried out to the initial data after rearrangement;
By treated, initial data cutting is training data and test data;
Wherein, the training data includes the characteristic information of training mission, the characteristic information of training edge device, the training times
It is engaged in matching score with the theory between the trained edge device;The test data includes the characteristic information of test assignment, test
The characteristic information of edge device, the test assignment match score with the theory between the test edge device.
9. according to the method described in claim 8, it is characterized in that, after the acquisition initial data, it is described will be described original
Before data are resequenced, further includes:
According to the proportionate relationship between preset Different matching score, the initial data is adjusted.
10. according to the method described in claim 8, it is characterized in that, described by treated initial data cutting is training number
According to and test data, comprising:
Whether missing data will fill judgement treated initial data if so, initial data is filled to treated
Initial data cutting afterwards is the training data and the test data;
Wherein, to treated, initial data is filled including average value filling, minimum value filling, maximum value filling, similar
Number filling.
11. a kind of edge device dispatches system characterized by comprising
First receiving module, for receiving the characteristic information of goal task;
Second receiving module, for receiving the characteristic information of each edge device in edge device cluster;
First computing module, for based on scheduling model parameter trained in advance, calculate the characteristic information of the goal task with
The matching score of the characteristic information of each edge device;
First output module, edge device corresponding to the matching score for the maximum preset quantity of output numerical value are candidate side
Edge equipment executes the goal task to select target edge device in the candidate edge equipment according to selection rule.
12. a kind of edge device dispatching device, which is characterized in that described device includes memory and processor, the memory
On be stored with the edge device scheduler program that can be run on the processor, the edge device scheduler program is by the processing
Device realizes method as described in any one of claim 1 to 10 when executing.
13. device according to claim 12, which is characterized in that described device is composition CDN network or block link network
The node of network.
14. a kind of computer readable storage medium, which is characterized in that be stored with edge on the computer readable storage medium and set
Standby scheduler program, the edge device scheduler program can be executed by one or more processor, with realize as claim 1 to
Edge device dispatching method described in any one of 10.
15. a kind of edge device dispatching method, which is characterized in that be applied to edge device, comprising:
The characteristic information of itself is sent, the characteristic information is used to calculate the matching score between the characteristic information of goal task,
The matching score is for assessing whether the edge device executes the goal task;
Judge whether to receive the goal task, if so, executing the goal task.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910229314.1A CN109918205B (en) | 2019-03-25 | 2019-03-25 | Edge equipment scheduling method, system, device and computer storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910229314.1A CN109918205B (en) | 2019-03-25 | 2019-03-25 | Edge equipment scheduling method, system, device and computer storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109918205A true CN109918205A (en) | 2019-06-21 |
CN109918205B CN109918205B (en) | 2023-11-17 |
Family
ID=66966748
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910229314.1A Active CN109918205B (en) | 2019-03-25 | 2019-03-25 | Edge equipment scheduling method, system, device and computer storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109918205B (en) |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110716806A (en) * | 2019-09-27 | 2020-01-21 | 深圳市网心科技有限公司 | Edge node computing capacity determination method, electronic device, system and medium |
CN111260076A (en) * | 2020-01-13 | 2020-06-09 | 天津大学 | Block chain-based edge node training method, block chain and storage medium |
CN111625354A (en) * | 2020-05-19 | 2020-09-04 | 南京乐贤智能科技有限公司 | Arrangement method of computing power of edge computing equipment and related equipment thereof |
CN111782931A (en) * | 2019-11-28 | 2020-10-16 | 北京京东尚科信息技术有限公司 | Method and device for establishing communication between users and computer readable storage medium |
CN112311820A (en) * | 2019-07-26 | 2021-02-02 | 腾讯科技(深圳)有限公司 | Edge device scheduling method, connection method, device and edge device |
WO2021087639A1 (en) * | 2019-11-04 | 2021-05-14 | Microsoft Technology Licensing, Llc | Cdn optimization platform |
CN113067714A (en) * | 2020-01-02 | 2021-07-02 | 中国移动通信有限公司研究院 | Content delivery network scheduling processing method, device and equipment |
CN113612650A (en) * | 2021-06-07 | 2021-11-05 | 北京东方通科技股份有限公司 | Monitoring method for edge computing equipment |
WO2022083293A1 (en) * | 2020-10-23 | 2022-04-28 | International Business Machines Corporation | Managing task flow in edge computing environment |
WO2022252456A1 (en) * | 2021-06-01 | 2022-12-08 | 苏州浪潮智能科技有限公司 | Task scheduling method and apparatus, electronic device, and readable storage medium |
CN116611536A (en) * | 2023-07-19 | 2023-08-18 | 支付宝(杭州)信息技术有限公司 | Model training method and device, electronic equipment and storage medium |
CN116739317A (en) * | 2023-08-15 | 2023-09-12 | 山东宇飞传动技术有限公司 | Mining winch automatic management and dispatching platform, method, equipment and medium |
Citations (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2187567A1 (en) * | 2008-11-18 | 2010-05-19 | Alcatel, Lucent | Predictive method and system for optimizing demands of connectivity services |
CN104346214A (en) * | 2013-07-30 | 2015-02-11 | 中国银联股份有限公司 | Device and method for managing asynchronous tasks in distributed environments |
US9411645B1 (en) * | 2015-08-26 | 2016-08-09 | International Business Machines Corporation | Scheduling MapReduce tasks based on estimated workload distribution |
US20170286180A1 (en) * | 2016-03-31 | 2017-10-05 | International Business Machines Corporation | Joint Network and Task Scheduling |
US10027594B1 (en) * | 2016-03-30 | 2018-07-17 | Amazon Technologies, Inc. | Congestion control for label switching traffic |
CN108319502A (en) * | 2018-02-06 | 2018-07-24 | 广东工业大学 | A kind of method and device of the D2D tasks distribution based on mobile edge calculations |
CN108320297A (en) * | 2018-03-09 | 2018-07-24 | 湖北工业大学 | A kind of video object method for real time tracking and system |
US20180288141A1 (en) * | 2014-11-11 | 2018-10-04 | Wangsu Science & Technology Co., Ltd. | Http scheduling system and method of content delivery network |
CN108874525A (en) * | 2018-06-22 | 2018-11-23 | 浙江大学 | A kind of service request distribution method towards edge calculations environment |
CN109002357A (en) * | 2018-06-07 | 2018-12-14 | 阿里巴巴集团控股有限公司 | Resource allocation methods, device and Internet of things system |
CN109189571A (en) * | 2018-07-30 | 2019-01-11 | 南京邮电大学 | Calculating task dispatching method and system, fringe node, storage medium and terminal |
CN109325541A (en) * | 2018-09-30 | 2019-02-12 | 北京字节跳动网络技术有限公司 | Method and apparatus for training pattern |
CN109358962A (en) * | 2018-08-15 | 2019-02-19 | 南京邮电大学 | The autonomous distributor of mobile computing resource |
CN109491790A (en) * | 2018-11-02 | 2019-03-19 | 中山大学 | Industrial Internet of Things edge calculations resource allocation methods and system based on container |
-
2019
- 2019-03-25 CN CN201910229314.1A patent/CN109918205B/en active Active
Patent Citations (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2187567A1 (en) * | 2008-11-18 | 2010-05-19 | Alcatel, Lucent | Predictive method and system for optimizing demands of connectivity services |
CN104346214A (en) * | 2013-07-30 | 2015-02-11 | 中国银联股份有限公司 | Device and method for managing asynchronous tasks in distributed environments |
US20180288141A1 (en) * | 2014-11-11 | 2018-10-04 | Wangsu Science & Technology Co., Ltd. | Http scheduling system and method of content delivery network |
US9411645B1 (en) * | 2015-08-26 | 2016-08-09 | International Business Machines Corporation | Scheduling MapReduce tasks based on estimated workload distribution |
US10027594B1 (en) * | 2016-03-30 | 2018-07-17 | Amazon Technologies, Inc. | Congestion control for label switching traffic |
US20170286180A1 (en) * | 2016-03-31 | 2017-10-05 | International Business Machines Corporation | Joint Network and Task Scheduling |
CN108319502A (en) * | 2018-02-06 | 2018-07-24 | 广东工业大学 | A kind of method and device of the D2D tasks distribution based on mobile edge calculations |
CN108320297A (en) * | 2018-03-09 | 2018-07-24 | 湖北工业大学 | A kind of video object method for real time tracking and system |
CN109002357A (en) * | 2018-06-07 | 2018-12-14 | 阿里巴巴集团控股有限公司 | Resource allocation methods, device and Internet of things system |
CN108874525A (en) * | 2018-06-22 | 2018-11-23 | 浙江大学 | A kind of service request distribution method towards edge calculations environment |
CN109189571A (en) * | 2018-07-30 | 2019-01-11 | 南京邮电大学 | Calculating task dispatching method and system, fringe node, storage medium and terminal |
CN109358962A (en) * | 2018-08-15 | 2019-02-19 | 南京邮电大学 | The autonomous distributor of mobile computing resource |
CN109325541A (en) * | 2018-09-30 | 2019-02-12 | 北京字节跳动网络技术有限公司 | Method and apparatus for training pattern |
CN109491790A (en) * | 2018-11-02 | 2019-03-19 | 中山大学 | Industrial Internet of Things edge calculations resource allocation methods and system based on container |
Cited By (24)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112311820A (en) * | 2019-07-26 | 2021-02-02 | 腾讯科技(深圳)有限公司 | Edge device scheduling method, connection method, device and edge device |
CN110716806A (en) * | 2019-09-27 | 2020-01-21 | 深圳市网心科技有限公司 | Edge node computing capacity determination method, electronic device, system and medium |
CN110716806B (en) * | 2019-09-27 | 2023-05-12 | 深圳市网心科技有限公司 | Edge node computing capability determining method, electronic equipment, system and medium |
CN113615137B (en) * | 2019-11-04 | 2024-04-12 | 微软技术许可有限责任公司 | CDN optimization platform |
US11856246B2 (en) | 2019-11-04 | 2023-12-26 | Microsoft Technology Licensing, Llc | CDN optimization platform |
WO2021087639A1 (en) * | 2019-11-04 | 2021-05-14 | Microsoft Technology Licensing, Llc | Cdn optimization platform |
CN113615137A (en) * | 2019-11-04 | 2021-11-05 | 微软技术许可有限责任公司 | CDN optimization platform |
CN111782931A (en) * | 2019-11-28 | 2020-10-16 | 北京京东尚科信息技术有限公司 | Method and device for establishing communication between users and computer readable storage medium |
CN113067714B (en) * | 2020-01-02 | 2022-12-13 | 中国移动通信有限公司研究院 | Content distribution network scheduling processing method, device and equipment |
CN113067714A (en) * | 2020-01-02 | 2021-07-02 | 中国移动通信有限公司研究院 | Content delivery network scheduling processing method, device and equipment |
CN111260076A (en) * | 2020-01-13 | 2020-06-09 | 天津大学 | Block chain-based edge node training method, block chain and storage medium |
CN111625354A (en) * | 2020-05-19 | 2020-09-04 | 南京乐贤智能科技有限公司 | Arrangement method of computing power of edge computing equipment and related equipment thereof |
CN111625354B (en) * | 2020-05-19 | 2023-09-19 | 南京乐贤智能科技有限公司 | Edge computing equipment calculation force arranging method and related equipment thereof |
GB2615683A (en) * | 2020-10-23 | 2023-08-16 | Ibm | Managing task flow in edge computing environment |
US11645111B2 (en) | 2020-10-23 | 2023-05-09 | International Business Machines Corporation | Managing task flow in edge computing environment |
WO2022083293A1 (en) * | 2020-10-23 | 2022-04-28 | International Business Machines Corporation | Managing task flow in edge computing environment |
WO2022252456A1 (en) * | 2021-06-01 | 2022-12-08 | 苏州浪潮智能科技有限公司 | Task scheduling method and apparatus, electronic device, and readable storage medium |
US11934871B1 (en) | 2021-06-01 | 2024-03-19 | Inspur Suzhou Intelligent Technology Co., Ltd. | Task scheduling method and apparatus, electronic device, and readable storage medium |
CN113612650A (en) * | 2021-06-07 | 2021-11-05 | 北京东方通科技股份有限公司 | Monitoring method for edge computing equipment |
CN113612650B (en) * | 2021-06-07 | 2022-09-30 | 北京东方通科技股份有限公司 | Monitoring method for edge computing equipment |
CN116611536A (en) * | 2023-07-19 | 2023-08-18 | 支付宝(杭州)信息技术有限公司 | Model training method and device, electronic equipment and storage medium |
CN116611536B (en) * | 2023-07-19 | 2023-09-29 | 支付宝(杭州)信息技术有限公司 | Model training method and device, electronic equipment and storage medium |
CN116739317A (en) * | 2023-08-15 | 2023-09-12 | 山东宇飞传动技术有限公司 | Mining winch automatic management and dispatching platform, method, equipment and medium |
CN116739317B (en) * | 2023-08-15 | 2023-10-31 | 山东宇飞传动技术有限公司 | Mining winch automatic management and dispatching platform, method, equipment and medium |
Also Published As
Publication number | Publication date |
---|---|
CN109918205B (en) | 2023-11-17 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109918205A (en) | A kind of edge device dispatching method, system, device and computer storage medium | |
US9798575B2 (en) | Techniques to manage virtual classes for statistical tests | |
CN104657194B (en) | Method and system for calculating the influence acted in network | |
CN109213597A (en) | Resource allocation methods, device, computer equipment and computer readable storage medium | |
US10782949B2 (en) | Risk aware application placement modeling and optimization in high turnover DevOps environments | |
CN110392882A (en) | It is created in batches in the network for including virtualization network function by pipe function | |
CN109388674A (en) | Data processing method, device, equipment and readable storage medium storing program for executing | |
CN110597719B (en) | Image clustering method, device and medium for adaptation test | |
CN107944000A (en) | Flight freight rate update method, device, electronic equipment, storage medium | |
CN108388512A (en) | Data processing method and server | |
CN109522923A (en) | Customer address polymerization, device and computer readable storage medium | |
CN115904539A (en) | Online generation method, device and equipment of segmentation strategy and storage medium | |
CN113850394B (en) | Federal learning method and device, electronic equipment and storage medium | |
CN110909034A (en) | Service data distribution method, device, terminal equipment and storage medium | |
CN109272402A (en) | Modeling method, device, computer equipment and the storage medium of scorecard | |
CN109828790A (en) | A kind of data processing method and system based on Shen prestige isomery many-core processor | |
CN109409746A (en) | A kind of production scheduling method and device | |
US20180068250A1 (en) | Using run time and historical customer profiling and analytics to determine customer test vs. production differences, and to enhance customer test effectiveness | |
CN115545577A (en) | Scheduling data processing method and device | |
CN113672375A (en) | Resource allocation prediction method, device, equipment and storage medium | |
CN110059083A (en) | A kind of data evaluation method, apparatus and electronic equipment | |
JP5206268B2 (en) | Rule creation program, rule creation method and rule creation device | |
CN111158827A (en) | Method and device for graphic configuration tool correlation calculation value information | |
CN110532267A (en) | Determination method, apparatus, storage medium and the electronic device of field | |
CN112579046B (en) | User story analysis method and device, electronic equipment 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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |