CN110391843B - Transmission quality prediction and path selection method and system for multi-domain optical network - Google Patents
Transmission quality prediction and path selection method and system for multi-domain optical network Download PDFInfo
- Publication number
- CN110391843B CN110391843B CN201910532179.8A CN201910532179A CN110391843B CN 110391843 B CN110391843 B CN 110391843B CN 201910532179 A CN201910532179 A CN 201910532179A CN 110391843 B CN110391843 B CN 110391843B
- Authority
- CN
- China
- Prior art keywords
- network
- path
- transmission quality
- neural network
- module
- 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.)
- Active
Links
Images
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
Abstract
The invention discloses a method and a system for transmission quality prediction and path selection of a multi-domain optical network, wherein the system comprises the following steps: the OPM in the transmission plane is used for inputting the currently acquired network data of the corresponding node into the first neural network module, outputting the abstracted network data of the node to the control plane of the multi-domain optical network after one layer of neural network operation; and the controller in the control plane is used for acquiring the abstracted network data output by the OPM corresponding to each node forming the path aiming at the path to be predicted, taking the acquired abstracted network data as the input of the second neural network module, and taking the output of the second neural network module as the transmission quality prediction result of the path. The invention realizes the cooperative intelligence of the transmission plane and the control plane, can overcome the autonomous and privacy constraints of the multi-domain optical network, simultaneously reduces the load of the controller, provides high-accuracy transmission quality prediction and ensures the reliability of services.
Description
Technical Field
The present invention relates to the field of network transmission, and in particular, to a method and a system for transmission quality prediction and path selection in a multi-domain optical network.
Background
Novel high-bandwidth applications, such as explosive expansion of cloud-based multimedia applications, promote exponential growth of internet traffic, and promote increasingly complex structures of optical communication networks, so that ensuring the transmission quality of services is an important and arduous task.
The current optical network operator generally guarantees the transmission quality of an optical path by considering the worst link condition and allocating a larger margin, so as to guarantee the performance degradation of the optical path which may occur in the life cycle, but the network resources are greatly reduced, so that an accurate optical path transmission quality (QoT) estimation model is crucial to improving the operation efficiency of the optical network.
In recent years researchers have applied artificial intelligence methods to the prediction of the transmission quality of optical networks, and artificial intelligence techniques have made it possible to represent high dimensional data and approximate complex functions. Such as an artificial neural network-based deep learning system that predicts transmission quality by monitoring channel power and a machine learning system that utilizes a Support Vector Machine (SVM) and k-nearest neighbor (KNN) algorithm, and so on. The schemes learn the damage parameters of the optical network through a training data set, so as to obtain the transmission quality of the optical path.
However, these models are clearly not directly applicable to the scenario of multi-domain optical networks, as they require access to the state of each optical component, which clearly violates multi-domain system autonomy and privacy. In fact, across multiple autonomous systems (multi-domain optical networks), because of the regulatory constraints, managers may keep some detailed network operational information secret, but only reveal very limited intra-domain information. Therefore, in the multi-domain optical network, providing transmission quality prediction for the inter-domain optical path is a very difficult task.
Disclosure of Invention
The invention provides a transmission quality prediction and path selection method and system of a multi-domain optical network, which can overcome the autonomous and privacy constraints of the multi-domain optical network, provide high-accuracy transmission quality prediction, lay a foundation for effective resource allocation in the follow-up process and ensure the reliability of services.
Based on the above object, the present invention provides a transmission quality prediction system for a multi-domain optical network, comprising:
the optical performance monitoring module OPM is arranged in a transmission plane of a multi-domain optical network and is used for inputting currently acquired network data of a corresponding node into a first neural network module to be subjected to operation of a layer of neural network by a non-reversible activation function and sending the abstracted network data of the node obtained by operation to a control plane of the multi-domain optical network;
the controller is arranged in the control plane and used for acquiring abstracted network data output by the OPM corresponding to each node forming the path aiming at the path to be predicted, taking the acquired abstracted network data as the input of the second neural network module, and taking the output of the second neural network module as the transmission quality prediction result of the path;
the second neural network module is obtained by training abstracted network data output by OPM history corresponding to nodes passed by a plurality of paths of the multi-domain optical network in advance through transmission Q factors of the paths.
The acquired network data of the corresponding node is specifically network characteristic data of the node; the network characteristic data is one or any combination of the following data:
congestion rate, power, link length, EDFA noise figure, WDM channel occupancy, modulation format.
Preferably, the number of neuron nodes of the first layer of hidden layer in the first neural network module is less than the number of input nodes.
Preferably, at least three layers of neural networks are arranged in the second neural network module.
The present invention also provides a path selection system of a multi-domain optical network, including:
a transmission quality prediction system for a multi-domain optical network as described above;
the transmission strategy module is arranged in an application plane of the multi-domain optical network and used for indicating a plurality of paths between a source node and a destination node as paths to be predicted one by one to the controller to carry out transmission quality prediction on the current paths to be predicted; and selecting a path meeting the requirement for the service according to the transmission quality prediction result of the path sent by the controller.
The invention also provides a method for predicting the transmission quality of the multi-domain optical network, which comprises the following steps:
an optical performance monitoring function module OPM arranged in a transmission plane of a multi-domain optical network inputs currently acquired network data of a corresponding node into a first neural network module to be subjected to operation of a layer of neural network by a non-reversible activation function, and sends abstracted network data of the node obtained by the operation to a control plane of the multi-domain optical network;
the controller arranged in the control plane acquires abstracted network data output by the OPM corresponding to each node forming the path aiming at the path to be predicted, the acquired abstracted network data is used as the input of the second neural network module, and the output of the second neural network module is used as the path transmission quality prediction result of the path;
the second neural network module is obtained by training abstracted network data of historical transmission Q factors of a plurality of paths in the multi-domain optical network and historical output of OPM corresponding to nodes passed by the paths in advance.
Preferably, before the OPM inputs the currently acquired network data of the corresponding node to the first neural network module, the method further includes:
the OPM standardizes the currently acquired network data; and
the OPM inputs currently acquired network data to a first neural network module, specifically:
the OPM inputs the normalized network data to a first neural network module.
The invention also provides a path selection method of the multi-domain optical network, which comprises the following steps:
the steps in the transmission quality prediction method as described above;
a transmission strategy module arranged in an application plane of the multi-domain optical network takes a plurality of paths between a source node and a destination node as paths to be predicted one by one and instructs the controller to carry out transmission quality prediction on the current paths to be predicted; and are
And selecting a path meeting the requirement for the service according to the transmission quality prediction result of the path sent by the controller.
Wherein, the selecting a path meeting the requirement for the service according to the transmission quality prediction result of the path sent by the controller specifically includes:
the transmission strategy module compares the transmission quality prediction result of the path sent by the controller with a set transmission quality threshold; and if the transmission quality prediction result of the path is greater than the transmission quality threshold, allocating the path for the service.
In the technical scheme of the invention, an optical performance monitoring function module OPM arranged in a transmission plane of a multi-domain optical network inputs currently acquired network data of a corresponding node into a first neural network module to be subjected to operation of a layer of neural network by a non-reversible activation function, and sends abstracted network data of the node obtained by the operation of the first neural network module to a control plane of the multi-domain optical network; the controller arranged in the control plane acquires abstracted network data output by the OPM corresponding to each node forming the path aiming at the path to be predicted, the acquired abstracted network data is used as the input of the second neural network module, and the output of the second neural network module is used as the transmission quality prediction result of the path; the second neural network module is obtained by training abstracted network data of transmission Q factors of a plurality of paths in the multi-domain optical network and historical OPM output corresponding to nodes passed by the paths in advance. In this way, the OPM in the transmission plane is operated by the first neural network module through the irreversible activation function, on one hand, the network data input to the first neural network module cannot be analyzed according to the output abstract network data, so that the encryption effect can be achieved, and the characteristics of autonomy, privacy and the like under the condition of a multi-domain optical network are met; on the other hand, the abstracted network data output by the OPM can reflect the characteristics of the network data collected in the transmission plane, and the abstracted network data can be further input into the controller in the control plane to carry out transmission quality prediction through the operation of the second neural network module, so that high-accuracy transmission quality prediction is provided, a foundation is laid for effective resource allocation in the follow-up process, and the reliability of the service is guaranteed.
Preferably, in the technical scheme of the invention, the number of the neuron nodes of the first hidden layer in the first neural network module is less than that of the input nodes, so that the effect of data compression is achieved, and the load is greatly reduced when data transmission and training of the second neural network module of the controller are carried out, thereby improving the network operation efficiency.
Drawings
Fig. 1 is an internal structure diagram of a transmission quality prediction and path selection system of a multi-domain optical network according to an embodiment of the present invention;
fig. 2 is a schematic internal structural diagram of a first neural network module according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of an internal structure of a second neural network module according to an embodiment of the present invention;
fig. 4 is a flowchart of a transmission quality prediction and path selection method for a multi-domain optical network according to an embodiment of the present invention;
fig. 5 is a flowchart of a method for selecting a satisfactory path for a service by a transmission policy module according to an embodiment of the present invention;
fig. 6 is a flowchart of a training method of the second neural network module according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to specific embodiments and the accompanying drawings.
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative only and should not be construed as limiting the invention.
As used herein, the singular forms "a", "an", "the" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that when an element is referred to as being "connected" or "coupled" to another element, it can be directly connected or coupled to the other element or intervening elements may also be present. Further, "connected" or "coupled" as used herein may include wirelessly connected or wirelessly coupled. As used herein, the term "and/or" includes all or any element and all combinations of one or more of the associated listed items.
It should be noted that all expressions using "first" and "second" in the embodiments of the present invention are used for distinguishing two entities with the same name but different names or different parameters, and it should be noted that "first" and "second" are merely for convenience of description and should not be construed as limitations of the embodiments of the present invention, and they are not described in any more detail in the following embodiments.
The inventor considers that the SDN (Software Defined Network) subverts the vertical integration characteristic of the traditional Network, and the SDN is decomposed into three surfaces, namely a transmission plane, a control plane and an application plane; the control plane is a deployment center of the whole network, all control logics are centralized on the control plane, the control logics of the network are decoupled from bottom hardware equipment (such as a router and a switch), and the network is evolved into a logically centralized control model. In the existing communication transmission quality prediction technology using the artificial intelligence algorithm, the complete artificial intelligence algorithm is usually operated in the SDN controller, so that the load of the controller is greatly increased; most schemes rely on data, however, under the constraint of a multi-domain optical network, a communication transmission quality prediction algorithm in a controller cannot directly acquire detailed data of bottom-layer hardware equipment, so that a transmission quality prediction result cannot be well guaranteed.
Therefore, in the technical scheme of the present invention, an optical performance monitoring function module OPM disposed in a transmission plane of a multi-domain optical network inputs currently acquired network data of a corresponding node into a first neural network module to perform a layer of neural network operation with a non-reversible activation function, and sends abstracted network data of the node obtained by the first neural network module operation to a control plane of the multi-domain optical network; the controller arranged in the control plane acquires abstracted network data output by the OPM corresponding to each node forming the path aiming at the path to be predicted, the acquired abstracted network data is used as the input of the second neural network module, and the output of the second neural network module is used as the transmission quality prediction result of the path; the second neural network module is obtained by training abstracted network data of transmission Q factors of a plurality of paths in the multi-domain optical network and historical OPM output corresponding to nodes passed by the paths in advance. In this way, the OPM in the transmission plane is operated by the first neural network module through the irreversible activation function, on one hand, the network data input to the first neural network module cannot be analyzed according to the output abstract network data, so that the encryption effect can be achieved, and the characteristics of autonomy, privacy and the like under the condition of a multi-domain optical network are met; on the other hand, the abstracted network data output by the OPM can reflect the characteristics of the network data collected in the transmission plane, and the abstracted network data can be further input into the controller in the control plane to carry out transmission quality prediction through the operation of the second neural network module, so that high-accuracy transmission quality prediction is provided, a foundation is laid for effective resource allocation in the follow-up process, and the reliability of the service is guaranteed.
The technical solution of the embodiments of the present invention is described in detail below with reference to the accompanying drawings.
The transmission quality prediction system of a multi-domain optical network provided by the embodiment of the present invention has a structure as shown in fig. 1, and specifically includes: an OPM (optical performance monitoring function) 101 disposed in a transport plane of the multi-domain optical network, a controller 102 disposed in a control plane of the multi-domain optical network.
The number of OPMs 101 may be multiple, and may be located in different optical network domains. For example, as shown in fig. 1, two optical nodes (optical nodes a1 and a2) and two OPMs, an OPM of a1 and a2, respectively, are provided in the first optical network domain; there are three optical nodes (optical nodes B1, B2, and B3) and three OPMs in the second optical network domain, namely the OPM of B1, the OPM of B2, and the OPM of B3. The OPM has intelligent operation capability through the external AI computing board.
The OPM101 is used for acquiring network data of a corresponding node, inputting the currently acquired network data into the first neural network module, performing operation on the network data through a layer of neural network by using a non-reversible activation function, and sending the abstract network data of the node obtained by the operation of the first neural network module to the control plane of the multi-domain optical network.
The first neural network module in the OPM101 may be built in the OPM101 or be externally hung on the OPM 101; the internal structure of the first neural network module may be as shown in fig. 2, and includes a layer of neural network, where X1-X6 represent input nodes, and a1-a 4 represent neuron nodes of a first hidden layer; the number of neuron nodes of the first layer hidden layer in the first neural network module may be less than the number of input nodes. Therefore, the dimension of the abstracted network data output by the first neural network module is less than that of the network data input to the first neural network module, so that the effect of data compression is achieved, the load can be greatly reduced when data transmission and training of the second neural network module of the controller are carried out, and the network operation efficiency is improved.
The network data collected by the OPM101 may specifically be network characteristic data of a node; the network characteristic data can be one or any combination of the following data: congestion rate, power, link length, noise figure of EDFA (erbium doped fiber amplifier), channel occupancy of WDM (wavelength division multiplexing), modulation format.
The controller 102 is configured to, for a path to be predicted, obtain abstracted network data output by the OPM corresponding to each node constituting the path, use the obtained abstracted network data as an input of the second neural network module, and use an output of the second neural network module as a transmission quality prediction result of the path.
The second neural network module in the controller 102 may be built in the controller 102, or may be externally hung on the controller 102; at least three layers of neural networks are arranged in the second neural network module, and the specific structure of the second neural network module can adopt a commonly used neural network structure, for example, as shown in fig. 3.
The second neural network module is obtained by training abstracted network data which is output by OPM101 history and corresponds to nodes passed by a plurality of paths of the multi-domain optical network in advance; and the abstracted network data output historically by the OPM101 is obtained by the operation of a first neural network module according to the network data collected historically by the OPM 101. The neural network parameters in the first neural network module may be set empirically by one skilled in the art without participation in training.
The path selection system of a multi-domain optical network according to an embodiment of the present invention is configured as shown in fig. 1, and includes the transmission quality prediction system of the multi-domain optical network, that is, includes the OPM (optical performance monitoring function module) 101 disposed in the transmission plane of the multi-domain optical network, and the controller 102 disposed in the control plane of the multi-domain optical network, and further includes: and the transmission strategy module 103 is arranged on an application plane of the multi-domain optical network.
The transmission policy module 103 is configured to instruct the controller to perform a transmission quality prediction result on a current path to be predicted by using a plurality of paths between a source node and a destination node as the path to be predicted one by one; and selecting a path meeting the requirement for the service according to the path transmission quality prediction result sent by the controller.
Based on the transmission quality prediction and path selection system of the multi-domain optical network, a method for transmission quality prediction and path selection of the multi-domain optical network provided by the embodiment of the present invention has a specific flow as shown in fig. 4, and includes the following steps:
step S401: the OPM101 collects current network characteristic data of the corresponding node.
Specifically, the OPM101 collects current network characteristic data; the network characteristic data can be one or any combination of the following data: congestion rate, power, link length, EDFA noise figure, WDM channel occupancy, modulation format.
Step S402: the OPM101 pre-processes the acquired network data.
Specifically, the OPM101 may perform standardization processing on currently acquired network data, including translation of each index of the network data, removal of redundant data, format conversion, making of a data type that can be directly used by the neural network module, and the like.
Step S403: the OPM101 inputs the currently acquired network data of the corresponding node into the first neural network module to be operated by a layer of neural network through a non-reversible activation function.
Specifically, the OPM101 inputs the network data after the normalization processing into the first neural network module, and obtains the abstracted network data of the node through the operation of a layer of neural network by using the irreversible activation function. The irreversible activation function may be, for example, ReLu, Ramp function, or the like.
Preferably, the number of neuron nodes of the first layer of hidden layer in the first neural network module is less than the number of input nodes; therefore, the dimension of the abstracted network data output by the first neural network module is less than that of the network data input to the first neural network module, so that the effect of data compression is achieved, the load can be greatly reduced when data transmission and training of the second neural network module of the controller are carried out, and the network operation efficiency is improved.
Step S404: the OPM101 transfers the abstracted network data of the nodes output by the first neural network module to the control plane.
Step S405: the controller 102 of the control plane obtains abstracted network data output by OPMs of different optical network domains in the multi-domain optical network.
Specifically, the controller 102 acquires, for a path to be predicted, abstracted network data output by the OPM corresponding to each node constituting the path.
Step S406: the controller 102 takes the obtained abstracted network data as an input of the second neural network module, and takes an output of the second neural network module as a transmission quality prediction result of the path.
The second neural network module for calculating the transmission quality prediction result of the path in this step is obtained by training the historical transmission Q factors of multiple paths in the multi-domain optical network and abstracted network data output historically by the OPM101 corresponding to the nodes passed by these paths in advance, and a specific training method will be introduced later.
The path transmission quality prediction result output by the second neural network module in this step may be a predicted transmission Q factor of the path.
Step S407: the controller 102 transmits the transmission quality prediction result of the path output by the second neural network module to the application plane.
Step S408: the transmission policy module 103 in the application plane selects a path that meets the requirements for the service according to the transmission quality prediction result of the path sent by the controller 102.
Specifically, the method for the transmission policy module 103 to select a path meeting the requirement for the service, where a specific flow may be as shown in fig. 5, includes the following steps:
step S501: the transmission strategy module 103 finds out several paths between the source node and the destination node of the traffic according to the network topology.
For example, as shown in FIG. 1, possible transmission paths between A1 and B3 are A1-A2-B1-B3 and A1-A2-B1-B2-B3.
Step S502: the transmission policy module 103 instructs the controller 102 to perform transmission quality prediction on the current path to be predicted one by one using a plurality of paths between the source node and the destination node as the path to be predicted.
For example, the transmission policy module 103 may take the paths a1-a2-B1-B3 and a1-a2-B1-B2-B3 as paths to be predicted in turn, and instruct the controller 102 to perform transmission quality prediction results on the paths a1-a2-B1-B3 and a1-a2-B1-B2-B3 in turn.
Step S503: the transmission policy module 103 selects a path meeting the requirement for the service according to the transmission quality prediction result of the path sent by the controller 102.
Specifically, the transmission policy module 103 compares the transmission quality prediction result of the path sent by the controller 102 with a set transmission quality threshold; and if the path transmission quality prediction result is greater than the transmission quality threshold, allocating the path for the service.
For example, if the transmission quality prediction result of the path a1-a2-B1-B2-B3 sent by the controller 102 is greater than the transmission quality threshold, the path allocated for the traffic is a1-a 2-B1-B2-B3.
The specific flow of the training method for the second neural network module is shown in fig. 6, and the method includes the following steps:
step S601: and collecting network data to generate a training set and a verification set.
Specifically, a training set and a verification set can be generated by using abstracted network data of corresponding nodes historically output by each OPM in the multi-domain optical network and historically acquired transmission Q factors of multiple paths in the multi-domain optical network; wherein the verification set comprises transmission Q factors of the paths, and the training set comprises abstracted network data of nodes passed by each path.
And the abstracted network data output by the OPM history is obtained by the operation of the first neural network module according to the network data collected by the OPM history. The neural network parameters in the first neural network module may be set empirically by one skilled in the art without participation in training.
Step S602: and inputting the abstracted network data of the nodes in the training set into the second neural network module, comparing the transmission Q factor of the path formed by the nodes in the verification set with the output of the second neural network module by taking the transmission Q factor as a label, and training the second neural network module by adopting a back propagation algorithm.
In the technical scheme of the invention, an optical performance monitoring function module OPM arranged in a transmission plane of a multi-domain optical network inputs currently acquired network data of a corresponding node into a first neural network module to be subjected to operation of a layer of neural network by a non-reversible activation function, and sends abstracted network data of the node obtained by the operation of the first neural network module to a control plane of the multi-domain optical network; the controller arranged in the control plane acquires abstracted network data output by the OPM corresponding to each node forming the path aiming at the path to be predicted, the acquired abstracted network data is used as the input of the second neural network module, and the output of the second neural network module is used as the transmission quality prediction result of the path; the second neural network module is obtained by training abstracted network data of transmission Q factors of a plurality of paths in the multi-domain optical network and historical OPM output corresponding to nodes passed by the paths in advance. Therefore, the OPM in the transmission plane is operated by the first neural network module through the irreversible activation function, on one hand, the network data input to the first neural network module cannot be analyzed according to the output abstract network data, so that the encryption effect can be achieved, the characteristics of autonomy, privacy and the like under the condition of a multi-domain optical network are met, and the load of the controller is reduced; on the other hand, the abstracted network data output by the OPM can reflect the characteristics of the network data collected in the transmission plane, and the abstracted network data can be further input into the controller in the control plane to carry out transmission quality prediction through the operation of the second neural network module, so that high-accuracy transmission quality prediction is provided, a foundation is laid for effective resource allocation in the follow-up process, and the reliability of the service is guaranteed.
Preferably, in the technical scheme of the invention, the number of the neuron nodes of the first hidden layer in the first neural network module is less than that of the input nodes, so that the effect of data compression is achieved, and the load is greatly reduced when data transmission and training of the second neural network module of the controller are carried out, thereby improving the network operation efficiency.
Those of skill in the art will appreciate that various operations, methods, steps in the processes, acts, or solutions discussed in the present application may be alternated, modified, combined, or deleted. Further, various operations, methods, steps in the flows, which have been discussed in the present application, may be interchanged, modified, rearranged, decomposed, combined, or eliminated. Further, steps, measures, schemes in the various operations, methods, procedures disclosed in the prior art and the present invention can also be alternated, changed, rearranged, decomposed, combined, or deleted.
Those of ordinary skill in the art will understand that: the discussion of any embodiment above is meant to be exemplary only, and is not intended to intimate that the scope of the disclosure, including the claims, is limited to these examples; within the idea of the invention, also features in the above embodiments or in different embodiments may be combined, steps may be implemented in any order, and there are many other variations of the different aspects of the invention as described above, which are not provided in detail for the sake of brevity. Therefore, any omissions, modifications, substitutions, improvements and the like that may be made without departing from the spirit and principles of the invention are intended to be included within the scope of the invention.
Claims (10)
1. A transmission quality prediction system for a multi-domain optical network, comprising:
the optical performance monitoring module OPM is arranged in a transmission plane of a multi-domain optical network and is used for inputting currently acquired network data of a corresponding node into a first neural network module to be subjected to operation of a layer of neural network by a non-reversible activation function and sending the abstracted network data of the node obtained by operation to a control plane of the multi-domain optical network;
the controller is arranged in the control plane and used for acquiring abstracted network data output by the OPM corresponding to each node forming the path aiming at the path to be predicted, taking the acquired abstracted network data as the input of the second neural network module, and taking the output of the second neural network module as the transmission quality prediction result of the path;
the second neural network module is obtained by training abstracted network data output by OPM history corresponding to nodes passed by a plurality of paths of the multi-domain optical network in advance through transmission Q factors of the paths.
2. The system according to claim 1, wherein the collected network data of the corresponding node is specifically network characteristic data of the node; the network characteristic data is one or any combination of the following data:
congestion rate, power, link length, EDFA noise figure, WDM channel occupancy, modulation format.
3. The system of claim 1, wherein the number of neuron nodes of the first hidden layer in the first neural network module is less than the number of input nodes.
4. The system of claim 1, wherein at least three layers of neural networks are disposed in the second neural network module.
5. A path selection system for a multi-domain optical network, comprising:
a transmission quality prediction system for a multi-domain optical network according to any of claims 1-4;
the transmission strategy module is arranged in an application plane of the multi-domain optical network and used for indicating a plurality of paths between a source node and a destination node as paths to be predicted one by one to the controller to carry out transmission quality prediction on the current paths to be predicted; and selecting a path meeting the requirement for the service according to the transmission quality prediction result of the path sent by the controller.
6. A method for predicting transmission quality of a multi-domain optical network is characterized by comprising the following steps:
an optical performance monitoring function module OPM arranged in a transmission plane of a multi-domain optical network inputs currently acquired network data of a corresponding node into a first neural network module to be subjected to operation of a layer of neural network by a non-reversible activation function, and sends abstracted network data of the node obtained by the operation to a control plane of the multi-domain optical network;
the controller arranged in the control plane acquires abstracted network data output by the OPM corresponding to each node forming the path aiming at the path to be predicted, the acquired abstracted network data is used as the input of the second neural network module, and the output of the second neural network module is used as the path transmission quality prediction result of the path;
the second neural network module is obtained by training abstracted network data of historical transmission Q factors of a plurality of paths in the multi-domain optical network and historical output of OPM corresponding to nodes passed by the paths in advance.
7. The method according to claim 6, wherein the collected network data of the corresponding node is specifically network characteristic data of the node; the network characteristic data is one or any combination of the following data:
congestion rate, power, link length, noise figure of EDFA, channel occupancy of WDM, modulation format.
8. The method of claim 6, before the OPM inputs the currently acquired network data of the corresponding node to the first neural network module, further comprising:
and the OPM standardizes the currently acquired network data.
9. A method for selecting a path of a multi-domain optical network, comprising:
the steps in the transmission quality prediction method according to any of claims 6-8;
a transmission strategy module arranged in an application plane of the multi-domain optical network takes a plurality of paths between a source node and a destination node as paths to be predicted one by one and instructs the controller to carry out transmission quality prediction on the current paths to be predicted; and are
And selecting a path meeting the requirement for the service according to the transmission quality prediction result of the path sent by the controller.
10. The method according to claim 9, wherein the selecting a path meeting the requirement for the service according to the transmission quality prediction result of the path sent by the controller specifically comprises:
the transmission strategy module compares the transmission quality prediction result of the path sent by the controller with a set transmission quality threshold; and if the transmission quality prediction result of the path is greater than the transmission quality threshold, allocating the path for the service.
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 CN110391843A (en) | 2019-10-29 |
| CN110391843B true 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) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
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 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110391843A (en) | 2019-10-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Fang et al. | Joint spectrum and IT resource allocation for efficient vNF service chaining in inter-datacenter elastic optical networks | |
| Ou et al. | New and improved approaches for shared-path protection in WDM mesh networks | |
| Suárez-Varela et al. | Routing in optical transport networks with deep reinforcement learning | |
| Shi et al. | On the effect of bandwidth fragmentation on blocking probability in elastic optical networks | |
| Chen et al. | Knowledge-based autonomous service provisioning in multi-domain elastic optical networks | |
| Baldine et al. | Traffic adaptive WDM networks: A study of reconfiguration issues | |
| Zhang et al. | Overview on routing and resource allocation based machine learning in optical networks | |
| US10681435B2 (en) | Method and apparatus for optimizing dynamically the operation of an optical network | |
| Hai et al. | QoS-aware protection in elastic optical networks with distance-adaptive and reconfigurable modulation formats | |
| Chen et al. | A multi-task-learning-based transfer deep reinforcement learning design for autonomic optical networks | |
| Ye et al. | Federated traffic engineering with supervised learning in multi-region networks | |
| Hans et al. | Controller placement in software defined Internet of Things using optimization algorithm | |
| CN110391843B (en) | Transmission quality prediction and path selection method and system for multi-domain optical network | |
| Hernández-Chulde et al. | Assessment of a latency-aware routing and spectrum assignment mechanism based on deep reinforcement learning | |
| US10778562B2 (en) | Path computation in a multi-tenant network | |
| Tang et al. | Link allocation, routing and scheduling of FSO augmented RF wireless mesh networks | |
| Castro et al. | Opportunities for AI/ML in telecommunications networks | |
| Dave et al. | Elastic optical networks based optimization using machine learning: state-of-art review | |
| KR102089425B1 (en) | SDN control method | |
| Angelov et al. | Modelling and Simulation Analysis of Routing Algorithms in Multichannel Optical Communication Networks | |
| Lourenço et al. | A deep neural network with a fuzzy multi-objective optimization model for fault analysis in an elastic optical network | |
| Forsgren et al. | Telecommunications network design | |
| Aravind et al. | Multi-Objective Controller Failure Aware Capacitated Controller Placement in Software-Defined Networks | |
| Chang et al. | Using reinforcement learning to the priority-based routing and call admission control in WDM networks | |
| Evseeva et al. | Method for optimal resource management of transport software-defined networks |
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 |