CN108449286A - Network bandwidth resources distribution method and device - Google Patents
Network bandwidth resources distribution method and device Download PDFInfo
- Publication number
- CN108449286A CN108449286A CN201810172531.7A CN201810172531A CN108449286A CN 108449286 A CN108449286 A CN 108449286A CN 201810172531 A CN201810172531 A CN 201810172531A CN 108449286 A CN108449286 A CN 108449286A
- Authority
- CN
- China
- Prior art keywords
- bandwidth
- business datum
- onu
- data
- hidden layer
- 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 43
- 238000009826 distribution Methods 0.000 title claims abstract description 22
- 238000013528 artificial neural network Methods 0.000 claims abstract description 88
- 238000013475 authorization Methods 0.000 claims abstract description 54
- 230000003287 optical effect Effects 0.000 claims abstract description 23
- 241001269238 Data Species 0.000 claims abstract description 11
- 230000005540 biological transmission Effects 0.000 claims description 67
- 230000006870 function Effects 0.000 claims description 55
- 238000004891 communication Methods 0.000 claims description 19
- 238000010606 normalization Methods 0.000 claims description 12
- 238000012545 processing Methods 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 9
- 238000004422 calculation algorithm Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 8
- 210000004218 nerve net Anatomy 0.000 claims description 2
- 238000003860 storage Methods 0.000 description 11
- 238000003062 neural network model Methods 0.000 description 8
- 230000008859 change Effects 0.000 description 5
- 230000008569 process Effects 0.000 description 4
- 238000012163 sequencing technique Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000013468 resource allocation Methods 0.000 description 3
- 210000005036 nerve Anatomy 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000000802 evaporation-induced self-assembly Methods 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
- 239000013307 optical fiber Substances 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/70—Admission control; Resource allocation
- H04L47/83—Admission control; Resource allocation based on usage prediction
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/147—Network analysis or design for predicting network behaviour
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/142—Network analysis or design using statistical or mathematical methods
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
An embodiment of the present invention provides a kind of network bandwidth resources distribution method and device, the controller being applied in SDN network framework, including：The first bandwidth request information that acquired previous period optical network unit ONU is sent obtains the predictions request band data of current period ONU as the input of preset neural network prediction model；According to the predictions request band data of the ONU of the current period, the first bandwidth authorization message is sent to the ONU.The embodiment of the present invention is identified according to business datum, in business datum input neural network prediction model corresponding with business datum mark in the first bandwidth request information that acquired previous period optical network unit ONU is sent, predict the ONU predictions request band datas of current period, there can be each business datum size in the prediction ONU of needle, then distributing authentication information can improve the accuracy rate of network bandwidth resources distribution to ONU.
Description
Technical field
The present invention relates to network resource management technical fields, more particularly to a kind of network bandwidth resources distribution method and dress
It sets.
Background technology
SDN (Software Defined Network, software defined network) includes：Ser (Server, server), CON
(controller, controller), ONU (Optical Network Unit, optical network unit) and OLT (optical line
Terminal, optical line terminal).Currently, the network bandwidth resources distribution method of SDN is as follows：ONU receives different user terminals hair
The business datum sent, the then bandwidth needed for business datum, to CON transmission bandwidth solicited messages.CON receives bandwidth and asks
After seeking information, consolidating for the business datum setting of different priorities can be sequentially allocated according to the priority of business datum from high to low
Determine bandwidth, network bandwidth resources distribution is carried out, to ONU transmission bandwidth authorization messages, wherein the bandwidth authorization message includes：Industry
The fixed-bandwidth of business data setting.After ONU reception bandwidth authorization messages, according to bandwidth authorization message to OLT transmission bandwidth data
Afterwards, server is transmitted to by OLT.
First, this network bandwidth resources distribution method, the bandwidth that CON is distributed to the minimum business datum of priority, far
The required bandwidth of the business datum minimum much smaller than priority causes the lower business datum of priority that can not transmit.
Secondly as the bandwidth that CON distributes to the different business datum of priority is fixed-bandwidth, and user terminal is sent
Business datum be real-time change, the bandwidth needed for different business data is also real-time change.For different business data
Priority distribution the fixed-bandwidth bandwidth difference actually required with business datum away from larger.Therefore, the Netowrk tape of the prior art
The accuracy rate of wide resource allocation methods, bandwidth allocation resource is not high.
Invention content
The embodiment of the present invention is designed to provide a kind of network bandwidth resources distribution method and device, to improve Netowrk tape
The accuracy rate of wide resource allocation.Specific technical solution is as follows：
In a first aspect, an embodiment of the present invention provides a kind of network bandwidth resources distribution method, it is applied to software defined network
Controller in network SDN frameworks, including：
The first bandwidth request information that acquired previous period optical network unit ONU is sent, as preset god
Input through Network Prediction Model obtains the predictions request bandwidth number of current period ONU by the neural network prediction model
According to；The first bandwidth request packet includes：Business datum identifies and business datum identifies corresponding bandwidth on demand data；
Wherein, the neural network prediction model is corresponded with business datum mark；The neural network prediction
Model includes：The business datum identifies the pass of the mathematical operation between corresponding bandwidth on demand and ONU predictions request band datas
System；The predictions request band data includes：The business datum mark and the business datum identify corresponding predictions request
Band data；
According to the predictions request band data of the ONU of the current period, the first bandwidth authorizing letter is sent to the ONU
Breath；
Wherein, first bandwidth authorization message includes：The business datum mark and business datum mark correspond to
Predictions request band data and transmission cycle in ONU can occupy the time slot size of transmission channel.
Optionally, the preset neural network prediction model is to train to obtain as follows in advance：
All second bandwidth request informations that the ONU being had received before current period is sent, it is described according to preset
Time sequencing and step-length, are respectively divided and data set and forecast set have occurred；
It is described data set to have occurred and forecast set includes respectively：ONU marks, business datum mark and the business datum mark
Know corresponding bandwidth on demand data；
Wherein, the corresponding three business datums mark of one ONU marks；
It is using preset normalization algorithm, the business datum occurred in data set and forecast set is corresponding
Bandwidth on demand data are normalized；
The business datum occurred in data set after normalized is identified into corresponding bandwidth on demand data as just
The input of beginning neural network prediction model；
Wherein, the initial neural network prediction model includes：Pre-set parameters；
Using the corresponding bandwidth on demand of the business datum in the forecast set as the training of initial neural network prediction model
Target；
According to the error function of output layer, the parameters of the initial neural network prediction model are adjusted；
Using the initial neural network prediction model after adjustment parameters as preset neural network prediction model.
Optionally, the initial neural network prediction model is：
Wherein, f (x) represents each layer of transmission function, xiOutput for the input of input layer, hidden layer 1 is yj, imply
The output of layer 2 is yk, output layer output is Om, the weight of input layer to hidden layer 1 is Wij, threshold value θj, hidden layer 1 arrives implicit
The weight of layer 2 is Wjk, threshold value θk, the weight of hidden layer 2 to output layer is Wkm, threshold value θm。
Optionally, described according to preset error function, adjust the items of the initial neural network prediction model
Parameter, including：
According to the weighed value adjusting function between the error function of output layer and each layer, initial neural network prediction model is adjusted
In each layer between weights；
According to the adjusting thresholds function between the error function of output layer and each layer, initial neural network prediction model is adjusted
In each layer between threshold value；
Error function is：
Wherein, RmIt represents business datum in ONU actual bandwidth solicited messages and identifies corresponding bandwidth on demand size of data；Om
It represents business datum in prediction bandwidth request information and identifies corresponding bandwidth on demand size of data；
The weighed value adjusting function of input layer to hidden layer 1 is：Wij(n+1)=Wij(n)+ηjδjxi；
The weighed value adjusting function of hidden layer 1 to hidden layer 2 is：Wjk(n+1)=Wjk(n)+ηkδkyj；
The weighed value adjusting function of hidden layer 2 to output layer is：Wkm(n+1)=Wkm(n)+ηmδmyk；
The adjusting thresholds function of input layer to hidden layer 1 is：θj(n+1)=θj(n)+λjδj；
The adjusting thresholds function of hidden layer 1 to hidden layer 2 is：θk(n+1)=θk(n)+λkδk；
The adjusting thresholds function of hidden layer 2 to output layer is：θm(n+1)=θm(n)+λmδm；
Wherein, ηjInput layer is to the weights learning rate of hidden layer 1, ηkRepresent the weights learning that hidden layer 1 arrives hidden layer 2
Rate, ηmRepresent the learning rate that hidden layer 2 arrives output layer；N represents the period, takes positive integer；Wij(n) input layer is represented to implicit
Weights before 1 adjustment of layer, Wij(n+1) weights after input layer is adjusted to hidden layer 1, W are representedjk(n) hidden layer 1 is represented to hidden
Weights before being adjusted containing layer 2, Wjk(n+1) weights after hidden layer 1 is adjusted to hidden layer 2, W are representedkm(n+1) hidden layer 2 is represented
Weights before being adjusted to output layer, Wkm(n) weights after hidden layer 2 to output layer adjustment are represented；λjInput layer is to hidden layer 1
Threshold learning rate, λkRepresent the threshold learning rate that hidden layer 1 arrives hidden layer 2, λmRepresent the threshold value that hidden layer 2 arrives output layer
Learning rate, θj(n) threshold value before input layer is adjusted to hidden layer 1, θ are representedj(n+1) after representing input layer to the adjustment of hidden layer 1
Threshold value, θk(n) threshold value before hidden layer 1 is adjusted to hidden layer 2, θ are representedk(n+1) after representing hidden layer 1 to the adjustment of hidden layer 2
Threshold value, θm(n+1) threshold value before hidden layer 2 to output layer adjustment, θ are representedm(n) after representing hidden layer 2 to output layer adjustment
Threshold value.
Optionally, the first bandwidth request information acquired previous period optical network unit ONU sent, as pre-
If neural network prediction model input, including：
Place is normalized in the first bandwidth request information that acquired previous period optical network unit ONU is sent
Reason；
Using the first bandwidth request information after normalization as the input of preset neural network prediction model.
Optionally, according to the predictions request band data of the ONU of the current period, the first bandwidth is sent to the ONU
Authorization message, including：
The predictions request band data of the ONU of the current period is subjected to anti-normalizing using preset renormalization algorithm
Change is handled；
It is asked the predictions request band data of the ONU of the current period after the anti-normalization processing as target prediction
Band data；
It is big to judge that each business datum in the target prediction bandwidth on demand data identifies corresponding bandwidth on demand data
It is small, if to be more than that respective business datum identifies corresponding pre-set bandwidths threshold value；
It is more than that own service Data Identification is corresponding default if business datum identifies corresponding bandwidth on demand size of data
Bandwidth threshold identifies corresponding bandwidth on demand data to the business datum, and setting is identified as heavy duty；
If business datum identifies corresponding bandwidth on demand size of data, it is corresponding pre- to be less than own service Data Identification
If bandwidth threshold, corresponding bandwidth on demand data are identified to the business datum, setting is identified as light load；
Corresponding bandwidth on demand Data Identification is identified according to the priority of pre-set business datum and business datum, it will
Business datum identifies corresponding bandwidth on demand data and is ranked up in the target prediction bandwidth on demand data, obtains and business number
According to the ranking results for identifying corresponding bandwidth on demand data；
Wherein, the ranking results that corresponding bandwidth on demand data are identified with business datum include：Highest priority
Light load, the light load of the heavy duty of highest priority, priority second, priority it is minimum gently load, priority second
The minimum heavy duty of heavy duty, priority；
The ranking results that corresponding bandwidth on demand data are identified according to the business datum send first band to the ONU
Wide authorization message.
Optionally, the predictions request band data of the ONU according to the current period sends first to the ONU
After the step of bandwidth authorization message, including：
The band data of OLT transmission in current period is obtained as target transmission bandwidth data；
The target transmission bandwidth data are first bandwidth authorization messages of the ONU according to reception, are transmitted from ONU to OLT
Band data；
According to the target transmission bandwidth data, the preset neural network prediction model is updated.
Optionally, the predictions request band data of the ONU according to the current period sends first to the ONU
After the step of bandwidth authorization message, further include：
It is passed according to the next cycle ONU received in current period the second bandwidth request informations sent and the target
Transmission of data sends the second bandwidth authorization message to the ONU；
Wherein, second bandwidth authorization message includes：The business datum mark and business datum mark correspond to
Predictions request band data and transmission cycle in ONU can occupy the time slot size of transmission channel.
Second aspect present embodiments provides a kind of network bandwidth resources distributor, is applied to software defined network SDN
Controller in framework, including：
Bandwidth prediction module, the first bandwidth request for sending acquired previous period optical network unit ONU
Information obtains current period ONU as the input of preset neural network prediction model by the neural network prediction model
Predictions request band data；
The first bandwidth request packet includes：Business datum identifies and business datum identifies corresponding bandwidth on demand number
According to；
Wherein, the neural network prediction model is corresponded with business datum mark；The neural network prediction
Model includes：The business datum identifies the pass of the mathematical operation between corresponding bandwidth on demand and ONU predictions request band datas
System；The predictions request band data includes：The business datum mark and the business datum identify corresponding predictions request
Band data；
Bandwidth authorizing module is used for the predictions request band data of the ONU according to the current period, is sent out to the ONU
Send the first bandwidth authorization message；
Wherein, first bandwidth authorization message includes：The business datum mark and business datum mark correspond to
Predictions request band data and transmission cycle in ONU can occupy the time slot size of transmission channel.
At the another aspect that the present invention is implemented, a kind of electronic equipment, including processor, communication interface, storage are additionally provided
Device and communication bus, wherein processor, communication interface, memory complete mutual communication by communication bus；
Memory, for storing computer program；
Processor when for executing the program stored on memory, realizes that any of the above-described described one kind estimating user
The method of age bracket.
At the another aspect that the present invention is implemented, a kind of computer readable storage medium is additionally provided, it is described computer-readable
Instruction is stored in storage medium, when run on a computer so that computer executes a kind of any of the above-described net
Network method for allocating bandwidth resources.
At the another aspect that the present invention is implemented, the embodiment of the present invention additionally provides a kind of computer program production comprising instruction
Product, when run on a computer so that computer executes a kind of any of the above-described network bandwidth resources distribution method.
A kind of network bandwidth resources distribution method provided in an embodiment of the present invention and device are applied in SDN network framework
Controller.The first bandwidth request information that acquired previous period optical network unit ONU is sent, as preset god
Input through Network Prediction Model obtains the predictions request band data of current period ONU；The first bandwidth request packet
It includes：Business datum identifies and business datum identifies corresponding bandwidth on demand data；Wherein, the neural network prediction model and institute
Business datum mark is stated to correspond；The neural network prediction model includes：The business datum identifies corresponding request band
Mathematical operation relationship between wide and ONU predictions request band datas；The predictions request band data includes：The business number
Corresponding predictions request band data is identified according to mark and the business datum；It is asked according to the prediction of the ONU of the current period
Band data is sought, the first bandwidth authorization message is sent to the ONU；Wherein, first bandwidth authorization message includes：The industry
Business Data Identification and the business datum, which identify ONU in corresponding predictions request band data and transmission cycle, can occupy transmission
The time slot size of channel.This programme is identified according to business datum, and acquired previous period optical network unit ONU is sent
Business datum in first bandwidth request information inputs in neural network prediction model corresponding with business datum mark, prediction
The ONU predictions request band datas of current period can have each business datum size in the prediction ONU of needle, then issue
Authorization message can improve the accuracy rate of network bandwidth resources distribution to ONU.Certainly, it implements any of the products of the present invention or square
Method does not necessarily require achieving all the advantages described above at the same time.
Description of the drawings
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
Obtain other attached drawings according to these attached drawings.
Fig. 1 is a kind of flow chart of network bandwidth resources distribution method of the embodiment of the present invention；
Fig. 2 is the flow chart that the embodiment of the present invention trains neural network model；
Fig. 3 is the flow chart that the embodiment of the present invention sends the first bandwidth authorization message；
Fig. 4 is a kind of structure chart of network bandwidth resources distributor of the embodiment of the present invention；
Fig. 5 is the structural schematic diagram of a kind of electronic equipment of the embodiment of the present invention.
Specific implementation mode
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation describes, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
The embodiment of the present invention is different according to the priority of business datum in order to solve the prior art, distributes this of fixed-bandwidth
Kind of network bandwidth resources distribution method, the not high problem of the accuracy rate of bandwidth allocation resource.It is understood that Internet resources band
Traffic data type is different in width, and the corresponding bandwidth on demand of different business data is different.Therefore, the present embodiment believes bandwidth request
Breath predicts the bandwidth on demand information of current period in the different corresponding neural network models of business datum, to ONU under hair band
Wide authorization message.
As shown in Figure 1, a kind of network bandwidth resources distribution method that the embodiment of the present invention is provided, is applied to software definition
Controller in network SDN frameworks, includes the following steps：
S101, the first bandwidth request information that acquired previous period optical network unit ONU is sent, as default
Neural network prediction model input, pass through the predictions request bandwidth number that neural network prediction model obtains current period ONU
According to；
First bandwidth request packet includes：Business datum identifies and business datum identifies corresponding bandwidth on demand data；
Wherein, neural network prediction model is corresponded with business datum mark；Neural network prediction model includes：Business
Mathematical operation relationship between the corresponding bandwidth on demand of Data Identification and ONU predictions request band datas；Predictions request bandwidth number
According to including：Business datum identifies and the business datum identifies corresponding predictions request band data；
S102 sends the first bandwidth authorization message according to the predictions request band data of the ONU of current period to ONU；
Wherein, the first bandwidth authorization message includes：Business datum identifies and the business datum identifies corresponding prediction and asks
Ask ONU in band data and transmission cycle that can occupy the time slot size of transmission channel.
The present embodiment is before ONU sends the bandwidth on demand data in current period, the predictions request of prediction current period ONU
Band data issues bandwidth authorization message in advance, saves and sends the round-trip of bandwidth on demand data and reception bandwidth authorization message
Time, this improves the real-times of bandwidth allocation.It is higher by method for allocating bandwidth resources real-time in this present embodiment,
And it uses business datum to identify one-to-one neural network model as bandwidth prediction model, ensure that each business
Data can be all transmitted, and targetedly predict the band of each business datum in current period bandwidth on demand information
Width issues bandwidth authorization message.Therefore, the present embodiment can improve the accuracy rate of bandwidth allocation resource.
Optionally, as shown in Fig. 2, the present embodiment before current period according to getting according to bandwidth request information, in advance
The preset neural network prediction model of training is saved so as to the predictions request band data of the ONU for subsequent prediction current period
Time.The preset neural network prediction model is to train to obtain as follows in advance：
S201, all second bandwidth request informations that the ONU being had received before current period is sent are described according to pre-
If time sequencing and step-length, be respectively divided and data set and forecast set occurred；
Wherein, data set has occurred and forecast set includes respectively：ONU marks, business datum mark and the business datum
Identify corresponding bandwidth on demand data；Wherein, the corresponding three business datums mark of an ONU mark；
Wherein, preset time sequencing be according to system time from front to back, step-length is the length being rule of thumb manually set
Degree, actual conditions step-length can set 100.
Assuming that ONU1, has 100 data, then can be divided into 100 groups.First group is 1-100 data and the 101st number
According to the 1-100 data is that data set has occurred, and the 101st data are predictive data set.
S202 is asked the business datum occurred in data set and forecast set is corresponding using preset normalization algorithm
Band data is asked to be normalized；
The business datum occurred in data set after normalized is identified corresponding bandwidth on demand data and made by S203
For the input of initial neural network prediction model；
Wherein, initial neural network prediction model includes：Pre-set parameters；Initial neural network prediction model
For：
Wherein, f (x) represents each layer of transmission function, xiOutput for the input of input layer, hidden layer 1 is yj, imply
The output of layer 2 is yk, output layer output is Om, the weight of input layer to hidden layer 1 is Wij, threshold value θj, hidden layer 1 arrives implicit
The weight of layer 2 is Wjk, threshold value θk, the weight of hidden layer 2 to output layer is Wkm, threshold value θm；
The initial neural network model of the present embodiment shares three layers, and any neural network model can also be used as initial nerve
Network model trains preset neural network prediction model according to data set has occurred.Here, the present embodiment is not limited.
S204, using the corresponding bandwidth on demand of the business datum in forecast set as the training of initial neural network prediction model
Target；
S205 adjusts the parameters of initial neural network prediction model according to the error function of output layer；
S206, using the initial neural network prediction model after adjustment parameters as preset neural network prediction mould
Type.
It is understood that the effect of normalized is the nerve so that the value of bandwidth on demand data is between 0 to 1
Network Prediction Model output valve is between 0 to 1.Bandwidth on demand data are normalized in the present embodiment so that follow-up number
It is simpler according to handling, the speed for training neural network model can be accelerated.The present embodiment before current period by will connect
Data set and forecast set have occurred for all second bandwidth request informations division that the ONU that receives is sent, according to data set has occurred
Targetedly business datum is trained to identify corresponding neural network model, business datum can be improved and identify corresponding nerve net
Network model accuracy.
Optionally, S205 can be realized according to following steps：
Step 1：According to the weighed value adjusting function between the error function of output layer and each layer, initial neural network is adjusted
The weights between each layer in prediction model；
Step 2：According to the adjusting thresholds function between the error function of output layer and each layer, initial neural network is adjusted
The threshold value between each layer in prediction model；
Wherein, error function is：
Wherein, RmIt represents business datum in ONU actual bandwidth solicited messages and identifies corresponding bandwidth on demand size of data；Om
It represents business datum in prediction bandwidth request information and identifies corresponding bandwidth on demand size of data；
The weighed value adjusting function of input layer to hidden layer 1 is：Wij(n+1)=Wij(n)+ηjδjxi；
The weighed value adjusting function of hidden layer 1 to hidden layer 2 is：Wjk(n+1)=Wjk(n)+ηkδkyj；
The weighed value adjusting function of hidden layer 2 to output layer is：Wkm(n+1)=Wkm(n)+ηmδmyk；
The adjusting thresholds function of input layer to hidden layer 1 is：θj(n+1)=θj(n)+λjδj；
The adjusting thresholds function of hidden layer 1 to hidden layer 2 is：θk(n+1)=θk(n)+λkδk；
The adjusting thresholds function of hidden layer 2 to output layer is：θm(n+1)=θm(n)+λmδm；
Wherein, ηjInput layer is to the weights learning rate of hidden layer 1, ηkRepresent the weights learning that hidden layer 1 arrives hidden layer 2
Rate, ηmRepresent the learning rate that hidden layer 2 arrives output layer；N represents the period, takes positive integer；Wij(n) input layer is represented to implicit
Weights before 1 adjustment of layer, Wij(n+1) weights after input layer is adjusted to hidden layer 1, W are representedjk(n) hidden layer 1 is represented to hidden
Weights before being adjusted containing layer 2, Wjk(n+1) weights after hidden layer 1 is adjusted to hidden layer 2, W are representedkm(n+1) hidden layer 2 is represented
Weights before being adjusted to output layer, Wkm(n) weights after hidden layer 2 to output layer adjustment are represented；λjInput layer is to hidden layer 1
Threshold learning rate, λkRepresent the threshold learning rate that hidden layer 1 arrives hidden layer 2, λmRepresent the threshold value that hidden layer 2 arrives output layer
Learning rate, θj(n) threshold value before input layer is adjusted to hidden layer 1, θ are representedj(n+1) after representing input layer to the adjustment of hidden layer 1
Threshold value, θk(n) threshold value before hidden layer 1 is adjusted to hidden layer 2, θ are representedk(n+1) after representing hidden layer 1 to the adjustment of hidden layer 2
Threshold value, θm(n+1) threshold value before hidden layer 2 to output layer adjustment, θ are representedm(n) after representing hidden layer 2 to output layer adjustment
Threshold value.
Optionally, the first bandwidth request sent acquired previous period optical network unit ONU in step S101
Information includes the following steps as the input of preset neural network prediction model：
Step 1：The first bandwidth request information that acquired previous period optical network unit ONU is sent is returned
One change is handled；
Step 2：Using the first bandwidth request information after normalization as the input of preset neural network prediction model.
It is understood that the present embodiment determines the method for normalized and the method phase of prior art normalized
Together, this will not be detailed here.
Optionally, as shown in figure 3, S102 can be obtained according to following steps：
S301 is returned using preset renormalization algorithm by the predictions request band data of the ONU of current period is counter
One change is handled；
S302 asks the predictions request band data of the ONU of the current period after anti-normalization processing as target prediction
Seek band data；
It is big to judge that each business datum in target prediction bandwidth on demand data identifies corresponding bandwidth on demand data by S303
It is small, if to be more than that respective business datum identifies corresponding pre-set bandwidths threshold value；
Wherein, business datum identifies corresponding pre-set bandwidths threshold value according to artificially rule of thumb setting, by calculating centainly
Each business datum size average value of the actual transmissions of period network and set.
S304 is more than that own service Data Identification corresponds to if business datum identifies corresponding bandwidth on demand size of data
Pre-set bandwidths threshold value, corresponding bandwidth on demand data are identified to business datum, setting is identified as heavy duty；
S305 is less than own service Data Identification pair if business datum identifies corresponding bandwidth on demand size of data
The pre-set bandwidths threshold value answered identifies corresponding bandwidth on demand data to business datum, and setting is identified as light load；
S306, according to the corresponding bandwidth on demand data of the mark of the priority of pre-set business datum and business datum
Business datum in target prediction bandwidth on demand data is identified corresponding bandwidth on demand data and is ranked up by mark, is obtained and industry
The ranking results for the corresponding bandwidth on demand data of Data Identification of being engaged in；
Wherein, the ranking results that corresponding bandwidth on demand data are identified with business datum include：The light of highest priority is born
The minimum light load of load, the light load of the heavy duty of highest priority, priority second, priority, priority second heavy burden
It carries, the heavy duty that priority is minimum；
S307 identifies the ranking results of corresponding bandwidth on demand data according to business datum, and sending the first bandwidth to ONU awards
Weigh information.
For example, destination request band data, which includes three kinds of traffic data types, is respectively：Accelerate forwarding
(expedited forwarding, EF), ensure forwarding (Assured Forwarding, AF) and (Best that does one's best
Effort, BE).The business datum highest priority of EF types；The business datum priority second of AF types；The business of BE types
Data priority is minimum.
Assuming that the corresponding bandwidth on demand size of data of the business datum of EF types is in target prediction bandwidth on demand data
The corresponding pre-set bandwidths threshold value of business datum of 100M, EF type is 80M, then the business datum of EF types is identified corresponding ask
Band data setting is asked to be identified as heavy duty；The corresponding request band of the business datum of AF types in target prediction bandwidth on demand data
Wide size of data is 50M, and the corresponding pre-set bandwidths threshold value of business datum of AF types is 60M, then by target prediction bandwidth on demand
The corresponding bandwidth on demand data setting of the business datum of AF types is identified as light load in data；Target prediction bandwidth on demand data
The corresponding bandwidth on demand size of data of business datum of middle BE types is 70M, the corresponding pre-set bandwidths of business datum of BE types
Threshold value is 60M, then by the corresponding bandwidth on demand data setting mark of the business datum of BE types in target prediction bandwidth on demand data
It is heavy duty to know；Business datum identifies the ranking results of corresponding bandwidth on demand data so in target prediction bandwidth on demand data
For：EF heavy dutys, AF gently load, BE heavy dutys, and the first bandwidth authorization message is sent to ONU according to ranking results.
It is exported in view of neural network model, the normalized knot of predictions request band data of the ONU of current period
Fruit.Therefore, the predictions request band data of the ONU for the current period that the present embodiment exports neural network carries out renormalization,
After renormalization, business datum in predictions request band data is identified into corresponding bandwidth on demand data setting mark, according to excellent
First grade and business datum identify corresponding bandwidth on demand Data Identification sequence, then send the first authorization message.The present embodiment is protected
Each business datum normal transmission in ONU has been demonstrate,proved, has avoided business datum that from can not being transmitted because priority is relatively low.Meanwhile this reality
It applies that example considers the priority of business datum and business datum identifies corresponding two factors of bandwidth on demand size of data, sends the
One bandwidth authorization message carries out bandwidth resource allocation, improves the accuracy of bandwidth allocation resource.
Optionally, after the S102 the step of, a kind of network bandwidth resources distribution method in the present embodiment further includes：
Step 1：The band data of OLT transmission in current period is obtained as target transmission bandwidth data；
Wherein, target transmission bandwidth data are first bandwidth authorization messages of the ONU according to reception, are transmitted from ONU to OLT
Band data；
Step 2：According to target transmission bandwidth data, preset neural network prediction model is updated.
For the band data that the present embodiment transmits OLT in current period as target transmission bandwidth data, update is preset
Neural network prediction model so that neural network prediction model prediction is more accurate.Neural network prediction model a cycle is more
New primary, using updated neural network prediction model, prediction ONU predictions request band data real-times are stronger.
Optionally, after the S102 the step of, a kind of network bandwidth resources distribution method in the present embodiment further includes：
According to the next cycle ONU received in current period the second bandwidth request informations sent and object transmission number
According to ONU the second bandwidth authorization messages of transmission；
Wherein, the second bandwidth authorization message includes：Business datum identifies and business datum identifies corresponding predictions request band
ONU can occupy the time slot size of transmission channel in wide data and transmission cycle.
It is continuous, controls it is understood that a transmission cycle internal controller receives the bandwidth request information that ONU is sent
During the entire process of device processed receives the first bandwidth request information of current period and sends the wide authorization message of first band, also it will receive
The ONU of next cycle sends the second bandwidth request information.The present embodiment is next according to being received in current current period
The second bandwidth request information and object transmission data, controller that period ONU sends adjust and are sent to the second band of ONU in real time
Wide authorization message, to prepare for next cycle transmission data.
For example, it is assumed that the business datum service identification pair in the bandwidth request information that prediction current period ONU is sent
In the predictions request band data and transmission cycle answered ONU can occupy transmission channel time slot it is smaller, and had sent to ONU
First bandwidth authorization message, ONU carry out data transmission according to the band data that the first bandwidth authorization message is transmitted to OLT.Currently
Period internal controller receives the second bandwidth request information of next cycle, and the business datum in the second bandwidth request information
Corresponding bandwidth is larger, then controller can adjust bandwidth authorization message in real time so that business mark in the second bandwidth authorization message
Know ONU in corresponding predictions request band data and transmission cycle and can occupy the time slot of transmission channel and become larger, reduces ONU end number
According to overstock, improve ONU transmission datas efficiency.
As shown in figure 4, a kind of network bandwidth resources distributor that the embodiment of the present invention is provided, is applied to software definition
Controller in network SDN frameworks, including：
Bandwidth prediction module 401, the first bandwidth for sending acquired previous period optical network unit ONU are asked
Information is sought, as the input of preset neural network prediction model, obtains current period ONU's by neural network prediction model
Predictions request band data；
Wherein, the first bandwidth request packet includes：Business datum identifies and business datum identifies corresponding bandwidth on demand number
According to；
Wherein, neural network prediction model is corresponded with business datum mark；Neural network prediction model includes：Business
Mathematical operation relationship between the corresponding bandwidth on demand of Data Identification and ONU predictions request band datas；Predictions request bandwidth number
According to including：Business datum identifies and business datum identifies corresponding predictions request band data；
Bandwidth authorizing module 402 is used for the predictions request band data of the ONU according to current period, and first is sent to ONU
Bandwidth authorization message；
Wherein, the first bandwidth authorization message includes：Business datum identifies and business datum identifies corresponding predictions request band
ONU can occupy the time slot size of transmission channel in wide data and transmission cycle.
Optionally, bandwidth prediction module 401, including：Model training submodule；
Model training submodule includes：
Set divides subelement, all second bandwidth requests that the ONU for will be had received before current period is sent
Information is respectively divided and data set and forecast set has occurred according to preset time sequencing and step-length；
Wherein, data set has occurred and forecast set includes respectively：ONU marks, business datum mark and business datum mark
Corresponding bandwidth on demand data；
Wherein, the corresponding three business datums mark of an ONU mark；
Subelement is normalized, for using preset normalization algorithm, the business that will occur in data set and forecast set
The corresponding bandwidth on demand data of data are normalized；
Mode input subelement, for the business datum mark occurred in data set after normalized is corresponding
Input of the bandwidth on demand data as initial neural network prediction model；
Wherein, initial neural network prediction model includes：Pre-set parameters；
Simulated target subelement, for using the corresponding bandwidth on demand of the business datum in forecast set as initial neural network
The training objective of prediction model；
Parameter adjustment subelement adjusts each of initial neural network prediction model for the error function according to output layer
Item parameter；
Prediction model subelement is used to adjust the initial neural network prediction model after parameters as preset god
Through Network Prediction Model.
Optionally, prediction model subelement is specifically used for：By the initial neural network prediction model after adjustment parameters
As preset neural network prediction model；
Initially neural network prediction model is：
Wherein, f (x) represents each layer of transmission function, xiOutput for the input of input layer, hidden layer 1 is yj, imply
The output of layer 2 is yk, output layer output is Om, the weight of input layer to hidden layer 1 is Wij, threshold value θj, hidden layer 1 arrives implicit
The weight of layer 2 is Wjk, threshold value θk, the weight of hidden layer 2 to output layer is Wkm, threshold value θm。
Optionally, parameter adjustment subelement is specifically used for：
According to the weighed value adjusting function between the error function of output layer and each layer, initial neural network prediction model is adjusted
In each layer between weights；
According to the adjusting thresholds function between the error function of output layer and each layer, initial neural network prediction model is adjusted
In each layer between threshold value；
Error function is：
Wherein, RmIt represents business datum in ONU actual bandwidth solicited messages and identifies corresponding bandwidth on demand size of data；Om
It represents business datum in prediction bandwidth request information and identifies corresponding bandwidth on demand size of data；
The weighed value adjusting function of input layer to hidden layer 1 is：Wij(n+1)=Wij(n)+ηjδjxi；
The weighed value adjusting function of hidden layer 1 to hidden layer 2 is：Wjk(n+1)=Wjk(n)+ηkδkyj；
The weighed value adjusting function of hidden layer 2 to output layer is：Wkm(n+1)=Wkm(n)+ηmδmyk；
The adjusting thresholds function of input layer to hidden layer 1 is：θj(n+1)=θj(n)+λjδj；
The adjusting thresholds function of hidden layer 1 to hidden layer 2 is：θk(n+1)=θk(n)+λkδk；
The adjusting thresholds function of hidden layer 2 to output layer is：θm(n+1)=θm(n)+λmδm；
Wherein, ηjInput layer is to the weights learning rate of hidden layer 1, ηkRepresent the weights learning that hidden layer 1 arrives hidden layer 2
Rate, ηmRepresent the learning rate that hidden layer 2 arrives output layer；N represents the period, takes positive integer；Wij(n) input layer is represented to implicit
Weights before 1 adjustment of layer, Wij(n+1) weights after input layer is adjusted to hidden layer 1, W are representedjk(n) hidden layer 1 is represented to hidden
Weights before being adjusted containing layer 2, Wjk(n+1) weights after hidden layer 1 is adjusted to hidden layer 2, W are representedkm(n+1) hidden layer 2 is represented
Weights before being adjusted to output layer, Wkm(n) weights after hidden layer 2 to output layer adjustment are represented；λjInput layer is to hidden layer 1
Threshold learning rate, λkRepresent the threshold learning rate that hidden layer 1 arrives hidden layer 2, λmRepresent the threshold value that hidden layer 2 arrives output layer
Learning rate, θj(n) threshold value before input layer is adjusted to hidden layer 1, θ are representedj(n+1) after representing input layer to the adjustment of hidden layer 1
Threshold value, θk(n) threshold value before hidden layer 1 is adjusted to hidden layer 2, θ are representedk(n+1) after representing hidden layer 1 to the adjustment of hidden layer 2
Threshold value, θm(n+1) threshold value before hidden layer 2 to output layer adjustment, θ are representedm(n) after representing hidden layer 2 to output layer adjustment
Threshold value.
Optionally, prediction model subelement is specifically used for：
Place is normalized in the first bandwidth request information that acquired previous period optical network unit ONU is sent
Reason；
Using the first bandwidth request information after normalization as the input of preset neural network prediction model.
Optionally, bandwidth authorizing module 402 is specifically used for：
The predictions request band data of the ONU of current period is carried out at renormalization using preset renormalization algorithm
Reason；
Using the predictions request band data of the ONU of the current period after anti-normalization processing as target prediction bandwidth on demand
Data；
Judge that each business datum in target prediction bandwidth on demand data identifies corresponding bandwidth on demand size of data, is
No is more than that respective business datum identifies corresponding pre-set bandwidths threshold value；
It is more than that own service Data Identification is corresponding default if business datum identifies corresponding bandwidth on demand size of data
Bandwidth threshold identifies corresponding bandwidth on demand data to business datum, and setting is identified as heavy duty；
If business datum identifies corresponding bandwidth on demand size of data, it is corresponding pre- to be less than own service Data Identification
If bandwidth threshold, corresponding bandwidth on demand data are identified to business datum, setting is identified as light load；
Corresponding bandwidth on demand Data Identification is identified according to the priority of pre-set business datum and business datum, it will
Business datum identifies corresponding bandwidth on demand data and is ranked up in target prediction bandwidth on demand data, obtains and business datum mark
Know the ranking results of corresponding bandwidth on demand data；
Wherein, the ranking results that corresponding bandwidth on demand data are identified with business datum include：The light of highest priority is born
The minimum light load of load, the light load of the heavy duty of highest priority, priority second, priority, priority second heavy burden
It carries, the heavy duty that priority is minimum；
The ranking results that corresponding bandwidth on demand data are identified according to business datum send the first bandwidth authorizing letter to ONU
Breath.
Optionally, bandwidth authorizing module 402, including：
Model modification submodule, for obtaining the band data of OLT transmission in current period as target transmission bandwidth number
According to；
Target transmission bandwidth data are first bandwidth authorization messages of the ONU according to reception, the bandwidth transmitted from ONU to OLT
Data；
According to target transmission bandwidth data, preset neural network prediction model is updated.
Optionally, bandwidth authorizing module 402, including：
Bandwidth authorizing submodule, the second bandwidth for being sent according to the next cycle ONU received in current period
Solicited message and object transmission data send the second bandwidth authorization message to ONU；
Wherein, the second bandwidth authorization message includes：Business datum identifies and business datum identifies corresponding predictions request band
ONU can occupy the time slot size of transmission channel in wide data and transmission cycle.
The embodiment of the present invention additionally provides a kind of electronic equipment, as shown in figure 5, including processor 501, communication interface 502,
Memory 503 and communication bus 504, wherein processor 501, communication interface 502, memory 503 are complete by communication bus 504
At mutual communication,
Memory 503, for storing computer program；
Processor 501 when for executing the program stored on memory 503, realizes following steps：
The first bandwidth request information that acquired previous period optical network unit ONU is sent, as preset god
Input through Network Prediction Model obtains the predictions request band data of current period ONU by neural network prediction model；The
One bandwidth request information includes：Business datum identifies and business datum identifies corresponding bandwidth on demand data；
Wherein, neural network prediction model is corresponded with business datum mark；Neural network prediction model includes：Business
Mathematical operation relationship between the corresponding bandwidth on demand of Data Identification and ONU predictions request band datas；Predictions request bandwidth number
According to including：Business datum identifies and business datum identifies corresponding predictions request band data；
According to the predictions request band data of the ONU of current period, the first bandwidth authorization message is sent to ONU；
Wherein, the first bandwidth authorization message includes：Business datum identifies and business datum identifies corresponding predictions request band
ONU can occupy the time slot size of transmission channel in wide data and transmission cycle.
The communication bus that above-mentioned electronic equipment is mentioned can be Peripheral Component Interconnect standard (Peripheral Component
Interconnect, PCI) bus or expanding the industrial standard structure (Extended Industry Standard
Architecture, EISA) bus etc..The communication bus can be divided into address bus, data/address bus, controlling bus etc..For just
It is only indicated with a thick line in expression, figure, it is not intended that an only bus or a type of bus.
Communication interface is for the communication between above-mentioned electronic equipment and other equipment.
Memory may include random access memory (Random Access Memory, RAM), can also include non-easy
The property lost memory (Non-Volatile Memory, NVM), for example, at least a magnetic disk storage.Optionally, memory may be used also
To be at least one storage device for being located remotely from aforementioned processor.
Above-mentioned processor can be general processor, including central processing unit (Central Processing Unit,
CPU), network processing unit (Network Processor, NP) etc.；It can also be digital signal processor (Digital Signal
Processing, DSP), it is application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), existing
It is field programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic device, discrete
Door or transistor logic, discrete hardware components.
In another embodiment provided by the invention, a kind of computer readable storage medium is additionally provided, which can
It reads to be stored with instruction in storage medium, when run on a computer so that computer executes any institute in above-described embodiment
A kind of network bandwidth resources distribution method stated.
In another embodiment provided by the invention, a kind of computer program product including instruction is additionally provided, when it
When running on computers so that computer executes any a kind of network bandwidth resources distribution side in above-described embodiment
Method.
In the above-described embodiments, can come wholly or partly by software, hardware, firmware or its arbitrary combination 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.When loading on computers and executing the computer program instructions, all or
It partly generates according to the flow or function described in the embodiment of the present invention.The computer can be all-purpose computer, special meter
Calculation machine, computer network or other programmable devices.The computer instruction can be stored in computer readable storage medium
In, or from a computer readable storage medium to the transmission of another computer readable storage medium, for example, the computer
Instruction can pass through wired (such as coaxial cable, optical fiber, number from a web-site, computer, server or data center
User's line (DSL)) or wireless (such as infrared, wireless, microwave etc.) mode to another web-site, computer, server or
Data center is transmitted.The computer readable storage medium can be any usable medium that computer can access or
It is comprising data storage devices such as one or more usable mediums integrated server, data centers.The usable medium can be with
It is magnetic medium, (for example, floppy disk, hard disk, tape), optical medium (for example, DVD) or semiconductor medium (such as solid state disk
Solid State Disk (SSD)) etc..
It should be noted that herein, relational terms such as first and second and the like are used merely to a reality
Body or operation are distinguished with another entity or operation, are deposited without necessarily requiring or implying between these entities or operation
In any actual relationship or order or sequence.Moreover, the terms "include", "comprise" or its any other variant are intended to
Non-exclusive inclusion, so that the process, method, article or equipment including a series of elements is not only wanted including those
Element, but also include other elements that are not explicitly listed, or further include for this process, method, article or equipment
Intrinsic element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that
There is also other identical elements in process, method, article or equipment including the element.
Each embodiment in this specification is all made of relevant mode and describes, identical similar portion between each embodiment
Point just to refer each other, and each embodiment focuses on the differences from other embodiments.Especially for device reality
For applying example, since it is substantially similar to the method embodiment, so description is fairly simple, related place is referring to embodiment of the method
Part explanation.
Claims (10)
1. a kind of network bandwidth resources distribution method, which is characterized in that the control being applied in software defined network SDN frameworks
Device, the method includes：
The first bandwidth request information that acquired previous period optical network unit ONU is sent, as preset nerve net
The input of network prediction model obtains the predictions request band data of current period ONU by the neural network prediction model；Institute
The first bandwidth request packet is stated to include：Business datum identifies and business datum identifies corresponding bandwidth on demand data；
Wherein, the neural network prediction model is corresponded with business datum mark；The neural network prediction model
Including：The business datum identifies the mathematical operation relationship between corresponding bandwidth on demand and ONU predictions request band datas；Institute
Stating predictions request band data includes：The business datum mark and the business datum identify corresponding predictions request bandwidth number
According to；
According to the predictions request band data of the ONU of the current period, the first bandwidth authorization message is sent to the ONU；
Wherein, first bandwidth authorization message includes：The business datum mark and business datum mark are corresponding pre-
The time slot size of transmission channel can be occupied by surveying ONU in bandwidth on demand data and transmission cycle.
2. according to the method described in claim 1, it is characterized in that, the preset neural network prediction model is by as follows
Training obtains step in advance：
All second bandwidth request informations that the ONU being had received before current period is sent, it is described according to the preset time
Sequence and step-length, are respectively divided and data set and forecast set have occurred；
It is described data set to have occurred and forecast set includes respectively：ONU marks, business datum mark and business datum mark pair
The bandwidth on demand data answered；
Wherein, the corresponding three business datums mark of one ONU marks；
Using preset normalization algorithm, by the corresponding request of the business datum occurred in data set and forecast set
Band data is normalized；
The business datum occurred in data set after normalized is identified into corresponding bandwidth on demand data as initial god
Input through Network Prediction Model；
Wherein, the initial neural network prediction model includes：Pre-set parameters；
Using the corresponding bandwidth on demand of the business datum in the forecast set as the training objective of initial neural network prediction model；
According to the error function of output layer, the parameters of the initial neural network prediction model are adjusted；
Using the initial neural network prediction model after adjustment parameters as preset neural network prediction model.
3. according to the method described in claim 2, it is characterized in that, the initial neural network prediction model is：
Wherein, f (x) represents each layer of transmission function, xiOutput for the input of input layer, hidden layer 1 is yj, hidden layer 2
Output is yk, output layer output is Om, the weight of input layer to hidden layer 1 is Wij, threshold value θj, hidden layer 1 to hidden layer 2
Weight is Wjk, threshold value θk, the weight of hidden layer 2 to output layer is Wkm, threshold value θm。
4. according to the method described in claim 3, it is characterized in that, described according to preset error function, described in adjustment
The parameters of initial neural network prediction model, including：
According to the weighed value adjusting function between the error function of output layer and each layer, adjust in initial neural network prediction model
Weights between each layer；
According to the adjusting thresholds function between the error function of output layer and each layer, adjust in initial neural network prediction model
Threshold value between each layer；
Error function is：
Wherein, RmIt represents business datum in ONU actual bandwidth solicited messages and identifies corresponding bandwidth on demand size of data；OmIt represents
Predict that business datum identifies corresponding bandwidth on demand size of data in bandwidth request information；
The weighed value adjusting function of input layer to hidden layer 1 is：Wij(n+1)=Wij(n)+ηjδjxi；
The weighed value adjusting function of hidden layer 1 to hidden layer 2 is：Wjk(n+1)=Wjk(n)+ηkδkyj；
The weighed value adjusting function of hidden layer 2 to output layer is：Wkm(n+1)=Wkm(n)+ηmδmyk；
The adjusting thresholds function of input layer to hidden layer 1 is：θj(n+1)=θj(n)+λjδj；
The adjusting thresholds function of hidden layer 1 to hidden layer 2 is：θk(n+1)=θk(n)+λkδk；
The adjusting thresholds function of hidden layer 2 to output layer is：θm(n+1)=θm(n)+λmδm；
Wherein, ηjInput layer is to the weights learning rate of hidden layer 1, ηkThe weights learning rate that hidden layer 1 arrives hidden layer 2 is represented,
ηmRepresent the learning rate that hidden layer 2 arrives output layer；N represents the period, takes positive integer；Wij(n) input layer is represented to adjust to hidden layer 1
Weights before whole, Wij(n+1) weights after input layer is adjusted to hidden layer 1, W are representedjk(n) it represents hidden layer 1 and arrives hidden layer 2
Weights before adjustment, Wjk(n+1) weights after hidden layer 1 is adjusted to hidden layer 2, W are representedkm(n+1) hidden layer 2 is represented to defeated
Go out the weights before layer adjustment, Wkm(n) weights after hidden layer 2 to output layer adjustment are represented；λjThreshold value of the input layer to hidden layer 1
Learning rate, λkRepresent the threshold learning rate that hidden layer 1 arrives hidden layer 2, λmRepresent the threshold learning that hidden layer 2 arrives output layer
Rate, θj(n) threshold value before input layer is adjusted to hidden layer 1, θ are representedj(n+1) threshold after input layer is adjusted to hidden layer 1 is represented
Value, θk(n) threshold value before hidden layer 1 is adjusted to hidden layer 2, θ are representedk(n+1) threshold after hidden layer 1 is adjusted to hidden layer 2 is represented
Value, θm(n+1) threshold value before hidden layer 2 to output layer adjustment, θ are representedm(n) threshold after hidden layer 2 to output layer adjustment is represented
Value.
5. according to the method described in claim 1, it is characterized in that,
First bandwidth request information that acquired previous period optical network unit ONU is sent, as preset god
Input through Network Prediction Model, including：
The first bandwidth request information that acquired previous period optical network unit ONU is sent is normalized；
Using the first bandwidth request information after normalization as the input of preset neural network prediction model.
6. according to the method described in claim 5, it is characterized in that, the predictions request of the ONU according to the current period
Band data sends the first bandwidth authorization message to the ONU, including：
The predictions request band data of the ONU of the current period is carried out at renormalization using preset renormalization algorithm
Reason；
Using the predictions request band data of the ONU of the current period after the anti-normalization processing as target prediction bandwidth on demand
Data；
Judge that each business datum in the target prediction bandwidth on demand data identifies corresponding bandwidth on demand size of data, is
No is more than that respective business datum identifies corresponding pre-set bandwidths threshold value；
It is more than the corresponding pre-set bandwidths of own service Data Identification if business datum identifies corresponding bandwidth on demand size of data
Threshold value identifies corresponding bandwidth on demand data to the business datum, and setting is identified as heavy duty；
If business datum identifies corresponding bandwidth on demand size of data, it is less than the corresponding default band of own service Data Identification
Wide threshold value identifies corresponding bandwidth on demand data to the business datum, and setting is identified as light load；
Corresponding bandwidth on demand Data Identification is identified according to the priority of pre-set business datum and business datum, it will be described
Business datum identifies corresponding bandwidth on demand data and is ranked up in target prediction bandwidth on demand data, obtains and business datum mark
Know the ranking results of corresponding bandwidth on demand data；
Wherein, the ranking results that corresponding bandwidth on demand data are identified with business datum include：The light of highest priority is born
The minimum light load of load, the light load of the heavy duty of highest priority, priority second, priority, priority second heavy burden
It carries, the heavy duty that priority is minimum；
The ranking results that corresponding bandwidth on demand data are identified according to the business datum send the first bandwidth to the ONU and award
Weigh information.
7. according to the method described in claim 1, it is characterized in that, the predictions request of the ONU according to the current period
Band data, to after the step of the ONU the first bandwidth authorization messages of transmission, including：
The band data of OLT transmission in current period is obtained as target transmission bandwidth data；
The target transmission bandwidth data are first bandwidth authorization messages of the ONU according to reception, the bandwidth transmitted from ONU to OLT
Data；
According to the target transmission bandwidth data, the preset neural network prediction model is updated.
8. the method according to the description of claim 7 is characterized in that the prediction of the ONU according to the current period
Bandwidth on demand data, to the ONU send the first bandwidth authorization message the step of after, further include：
According to the next cycle ONU received in current period the second bandwidth request informations sent and the object transmission number
According to the ONU the second bandwidth authorization messages of transmission；
Wherein, second bandwidth authorization message includes：The business datum mark and business datum mark are corresponding pre-
The time slot size of transmission channel can be occupied by surveying ONU in bandwidth on demand data and transmission cycle.
9. a kind of network bandwidth resources distributor, which is characterized in that the control being applied in software defined network SDN frameworks
Device, described device include：
Bandwidth prediction module, the first bandwidth request information for sending acquired previous period optical network unit ONU,
As the input of preset neural network prediction model, the pre- of current period ONU is obtained by the neural network prediction model
Survey bandwidth on demand data；
The first bandwidth request packet includes：Business datum identifies and business datum identifies corresponding bandwidth on demand data；
Wherein, the neural network prediction model is corresponded with business datum mark；The neural network prediction model
Including：The business datum identifies the mathematical operation relationship between corresponding bandwidth on demand and ONU predictions request band datas；Institute
Stating predictions request band data includes：The business datum mark and the business datum identify corresponding predictions request bandwidth number
According to；
Bandwidth authorizing module sends for the predictions request band data according to the ONU of the current period to the ONU
One bandwidth authorization message；
Wherein, first bandwidth authorization message includes：The business datum mark and business datum mark are corresponding pre-
The time slot size of transmission channel can be occupied by surveying ONU in bandwidth on demand data and transmission cycle.
10. a kind of electronic equipment, which is characterized in that including processor, communication interface, memory and communication bus, wherein processing
Device, communication interface, memory complete mutual communication by communication bus；
Memory, for storing computer program；
Processor when for executing the program stored on memory, realizes any method and steps of claim 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810172531.7A CN108449286B (en) | 2018-03-01 | 2018-03-01 | Network bandwidth resource allocation method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810172531.7A CN108449286B (en) | 2018-03-01 | 2018-03-01 | Network bandwidth resource allocation method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108449286A true CN108449286A (en) | 2018-08-24 |
| CN108449286B CN108449286B (en) | 2020-07-03 |
Family
ID=63193462
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810172531.7A Active CN108449286B (en) | 2018-03-01 | 2018-03-01 | Network bandwidth resource allocation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108449286B (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109598378A (en) * | 2018-11-30 | 2019-04-09 | 平安医疗健康管理股份有限公司 | Medical insurance prediction technique, device, equipment and computer readable storage medium |
| CN110198280A (en) * | 2019-05-28 | 2019-09-03 | 华南理工大学 | A kind of SDN link allocation method based on BP neural network |
| CN110535803A (en) * | 2019-09-03 | 2019-12-03 | 西南交通大学 | A kind of passive optical network uplink transmission receiving end demodulation method |
| CN111464323A (en) * | 2019-01-18 | 2020-07-28 | 北京沃东天骏信息技术有限公司 | Node bandwidth scheduling method and scheduling device |
| CN111491312A (en) * | 2019-01-28 | 2020-08-04 | 中国移动通信有限公司研究院 | Method and equipment for predicting, allocating, acquiring and training neural network of wireless resources |
| CN111565323A (en) * | 2020-03-23 | 2020-08-21 | 视联动力信息技术股份有限公司 | Flow control method and device, electronic equipment and storage medium |
| WO2020169031A1 (en) * | 2019-02-19 | 2020-08-27 | 中兴通讯股份有限公司 | Data transmission method and apparatus, ap, onu pon, networking, and storage medium |
| WO2021003811A1 (en) * | 2019-07-11 | 2021-01-14 | 网宿科技股份有限公司 | Deep learning-based bandwidth filtering method and device, server, and storage medium |
| CN112367708A (en) * | 2020-10-30 | 2021-02-12 | 新华三技术有限公司 | Network resource allocation method and device |
| CN112532459A (en) * | 2020-12-07 | 2021-03-19 | 郑州师范学院 | Bandwidth resource adjusting method, device and equipment |
| US11005689B2 (en) | 2019-07-11 | 2021-05-11 | Wangsu Science & Technology Co., Ltd. | Method and apparatus for bandwidth filtering based on deep learning, server and storage medium |
| CN114244767A (en) * | 2021-11-01 | 2022-03-25 | 北京邮电大学 | Load balancing-based link minimum end-to-end delay routing algorithm |
| CN114465915A (en) * | 2021-12-22 | 2022-05-10 | 天翼云科技有限公司 | CDN bandwidth prediction method, device and system and electronic equipment |
| CN114501353A (en) * | 2020-10-23 | 2022-05-13 | 维沃移动通信有限公司 | Method for sending and receiving communication information and communication equipment |
| CN114866145A (en) * | 2021-01-20 | 2022-08-05 | 上海诺基亚贝尔股份有限公司 | Method, apparatus, device and computer readable medium for optical communication |
| CN115002215A (en) * | 2022-04-11 | 2022-09-02 | 北京邮电大学 | Cloud-based government and enterprise-oriented resource allocation model training method and resource allocation method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1697348A (en) * | 2004-05-14 | 2005-11-16 | 上海贝尔阿尔卡特股份有限公司 | Dynamic bandwidth allocation method and device in multiple operation types, and optical line terminal |
| CN101512970A (en) * | 2005-04-15 | 2009-08-19 | 新泽西理工学院 | Dynamic bandwidth allocation and service differentiation for broadband passive optical networks |
| WO2011020918A1 (en) * | 2009-08-21 | 2011-02-24 | Telefonaktiebolaget L M Ericsson (Publ) | Method for soft bandwidth limiting in dynamic bandwidth allocation |
| CN105933064A (en) * | 2016-07-05 | 2016-09-07 | 北京邮电大学 | Dynamic bandwidth allocation method and apparatus |
| CN106209687A (en) * | 2016-07-12 | 2016-12-07 | 重庆邮电大学 | A kind of hybrid multiplex efficient distribution method of PON global resource |
-
2018
- 2018-03-01 CN CN201810172531.7A patent/CN108449286B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1697348A (en) * | 2004-05-14 | 2005-11-16 | 上海贝尔阿尔卡特股份有限公司 | Dynamic bandwidth allocation method and device in multiple operation types, and optical line terminal |
| CN101512970A (en) * | 2005-04-15 | 2009-08-19 | 新泽西理工学院 | Dynamic bandwidth allocation and service differentiation for broadband passive optical networks |
| WO2011020918A1 (en) * | 2009-08-21 | 2011-02-24 | Telefonaktiebolaget L M Ericsson (Publ) | Method for soft bandwidth limiting in dynamic bandwidth allocation |
| CN105933064A (en) * | 2016-07-05 | 2016-09-07 | 北京邮电大学 | Dynamic bandwidth allocation method and apparatus |
| CN106209687A (en) * | 2016-07-12 | 2016-12-07 | 重庆邮电大学 | A kind of hybrid multiplex efficient distribution method of PON global resource |
Cited By (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109598378A (en) * | 2018-11-30 | 2019-04-09 | 平安医疗健康管理股份有限公司 | Medical insurance prediction technique, device, equipment and computer readable storage medium |
| CN111464323A (en) * | 2019-01-18 | 2020-07-28 | 北京沃东天骏信息技术有限公司 | Node bandwidth scheduling method and scheduling device |
| CN111491312A (en) * | 2019-01-28 | 2020-08-04 | 中国移动通信有限公司研究院 | Method and equipment for predicting, allocating, acquiring and training neural network of wireless resources |
| WO2020169031A1 (en) * | 2019-02-19 | 2020-08-27 | 中兴通讯股份有限公司 | Data transmission method and apparatus, ap, onu pon, networking, and storage medium |
| US11743622B2 (en) | 2019-02-19 | 2023-08-29 | Zte Corporation | Data transmission method and apparatus, AP, ONU PON, networking and storage medium |
| CN110198280A (en) * | 2019-05-28 | 2019-09-03 | 华南理工大学 | A kind of SDN link allocation method based on BP neural network |
| US11005689B2 (en) | 2019-07-11 | 2021-05-11 | Wangsu Science & Technology Co., Ltd. | Method and apparatus for bandwidth filtering based on deep learning, server and storage medium |
| WO2021003811A1 (en) * | 2019-07-11 | 2021-01-14 | 网宿科技股份有限公司 | Deep learning-based bandwidth filtering method and device, server, and storage medium |
| EP3783840A4 (en) * | 2019-07-11 | 2021-08-11 | Wangsu Science & Technology Co., Ltd. | Deep learning-based bandwidth filtering method and device, server, and storage medium |
| CN110535803A (en) * | 2019-09-03 | 2019-12-03 | 西南交通大学 | A kind of passive optical network uplink transmission receiving end demodulation method |
| CN111565323A (en) * | 2020-03-23 | 2020-08-21 | 视联动力信息技术股份有限公司 | Flow control method and device, electronic equipment and storage medium |
| CN114501353A (en) * | 2020-10-23 | 2022-05-13 | 维沃移动通信有限公司 | Method for sending and receiving communication information and communication equipment |
| CN114501353B (en) * | 2020-10-23 | 2024-01-05 | 维沃移动通信有限公司 | Communication information sending and receiving method and communication equipment |
| CN112367708B (en) * | 2020-10-30 | 2023-05-26 | 新华三技术有限公司 | Network resource allocation method and device |
| CN112367708A (en) * | 2020-10-30 | 2021-02-12 | 新华三技术有限公司 | Network resource allocation method and device |
| CN112532459A (en) * | 2020-12-07 | 2021-03-19 | 郑州师范学院 | Bandwidth resource adjusting method, device and equipment |
| CN114866145A (en) * | 2021-01-20 | 2022-08-05 | 上海诺基亚贝尔股份有限公司 | Method, apparatus, device and computer readable medium for optical communication |
| CN114866145B (en) * | 2021-01-20 | 2024-02-09 | 上海诺基亚贝尔股份有限公司 | Method, apparatus, device and computer readable medium for optical communication |
| CN114244767A (en) * | 2021-11-01 | 2022-03-25 | 北京邮电大学 | Load balancing-based link minimum end-to-end delay routing algorithm |
| CN114244767B (en) * | 2021-11-01 | 2023-09-26 | 北京邮电大学 | Link minimum end-to-end delay routing algorithm based on load balancing |
| CN114465915A (en) * | 2021-12-22 | 2022-05-10 | 天翼云科技有限公司 | CDN bandwidth prediction method, device and system and electronic equipment |
| CN115002215A (en) * | 2022-04-11 | 2022-09-02 | 北京邮电大学 | Cloud-based government and enterprise-oriented resource allocation model training method and resource allocation method |
| CN115002215B (en) * | 2022-04-11 | 2023-12-05 | 北京邮电大学 | Cloud government enterprise oriented resource allocation model training method and resource allocation method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108449286B (en) | 2020-07-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108449286A (en) | Network bandwidth resources distribution method and device | |
| Bunyakitanon et al. | End-to-end performance-based autonomous VNF placement with adopted reinforcement learning | |
| CN113055308B (en) | Bandwidth scheduling method, traffic transmission method and related products | |
| CN109617826B (en) | Storm dynamic load balancing method based on cuckoo search | |
| WO2019184445A1 (en) | Service resource allocation | |
| CN102724103B (en) | Proxy server, hierarchical network system and distributed workload management method | |
| CN105471759A (en) | Network traffic scheduling method and apparatus for data centers | |
| CN111106999A (en) | IP-optical network communication service joint distribution method and device | |
| CN108965014A (en) | The service chaining backup method and system of QoS perception | |
| CN112887217B (en) | Control data packet sending method, model training method, device and system | |
| CN113408797A (en) | Method for generating flow-traffic prediction multi-time-sequence model, information sending method and device | |
| CN109802997A (en) | A kind of selection method and its relevant device of node device | |
| CN109361547B (en) | Network slice link deployment method and device | |
| CN109714795A (en) | A kind of method for managing resource, resource management system and device based on SDN network slice | |
| WO2020059763A1 (en) | Resource allocation calculation device and resource allocation calculation method | |
| Cao et al. | A deep reinforcement learning approach to multi-component job scheduling in edge computing | |
| CN113543160A (en) | 5G slice resource allocation method and device, computing equipment and computer storage medium | |
| CN114116207A (en) | Flow control method, device, equipment and system | |
| CN105335376B (en) | A kind of method for stream processing, apparatus and system | |
| Mobasheri et al. | Toward developing fog decision making on the transmission rate of various IoT devices based on reinforcement learning | |
| CN104486254A (en) | SDN network bandwidth control method and SDN network bandwidth control system | |
| CN115499306A (en) | Method and device for constructing traffic scheduling model, electronic equipment and storage medium | |
| CN114448838B (en) | System reliability evaluation method | |
| CN113949666A (en) | Flow control method, device, equipment and system | |
| CN113269339A (en) | Method and system for automatically creating and distributing network appointment tasks |
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 |