CN110391843A - Transmission quality prediction, routing resource and the system of multi-area optical network - Google Patents
Transmission quality prediction, routing resource and the system of multi-area optical network Download PDFInfo
- Publication number
- CN110391843A CN110391843A CN201910532179.8A CN201910532179A CN110391843A CN 110391843 A CN110391843 A CN 110391843A CN 201910532179 A CN201910532179 A CN 201910532179A CN 110391843 A CN110391843 A CN 110391843A
- Authority
- CN
- China
- Prior art keywords
- network
- path
- node
- module
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B10/00—Transmission systems employing electromagnetic waves other than radio-waves, e.g. infrared, visible or ultraviolet light, or employing corpuscular radiation, e.g. quantum communication
- H04B10/07—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems
- H04B10/075—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems using an in-service signal
- H04B10/079—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems using an in-service signal using measurements of the data signal
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B10/00—Transmission systems employing electromagnetic waves other than radio-waves, e.g. infrared, visible or ultraviolet light, or employing corpuscular radiation, e.g. quantum communication
- H04B10/07—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems
- H04B10/075—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems using an in-service signal
- H04B10/079—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems using an in-service signal using measurements of the data signal
- H04B10/0793—Network aspects, e.g. central monitoring of transmission parameters
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B10/00—Transmission systems employing electromagnetic waves other than radio-waves, e.g. infrared, visible or ultraviolet light, or employing corpuscular radiation, e.g. quantum communication
- H04B10/07—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems
- H04B10/075—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems using an in-service signal
- H04B10/079—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems using an in-service signal using measurements of the data signal
- H04B10/0795—Performance monitoring; Measurement of transmission parameters
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B10/00—Transmission systems employing electromagnetic waves other than radio-waves, e.g. infrared, visible or ultraviolet light, or employing corpuscular radiation, e.g. quantum communication
- H04B10/07—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems
- H04B10/075—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems using an in-service signal
- H04B10/079—Arrangements for monitoring or testing transmission systems; Arrangements for fault measurement of transmission systems using an in-service signal using measurements of the data signal
- H04B10/0795—Performance monitoring; Measurement of transmission parameters
- H04B10/07955—Monitoring or measuring power
Landscapes
- Physics & Mathematics (AREA)
- Electromagnetism (AREA)
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a kind of prediction of the transmission quality of multi-area optical network, routing resource and systems, the system comprises the OPM in transport plane, for by the network data of the corresponding node currently acquired be input to first nerves network module export after one layer of neural network computing the node abstract network data to the multi-area optical network control plane；Control the controller in plane, for being directed to path to be predicted, obtain the network data for constituting the abstract of the output of OPM corresponding to each node in the path, and input of the network data for the abstract that will acquire as nervus opticus network module, the transmission quality prediction result by the output of nervus opticus network module as the path.Transport plane and control plane cooperative intelligent are realized using the present invention, the independence of multi-area optical network and privacy can be overcome to constrain while reducing controller load, and the transmission quality prediction of high accuracy is provided, ensures the reliability of business.
Description
Technical field
The present invention relates to field of network transmission, particularly relate to transmission quality prediction, the Path selection of a kind of multi-area optical network
Method and system.
Background technique
Novel bandwidth applications, such as explosive expansion of multimedia application based on cloud, have pushed the internet traffic to be in
Exponential growth promotes the structure of optical communication network also to become increasingly complex, therefore guarantees that the transmission quality of business is a Xiang Shifen
Important and difficult task.
Current optical-fiber network operator usually guarantees optical path by considering the worst link condition and distributing biggish nargin
Transmission quality, to guarantee the performance decline that optical path is likely to occur in life cycle, but network money can be reduced significantly in this way
Source, therefore accurately optic path quality (QoT) estimation model is most important for improving optical-fiber network operational efficiency.
Recent study person by the approach application of artificial intelligence in the prediction of the transmission quality of optical-fiber network, artificial intelligence skill
Art to indicate high dimensional data and approximate complicated function.Such as the deep learning system based on artificial neural network passes through
Channel power is monitored to predict transmission quality and utilize the machine learning system of support vector machines (SVM) and k neighbour (KNN) algorithm
System is to predict transmission quality etc..These schemes learn the impairment parameter of optical-fiber network by training dataset, to obtain light
The transmission quality on road.
However, these models are obviously not directly applicable the scene of multi-area optical network, because they need to access each
The state of optical module, this obviously violates multiregion system independence and privacy.In fact, across multiple autonomous system (multiple domains
Optical-fiber network) because being limited by management constraint, administrative staff may maintain secrecy to some detailed network operations informations, and only disclose non-
Information in normal limited domain.Therefore, in multi-area optical network, it is one very arduous that optical path, which provides transmission quality prediction, between domain
Task.
Summary of the invention
The invention proposes a kind of prediction of the transmission quality of multi-area optical network, routing resource and systems, can overcome
The independence and privacy of multi-area optical network constrain, and provide the transmission quality prediction of high accuracy, are effectively provided to be subsequent
Source distribution lays the foundation, and ensures the reliability of business.
Based on above-mentioned purpose, the present invention provides a kind of transmission quality forecasting system of multi-area optical network, comprising:
The optical information networks functional module OPM that is set in the transport plane of multi-area optical network, for will currently acquire
The network data of corresponding node is input in first nerves network module with non-reversible activation primitive by one layer of neural network
Operation, and the control that the network data of the abstract for the node that operation obtains is sent to the multi-area optical network is put down
Face；
The controller being set in the control plane obtains for being directed to path to be predicted and constitutes each of the path
The network data of the abstract of the output of OPM corresponding to node, and the network data for the abstract that will acquire is as nervus opticus
The input of network module, the transmission quality prediction result by the output of nervus opticus network module as the path；
Wherein, nervus opticus network module is the pre- transmission Q factor for first passing through multiple paths in the multi-area optical network, with
And the network data training of the abstract of the output of OPM history corresponding to the node that is passed through of these paths obtains.
Wherein, the network data of the corresponding node of the acquisition is specially the network characterization data of the node；The net
Network characteristic is one of following data or any combination:
Congestion ratio, power, linkage length, EDFA noise coefficient, the channel occupancy of WDM, modulation format.
Preferably, the neuron node number of the first layer hidden layer in first nerves network module is less than input node
Number.
Preferably, being provided at least three-layer neural network in nervus opticus network module.
The present invention also provides a kind of route selection systems of multi-area optical network, comprising:
The transmission quality forecasting system of multi-area optical network as described above；
Be set to the multi-area optical network using the transmission policy module in plane, for by source node and destination node
Between several paths indicate that the controller carries out transmission quality to current path to be predicted as path to be predicted one by one
Prediction result；It is service selection one satisfactory according to the transmission quality prediction result in the path that the controller is sent
Path.
The present invention also provides a kind of transmission quality prediction techniques of multi-area optical network, comprising:
The corresponding section that the optical information networks functional module OPM being set in the transport plane of multi-area optical network will be acquired currently
The network data of point is input to the operation in first nerves network module with non-reversible activation primitive Jing Guo one layer of neural network,
And the network data of the abstract for the node that operation obtains is sent to the control plane of the multi-area optical network；
The controller in the control plane is set to for path to be predicted, obtains each node institute for constituting the path
The network data of the abstract of corresponding OPM output, and the network data for the abstract that will acquire is as nervus opticus network mould
The input of block, the path transmission prediction of quality result by the output of nervus opticus network module as the path；
Wherein, nervus opticus network module be the pre- historic transmission Q for first passing through multiple paths in the multi-area optical network because
What the network data training of the abstract of the output of OPM history corresponding to the node that son and these paths are passed through obtained.
Preferably, the network data of the corresponding node currently acquired is input to first nerves network module in the OPM
Before, further includes:
The network data currently acquired is standardized by the OPM；And
The network data currently acquired is input to first nerves network module by the OPM, specifically:
Network data after standardization is input to first nerves network module by the OPM.
The present invention also provides a kind of routing resources of multi-area optical network, comprising:
Each step in transmission quality prediction technique as described above；
Be set to the multi-area optical network will be between source node and destination node using the transmission policy module in plane
Several paths indicate that the controller carries out transmission quality prediction to current path to be predicted as path to be predicted one by one
As a result；And
It is one satisfactory road of service selection according to the transmission quality prediction result in the path that the controller is sent
Diameter.
Wherein, the transmission quality prediction result in the path sent according to the controller, for service selection one symbol
Desired path is closed, is specifically included:
The transmission quality prediction result in the path that the transmission policy module sends the controller and the transmission of setting
Quality threshold is compared；If the transmission quality prediction result in the path is greater than the transmission quality threshold value, for the industry
Business distributes the path.
In technical solution of the present invention, the optical information networks functional module that is set in the transport plane of multi-area optical network
The network data of the corresponding node currently acquired is input in first nerves network module and is passed through with non-reversible activation primitive by OPM
Cross the operation of one layer of neural network, and the network data of the abstract for the node that first nerves network module operation is obtained
It is sent to the control plane of the multi-area optical network；The controller in the control plane is set to for path to be predicted, is obtained
Take the network data for constituting the abstract of the output of OPM corresponding to each node in the path, and the net for the abstract that will acquire
Input of the network data as nervus opticus network module, the transmission matter by the output of nervus opticus network module as the path
Measure prediction result；Wherein, nervus opticus network module is the pre- transmission Q factor for first passing through multiple paths in multi-area optical network, with
And the network data training of the abstract of the output of OPM history corresponding to the node that is passed through of these paths obtains.In this way, passing
OPM in plane is sent to carry out operation by first nerves network module with non-reversible activation primitive, on the one hand due to can not root
The network data for being input to first nerves network module is parsed according to the network data of the abstract of output, can play encryption effect
Fruit, thus the features such as independence and privacy in the case of meeting multi-area optical network；On the other hand, the abstract that OPM is exported
Network data can reflect the feature of the network data acquired in transport plane, and the network data of these abstracts further may be used
Transmission quality prediction is carried out by the operation of nervus opticus network module with the controller being input into control plane, height is provided
The transmission quality of accuracy is predicted, is laid the foundation for the subsequent effective resource allocation of progress, is ensured the reliability of business.
More preferably, in technical solution of the present invention, the neuron node of the first layer hidden layer in first nerves network module
Number is less than input node number, the effect of data compression is played, in the nervus opticus for carrying out data transmission and carrying out controller
Load is substantially reduced when network module training, to improve network operation efficiency.
Detailed description of the invention
Fig. 1 be a kind of transmission quality prediction of multi-area optical network provided in an embodiment of the present invention, route selection system it is interior
Portion's structure chart；
Fig. 2 is the schematic diagram of internal structure of first nerves network module provided in an embodiment of the present invention；
Fig. 3 is the schematic diagram of internal structure of nervus opticus network module provided in an embodiment of the present invention；
Fig. 4 is a kind of transmission quality prediction of multi-area optical network provided in an embodiment of the present invention, routing resource process
Figure；
Fig. 5 is the method that transmission policy module provided in an embodiment of the present invention is one satisfactory path of service selection
Flow chart；
Fig. 6 is the training method flow chart of nervus opticus network module provided in an embodiment of the present invention.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with specific embodiment, and reference
Attached drawing, the present invention is described in more detail.
The embodiment of the present invention is described below in detail, examples of the embodiments are shown in the accompanying drawings, wherein from beginning to end
Same or similar label indicates same or similar element or element with the same or similar functions.Below with reference to attached
The embodiment of figure description is exemplary, and for explaining only the invention, and is not construed as limiting the claims.
Those skilled in the art of the present technique are appreciated that unless expressly stated, singular " one " used herein, " one
It is a ", " described " and "the" may also comprise plural form.It is to be further understood that when we claim element to be " connected " or " coupling
Connect " to another element when, it can be directly connected or coupled to other elements, or there may also be intermediary elements.In addition, this
In " connection " or " coupling " that uses may include being wirelessly connected or wireless coupling.Wording "and/or" used herein includes one
A or more associated whole for listing item or any cell and all combination.
It should be noted that all statements for using " first " and " second " are for differentiation two in the embodiment of the present invention
The non-equal entity of a same names or non-equal parameter, it is seen that " first " " second " only for the convenience of statement, does not answer
It is interpreted as the restriction to the embodiment of the present invention, subsequent embodiment no longer illustrates this one by one.
The present inventor is it is considered that SDN (Software Defined Network, software defined network) is overturned
The characteristic of traditional network vertical integration, SDN network are broken down into three faces i.e. transport plane, control plane, using plane；Wherein
Control plane is the allotment center of whole network, and all control logics all concentrate on control plane, by network-based control logic
It is decoupling (such as router, interchanger) with bottom hardware equipment, it is the Controlling model concentrated in logic by network evolution.And it is existing
The communication transmission quality Predicting Technique using intelligent algorithm often by complete intelligent algorithm operate in SDN control
In device processed, the load of controller is increased significantly；And most of scheme depends on data itself unduly, however multi-area optical network about
Under beam, communication transmission quality prediction algorithm in controller can not directly acquire the data of detailed bottom hardware equipment, so
It cannot guarantee transmission quality prediction result well.
Therefore, in technical solution of the present invention, the optical information networks function that is set in the transport plane of multi-area optical network
The network data of the corresponding node currently acquired is input in first nerves network module with non-reversible activation letter by module OPM
Operation of the number Jing Guo one layer of neural network, and the network of the abstract for the node that first nerves network module operation is obtained
Data are sent to the control plane of the multi-area optical network；The controller in the control plane is set to for road to be predicted
Diameter obtains the network data of the abstract of the output of OPM corresponding to each node for constituting the path, and the abstract that will acquire
Input of the network data as nervus opticus network module, the biography by the output of nervus opticus network module as the path
Transmission quality prediction result；Wherein, nervus opticus network module be the pre- transmission Q for first passing through multiple paths in multi-area optical network because
What the network data training of the abstract of the output of OPM history corresponding to the node that son and these paths are passed through obtained.This
Sample, the OPM in transport plane carries out operation by first nerves network module with non-reversible activation primitive, on the one hand due to nothing
Method parses the network data for being input to first nerves network module according to the network data of the abstract of output, can play and add
Close effect, thus the features such as independence and privacy in the case of meeting multi-area optical network；On the other hand, OPM output is abstract
The network data of change can reflect the feature of the network data acquired in transport plane, and the network data of these abstracts is into one
The controller that step can be input into control plane carries out transmission quality prediction by the operation of nervus opticus network module, mentions
For the transmission quality prediction of high accuracy, lays the foundation for the subsequent effective resource allocation of progress, ensure the reliability of business.
The technical solution for embodiment that the invention will now be described in detail with reference to the accompanying drawings.
A kind of transmission quality forecasting system of multi-area optical network provided in an embodiment of the present invention, structure is as shown in Figure 1, specific
Include: the OPM (optical information networks functional module) 101 being set in the transport plane of multi-area optical network, be set to multi-domain optical net
Controller 102 in the control plane of network.
Wherein, OPM101 can be multiple, can be set in different optical-fiber network domains.For example, as shown in Figure 1, first
Two optical nodes (optical node A1 and A2) and two OPM, the OPM of respectively A1 and A2 are provided in a optical-fiber network domain；Second
In a optical-fiber network domain there are three optical node (optical node B1, B2 and B3) and three OPM, i.e. the OPM and B3 of OPM, B2 of B1
OPM.OPM calculates board by plug-in AI and all has intelligent operational capability.
OPM101 is used to acquire the network data of corresponding node, and the network data currently acquired is input to first nerves
Operation in network module with non-reversible activation primitive Jing Guo one layer of neural network, and first nerves network module operation is obtained
The control plane of the multi-area optical network is sent to the network data of the abstract of the node.
Wherein, the first nerves network module in OPM101, which can be, is built in OPM101 and is also possible to outer be hung on
OPM101；The internal structure of first nerves network module can be as shown in Fig. 2, include one layer of neural network, wherein X1~X6 table
Show input node, a1~a4 indicates the neuron node of first layer hidden layer；First layer in first nerves network module is hidden
The neuron node number of layer can be less than input node number.In this way, the net of the abstract of first nerves network module output
The dimension of network data is input to the dimension of the network data of first nerves network module by being less than, to play the effect of data compression
Fruit can substantially reduce load when carrying out data transmission and carrying out the nervus opticus network module training of controller, to mention
High network operation efficiency.
Wherein, the network data of OPM101 acquisition specifically can be the network characterization data of node；The network characterization number
According to can be one of following data or any combination: congestion ratio, power, linkage length, EDFA (erbium-doped fiber amplifier) are made an uproar
Sonic system number, the channel occupancy of WDM (wavelength-division multiplex), modulation format.
Controller 102 is used to be directed to path to be predicted, obtains the output of OPM corresponding to each node for constituting the path
The network data of abstract, and input of the network data for the abstract that will acquire as nervus opticus network module, by second
Transmission quality prediction result of the output of neural network module as the path.
Wherein, the nervus opticus network module in controller 102, which can be, is built in controller 102, is also possible to outer
It is hung on controller 102；Be provided at least three-layer neural network in nervus opticus network module, nervus opticus network module it is specific
Structure can use common neural network structure, for example, as shown in Figure 3.
Nervus opticus network module be the pre- transmission Q factor for first passing through multiple paths in the multi-area optical network and this
What the network data training of the abstract of OPM101 history output corresponding to the node that a little paths are passed through obtained；And OPM101
The network data of the abstract of history output is to be transported according to the network data of OPM101 history acquisition through first nerves network module
It obtains.Neural network parameter in first nerves network module can be what those skilled in the art were rule of thumb arranged,
And do not have to participate in training.
A kind of route selection system of multi-area optical network provided in an embodiment of the present invention, structure is as shown in Figure 1, on including
The transmission quality forecasting system for the multi-area optical network stated includes above-mentioned is set in the transport plane of multi-area optical network
The controller 102 that OPM (optical information networks functional module) 101, is set in the control plane of multi-area optical network, is additionally wrapped
It includes: being set to the transmission policy module 103 using plane of multi-area optical network.
Transmission policy module 103 is used to regard several paths between source node and destination node as path to be predicted one by one
Indicate that the controller carries out transmission quality prediction result to current path to be predicted；The path sent according to the controller
Transmission quality prediction result is one satisfactory path of service selection.
Transmission quality prediction based on above-mentioned multi-area optical network, route selection system, provided in an embodiment of the present invention one
The transmission quality prediction of kind multi-area optical network, routing resource, detailed process is as shown in figure 4, include the following steps:
The current network characterization data of step S401:OPM101 acquisition corresponding node.
Specifically, OPM101 acquires current network characterization data；The network characterization data can be following data it
One or any combination: congestion ratio, power, linkage length, EDFA noise coefficient, the channel occupancy of WDM, modulation format.
Step S402:OPM101 pre-processes the network data of acquisition.
Specifically, the network data currently acquired can be standardized by OPM101, including network data is every
Index is translated, removal redundant data, format are converted, are fabricated to the data type that neural network module can be used directly etc..
The network data of the corresponding node currently acquired is input in first nerves network module by step S403:OPM101
Operation with non-reversible activation primitive Jing Guo one layer of neural network.
Specifically, the network data after standardization is input in first nerves network module with non-by OPM101
Operation of the reversible activation primitive Jing Guo one layer of neural network, obtains the network data of the abstract of the node.Wherein, non-can
Inverse activation primitive, for example can be ReLu, Ramp function etc..
Preferably, the neuron node number of the first layer hidden layer in first nerves network module is less than input node
Number；In this way, the dimension of the network data of the abstract of first nerves network module output, which will be less than, is input to first nerves network
The dimension of the network data of module is carrying out data transmission and is carrying out the second of controller to play the effect of data compression
Load can be substantially reduced when neural network module training, to improve network operation efficiency.
The network data of the abstract for the node that step S404:OPM101 exports first nerves network module transmits
To control plane.
Step S405: the controller 102 for controlling plane obtains the OPM output in different optical-fiber networks domain in the multi-area optical network
Abstract network data.
Specifically, controller 102 is directed to path to be predicted, and it is defeated to obtain OPM corresponding to each node for constituting the path
The network data of abstract out.
Step S406: input of the network data for the abstract that controller 102 will acquire as nervus opticus network module,
Transmission quality prediction result by the output of nervus opticus network module as the path.
The nervus opticus network module for the transmission quality prediction result that this step is used to calculate path is described in pre- first pass through
OPM101 corresponding to the node that the historic transmission Q factor in multiple paths and these paths are passed through in multi-area optical network is gone through
What the network data training of the abstract of history output obtained, specific training method will be in subsequent introduction.
The path transmission prediction of quality result that nervus opticus network module exports in this step can be the road of prediction
The transmission Q factor of diameter.
Step S407: the transmission quality prediction result transmission in the path that controller 102 exports nervus opticus network module
To using plane.
Step S408: using the transmission quality in the path that the transmission policy module 103 in plane is sent according to controller 102
Prediction result is one satisfactory path of service selection.
Specifically, transmission policy module 103 is the method in one satisfactory path of service selection, and detailed process can be with
As shown in figure 5, including the following steps:
Step S501: transmission policy module 103 according to network topology structure find out business source node and destination node it
Between several paths.
For example, as shown in Figure 1, possible transmission path has A1-A2-B1-B3 and A1-A2-B1-B2-B3 between A1 and B3.
Step S502: several paths between source node and destination node are used as one by one to pre- by transmission policy module 103
It surveys path instruction controller 102 and transmission quality prediction result is carried out to current path to be predicted.
For example, path A1-A2-B1-B3 and A1-A2-B1-B2-B3 can be successively used as to pre- by transmission policy module 103
Path is surveyed, instruction controller 102 successively carries out transmission quality prediction knot to path A1-A2-B1-B3 and A1-A2-B1-B2-B3
Fruit.
Step S503: the transmission quality prediction result in the path that policy module 103 is sent according to controller 102 is transmitted, is
One satisfactory path of service selection.
Specifically, transmission policy module 103 by the transmission quality prediction result in the path that the controller 102 is sent with set
Fixed transmission quality threshold value is compared；If the path transmission prediction of quality result is greater than the transmission quality threshold value, for
The traffic assignments path.
For example, if the transmission quality prediction result for the path A1-A2-B1-B2-B3 that controller 102 is sent is greater than the biography
Transmission quality threshold value, then the path for the traffic assignments is A1-A2-B1-B2-B3.
The training method of above-mentioned nervus opticus network module, detailed process is as shown in fig. 6, include the following steps:
Step S601: acquisition network data generates training set and verifying collection.
Specifically, it can use the network number of the abstract of the corresponding node of each OPM history output in multi-area optical network
According to and history acquisition multi-area optical network in multiple paths transmission Q factor generate training set and verifying collection；Wherein, it verifies
The transmission Q factor including these paths is concentrated, includes the network number of the abstract for the node that each path is passed through in training set
According to.
And the network data of the abstract of OPM history output is the network data according to the acquisition of OPM history through first nerves
Network module operation obtains.Neural network parameter in first nerves network module can be those skilled in the art according to warp
Setting is tested, and does not have to participate in training.
Step S602: being input to nervus opticus network module for the network data of the abstract of the node in training set, will
Verifying concentrates the transmission Q factor in the path being made of these nodes as label, and the output with nervus opticus network module is done
Compare, nervus opticus network module is trained using back-propagation algorithm.
In technical solution of the present invention, the optical information networks functional module that is set in the transport plane of multi-area optical network
The network data of the corresponding node currently acquired is input in first nerves network module and is passed through with non-reversible activation primitive by OPM
Cross the operation of one layer of neural network, and the network data of the abstract for the node that first nerves network module operation is obtained
It is sent to the control plane of the multi-area optical network；The controller in the control plane is set to for path to be predicted, is obtained
Take the network data for constituting the abstract of the output of OPM corresponding to each node in the path, and the net for the abstract that will acquire
Input of the network data as nervus opticus network module, the transmission matter by the output of nervus opticus network module as the path
Measure prediction result；Wherein, nervus opticus network module is the pre- transmission Q factor for first passing through multiple paths in multi-area optical network, with
And the network data training of the abstract of the output of OPM history corresponding to the node that is passed through of these paths obtains.In this way, passing
OPM in plane is sent to carry out operation by first nerves network module with non-reversible activation primitive, on the one hand due to can not root
The network data for being input to first nerves network module is parsed according to the network data of the abstract of output, can play encryption effect
Fruit, to the features such as independence and privacy in the case of meeting multi-area optical network while reduce controller load；On the other hand,
The network data of the abstract of OPM output can reflect the feature of the network data acquired in transport plane, and these are abstracted
Network data may further be input into control plane in controller by nervus opticus network module operation carry out
Transmission quality prediction provides the transmission quality prediction of high accuracy, lays the foundation, ensures for the subsequent effective resource allocation of progress
The reliability of business.
More preferably, in technical solution of the present invention, the neuron node of the first layer hidden layer in first nerves network module
Number is less than input node number, the effect of data compression is played, in the nervus opticus for carrying out data transmission and carrying out controller
Load is substantially reduced when network module training, to improve network operation efficiency.
Those skilled in the art of the present technique have been appreciated that in the present invention the various operations crossed by discussion, method, in process
Steps, measures, and schemes can be replaced, changed, combined or be deleted.Further, each with having been crossed by discussion in the present invention
Kind of operation, method, other steps, measures, and schemes in process may also be alternated, changed, rearranged, decomposed, combined or deleted.
Further, in the prior art to have and the step in various operations, method disclosed in the present invention, process, measure, scheme
It may also be alternated, changed, rearranged, decomposed, combined or deleted.
It should be understood by those ordinary skilled in the art that: the discussion of any of the above embodiment is exemplary only, not
It is intended to imply that the scope of the present disclosure (including claim) is limited to these examples；Under thinking of the invention, above embodiments
Or can also be combined between the technical characteristic in different embodiments, step can be realized with random order, and be existed such as
Many other variations of the upper different aspect of the invention, for simplicity, they are not provided in details.Therefore, it is all
Within the spirit and principles in the present invention, any omission, modification, equivalent replacement, improvement for being made etc. be should be included in of the invention
Within protection scope.
Claims (10)
1. a kind of transmission quality forecasting system of multi-area optical network characterized by comprising
The optical information networks functional module OPM being set in the transport plane of multi-area optical network, the correspondence for will currently acquire
The network data of node is input to the fortune in first nerves network module with non-reversible activation primitive Jing Guo one layer of neural network
It calculates, and the network data of the abstract for the node that operation obtains is sent to the control plane of the multi-area optical network；
The controller being set in the control plane obtains each node for constituting the path for being directed to path to be predicted
The network data of the abstract of corresponding OPM output, and the network data for the abstract that will acquire is as nervus opticus network
The input of module, the transmission quality prediction result by the output of nervus opticus network module as the path；
Wherein, nervus opticus network module be the pre- transmission Q factor for first passing through multiple paths in the multi-area optical network and this
What the network data training of the abstract of OPM history output corresponding to the node that a little paths are passed through obtained.
2. system according to claim 1, which is characterized in that the network data of the corresponding node of the acquisition is specially institute
State the network characterization data of node；The network characterization data are one of following data or any combination:
Congestion ratio, power, linkage length, EDFA noise coefficient, the channel occupancy of WDM, modulation format.
3. system according to claim 1, which is characterized in that the mind of the first layer hidden layer in first nerves network module
Input node number is less than through first node number.
4. system according to claim 1, which is characterized in that be provided at least three layers of nerve in nervus opticus network module
Network.
5. a kind of route selection system of multi-area optical network characterized by comprising
The transmission quality forecasting system of multi-area optical network as described in claim 1-4 is any；
Be set to the multi-area optical network using the transmission policy module in plane, for will be between source node and destination node
Several paths indicate that the controller carries out transmission quality prediction to current path to be predicted as path to be predicted one by one
As a result；It is one satisfactory path of service selection according to the transmission quality prediction result in the path that the controller is sent.
6. a kind of transmission quality prediction technique of multi-area optical network characterized by comprising
The optical information networks functional module OPM being set in the transport plane of multi-area optical network is by the corresponding node currently acquired
Network data is input to the operation in first nerves network module with non-reversible activation primitive Jing Guo one layer of neural network, and will
The network data of the abstract for the node that operation obtains is sent to the control plane of the multi-area optical network；
The controller in the control plane is set to for path to be predicted, is obtained corresponding to each node for constituting the path
OPM output abstract network data, and the network data for the abstract that will acquire is as nervus opticus network module
Input, the path transmission prediction of quality result by the output of nervus opticus network module as the path；
Wherein, nervus opticus network module is the pre- historic transmission Q factor for first passing through multiple paths in the multi-area optical network, with
And the network data training of the abstract of the output of OPM history corresponding to the node that is passed through of these paths obtains.
7. according to the method described in claim 6, it is characterized in that, the network data of the corresponding node of the acquisition is specially institute
State the network characterization data of node；The network characterization data are one of following data or any combination:
Congestion ratio, power, linkage length, EDFA noise coefficient, the channel occupancy of WDM, modulation format.
8. according to the method described in claim 6, it is characterized in that, in the OPM by the network of the corresponding node currently acquired
Data are input to before first nerves network module, further includes:
The network data currently acquired is standardized by the OPM；And
The network data currently acquired is input to first nerves network module by the OPM, specifically:
Network data after standardization is input to first nerves network module by the OPM.
9. a kind of routing resource of multi-area optical network characterized by comprising
Such as each step in transmission quality prediction technique as claimed in claim 6 to 8；
If be set to the multi-area optical network will be between source node and destination node using the transmission policy module in plane
Main line diameter indicates that the controller carries out transmission quality prediction result to current path to be predicted as path to be predicted one by one；
And
It is one satisfactory path of service selection according to the transmission quality prediction result in the path that the controller is sent.
10. according to the method described in claim 9, it is characterized in that, the transmission in the path sent according to the controller
Prediction of quality specifically includes as a result, for one satisfactory path of service selection:
The transmission quality prediction result in the path that the transmission policy module sends the controller and the transmission quality of setting
Threshold value is compared；If the transmission quality prediction result in the path is greater than the transmission quality threshold value, for the business point
With the path.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910532179.8A CN110391843B (en) | 2019-06-19 | 2019-06-19 | Transmission quality prediction and path selection method and system for multi-domain optical network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910532179.8A CN110391843B (en) | 2019-06-19 | 2019-06-19 | Transmission quality prediction and path selection method and system for multi-domain optical network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110391843A true CN110391843A (en) | 2019-10-29 |
| CN110391843B CN110391843B (en) | 2021-01-05 |
Family
ID=68285607
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910532179.8A Active CN110391843B (en) | 2019-06-19 | 2019-06-19 | Transmission quality prediction and path selection method and system for multi-domain optical network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110391843B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113627541A (en) * | 2021-08-13 | 2021-11-09 | 北京邮电大学 | Light path transmission quality prediction method based on sample migration screening |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104654024A (en) * | 2015-02-12 | 2015-05-27 | 常州大学 | Method for locating and analyzing leakage of city gas pipeline based on GRNN (Generalized Regression Neural Network) |
| CN107749779A (en) * | 2017-09-29 | 2018-03-02 | 北京邮电大学 | A kind of pulse amplitude modulation PAM n-signals optical information networks method and apparatus |
| CN109241291A (en) * | 2018-07-18 | 2019-01-18 | 华南师范大学 | Knowledge mapping optimal path inquiry system and method based on deeply study |
-
2019
- 2019-06-19 CN CN201910532179.8A patent/CN110391843B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104654024A (en) * | 2015-02-12 | 2015-05-27 | 常州大学 | Method for locating and analyzing leakage of city gas pipeline based on GRNN (Generalized Regression Neural Network) |
| CN107749779A (en) * | 2017-09-29 | 2018-03-02 | 北京邮电大学 | A kind of pulse amplitude modulation PAM n-signals optical information networks method and apparatus |
| CN109241291A (en) * | 2018-07-18 | 2019-01-18 | 华南师范大学 | Knowledge mapping optimal path inquiry system and method based on deeply study |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113627541A (en) * | 2021-08-13 | 2021-11-09 | 北京邮电大学 | Light path transmission quality prediction method based on sample migration screening |
| CN113627541B (en) * | 2021-08-13 | 2023-07-21 | 北京邮电大学 | Optical path transmission quality prediction method based on sample migration screening |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110391843B (en) | 2021-01-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yang et al. | BrainIoT: Brain-like productive services provisioning with federated learning in industrial IoT | |
| CA2461069C (en) | System and method for fault diagnosis using distributed alarm correlation | |
| CN104486194B (en) | A kind of more reliability step virtual network controls system and method | |
| CN106302153A (en) | Multiple domain controller, single domain controller, software definition optical network system and method | |
| CN109995435A (en) | A kind of mapping method of virtual network and system based on multi-fiber core optical-fiber network | |
| CN105703973B (en) | A kind of power communication fiber optic network reliability consideration method based on composite measure | |
| Liu et al. | DL-assisted cross-layer orchestration in software-defined IP-over-EONs: From algorithm design to system prototype | |
| CN108650011A (en) | Multiple no-manned plane and ground network efficient collaboration, both data relay method and system | |
| Chen et al. | ADMIRE: demonstration of collaborative data-driven and model-driven intelligent routing engine for IP/optical cross-layer optimization in X-Haul networks | |
| Panayiotou et al. | Centralized and distributed machine learning-based QoT estimation for sliceable optical networks | |
| CN105430538B (en) | A kind of inter-domain routing method based on optical-fiber network subtopology figure | |
| CN103441956B (en) | A kind of cognitive network resource management method based on resource space vector calculus | |
| CN110391843A (en) | Transmission quality prediction, routing resource and the system of multi-area optical network | |
| Alvizu et al. | Machine-learning-based prediction and optimization of mobile metro-core networks | |
| Wu et al. | Dynamic network topology portrait for digital twin optical network | |
| CN104202262A (en) | Method and device for spectrum allocation in spectral flexible optical network | |
| US11387926B2 (en) | Efficient transfer of sensor data on dynamic software defined network (SDN) controlled optical network | |
| Goścień | Traffic-aware dedicated path protection in elastic optical networks | |
| CN107181680A (en) | A kind of method for realizing SDO functions, system and SDON systems | |
| Savva et al. | Deep graph learning for QoT estimation of unseen optical sub-network states: Capturing the crosstalk impact on the in-service lightpaths | |
| Meng et al. | Intelligent routing orchestration for ultra-low latency transport networks | |
| Mobasheri et al. | Toward developing fog decision making on the transmission rate of various IoT devices based on reinforcement learning | |
| CN107864093A (en) | A kind of multilayer union route and Survivability Strategy based on economic factors | |
| CN109842522A (en) | Communicate the configuration method and system of route resource | |
| CN106716935A (en) | Cross-layer service configuration method and controller |
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 |