Uplink signal processing method for passive optical network

Abstract

Provided in the embodiments of the present disclosure is an uplink signal processing method for a passive optical network. The method comprises: an optical line terminal (OLT) collecting, on an uplink bandwidth allocated to a target ONU in advance, a target uplink signal from the target ONU, so as to obtain characteristic data of the target uplink signal; and determining identity information of the target ONU on the basis of the uplink bandwidth, and determining a processing mode for the characteristic data.

Description

A method for processing upstream signals in a passive optical network

[0001] Cross-reference to Related Applications

[0002] The present disclosure is based on and claims priority to Chinese Patent Application No. CN202410701783.X entitled "A method for processing upstream signals in a passive optical network" filed on May 31, 2024, the disclosure of which is incorporated herein by reference in its entirety. TECHNICAL FIELD

[0003] The present disclosure relates to the field of communications, and in particular, to a method for processing upstream signals in a passive optical network. BACKGROUND

[0004] A passive optical network (PON) is a point-to-multipoint network topology, in which one optical line terminal (OLT) connects multiple optical network units (ONUs) through an optical distribution network (ODN).

[0005] In a PON system, an ONU needs to complete an activation process to enjoy service, and in the activation process, the ONU needs to send a serial number (SN) to the OLT so that the OLT can distinguish different ONUs. The activation process is completed through physical layer operation management and maintenance (PLOAM) messages and bandwidth allocation messages between the ONU and the OLT. From the start of the activation process, all management information (such as PLOAM messages and bandwidth allocation information) exchanged between the ONU and the OLT is sent in plaintext, and the resource allocation obtained by the ONU, such as an optical network unit identifier (ONU-ID), an allocation identifier (Alloc-ID), and a GEM port identifier (GEM Port-ID), is also in plaintext or is a default configuration.

[0006] Therefore, a malicious ONU can easily listen to the management information between the ONU and the OLT or the use of ONU resources, thereby obtaining the online and offline status of a normal ONU. When the normal ONU is offline or not online, the malicious ONU can initiate and complete the activation process by imitating the SN of the normal ONU, thereby stealing the service of the normal ONU.

[0007] In conclusion, there is still no good solution to the above problems. Summary of the Invention

[0008] This disclosure provides a method for processing uplink signals on a passive optical network, which at least solves the problem in related technologies where the SN of a normal ONU is counterfeited, resulting in the unauthorized use of services by the normal ONU.

[0009] According to one embodiment of this disclosure, a passive optical network uplink signal processing method is provided, applied to an optical line terminal (OLT). The method includes: acquiring a target uplink signal from a target ONU on an uplink bandwidth pre-allocated to the target ONU, obtaining feature data of the target uplink signal; determining the identity information of the target ONU based on the uplink bandwidth, and determining the processing method of the feature data.

[0010] According to another embodiment of this disclosure, a passive optical network uplink signal processing method is provided, applied to a computing center. The method includes: acquiring feature data of a target uplink signal, identity information of a target ONU, and a processing method for the feature data from an optical line terminal (OLT).

[0011] According to another embodiment of this disclosure, a passive optical network architecture is provided, the architecture comprising: an optical line terminal (OLT) configured to perform the steps in any of the above-described method embodiments applied to the OLT; and a computing center configured to perform the steps in any of the above-described method embodiments applied to the computing center.

[0012] According to yet another embodiment of this disclosure, a computer-readable storage medium is also provided, which stores a computer program, wherein the computer program is executed by a processor to perform the steps in any of the above method embodiments.

[0013] According to yet another embodiment of this disclosure, an electronic device is also provided, including a memory and a processor, the memory storing a computer program, the processor being configured to run the computer program to perform the steps in any of the above method embodiments.

[0014] According to yet another embodiment of this disclosure, a computer program product is also provided, including a computer program that, when executed by a processor, implements the steps in any of the above method embodiments. Attached Figure Description

[0015] Figure 1 is a flowchart of a passive optical network uplink signal processing method of an OLT according to an embodiment of the present disclosure;

[0016] Figure 2 is a flowchart of a passive optical network uplink signal processing method for a computing center according to an embodiment of the present disclosure;

[0017] FIG. 3 is a schematic diagram of a passive optical network architecture according to an embodiment of the present disclosure;

[0018] FIG. 4 is a schematic diagram of a functional distribution of a passive optical network architecture according to an embodiment of the present disclosure;

[0019] FIG. 5 is a schematic diagram of a functional distribution of a passive optical network architecture according to another embodiment of the present disclosure. DETAILED DESCRIPTION

[0020] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings and in conjunction with embodiments.

[0021] It should be noted that the terms "first", "second", and the like in the description of the present disclosure and the claims and the above-described accompanying drawings are used to distinguish similar objects, and do not necessarily have to describe a specific order or sequence.

[0022] The method in the embodiments of the present disclosure is applied to a passive optical network. The passive optical network generally includes an OLT and an ONU. One OLT can connect and control multiple ONUs.

[0023] In some embodiments, the passive optical network further has a computing center deployed inside or outside the OLT.

[0024] A method for processing an upstream signal of a passive optical network is provided in an embodiment of the present disclosure and is applied to an optical line terminal (OLT). FIG. 1 is a flowchart of a method for processing an upstream signal of a passive optical network of an OLT according to an embodiment of the present disclosure. As shown in FIG. 1, the flowchart includes the following steps:

[0025] In step S102, a target upstream signal from a target ONU is collected on an upstream bandwidth pre-assigned to the target ONU, to obtain characteristic data of the target upstream signal.

[0026] In step S104, identity information of the target ONU is determined according to the upstream bandwidth, and a processing mode of the characteristic data is determined.

[0027] In the present embodiment, the characteristic data is data obtained by the OLT collecting a signal. Exemplarily, the characteristic data can be a series of waveform amplitude values at corresponding time points obtained by sampling a time-domain waveform of a signal at a certain sampling rate. Alternatively, the series of data can also be converted to obtain other formats of characteristic data, such as image data of an eye diagram.

[0028] In this embodiment, the feature data can be understood as the "fingerprint information" of a certain ONU, different ONUs have different "fingerprint information", taking the waveform feature as an example, the present disclosure can learn the waveform feature of a certain ONU, identify the identity of the signal sender based on the waveform feature, and further determine the legitimacy of the ONU identity, that is, whether the ONU identity is legal or is being impersonated.

[0029] In this embodiment, the ONU identity information is the unique identity of the ONU, which can include but is not limited to the ONU serial number (SN), ONU identifier, etc. In the interaction process between the OLT and the ONU, the identity information is sent in plaintext, so there is a certain risk of being impersonated, while the "fingerprint information" is hidden in the feature data of the signal (such as waveform, amplitude value, etc.), which is difficult to imitate.

[0030] In the embodiment of the present disclosure, through the above steps S102 and S104, the identity of the ONU can be determined according to the feature data in the uplink signal, which can effectively and reliably prevent the ONU identity from being impersonated, and further solve the problem that the SN of the normal ONU is imitated in the related art, resulting in the problem that the normal ONU service is stolen and enjoyed.

[0031] In some embodiments, the processing mode includes ONU identification and model training. According to the processing mode, the OLT can use the collected feature data for ONU identification, or can train a model according to the feature data. In some other embodiments, the OLT can not further process the feature data, but send the processing mode and the feature data to the computing center for further processing.

[0032] In an exemplary embodiment, when the ONU is initially online, the OLT can collect uplink signal data and establish a correspondence between the uplink signal data and the identity information through training. In the subsequent ONU working process, including continuously staying online, power-on restart, reactivation, etc., the uplink signal of the ONU needs to be collected, and the matching relationship between the uplink signal and the identity information is identified to determine the legitimacy of the ONU, that is, whether it is legal or is being impersonated.

[0033] In some embodiments, the function distribution of the OLT and the computing center includes the following cases: A, the OLT only collects data, and the computing center performs model training and ONU identification; B, the OLT can collect data and perform ONU identification, and the computing center can perform model training; C, the OLT can collect data, train a model, and perform ONU identification. The method on the OLT side will be described in detail based on the above function distribution.

[0034] Embodiment A, the OLT only collects data, and the computing center performs model training and ONU identification.

[0035] In some embodiments, after step S104, the method further comprises: step S106A, in the case that the processing mode is ONU identification, sending the feature data, the identity information and the processing mode to the computing center, so that the computing center identifies the legitimacy of the identity information of the target ONU.

[0036] In some embodiments, step S106A can further comprise, in the case that the processing mode is ONU identification, sending the signal type of the target upstream signal to the computing center.

[0037] In some embodiments, after step S106A, the method further comprises: step S108A, receiving the identification result of the computing center, wherein the identification result is used to indicate the legitimacy of the identity information of the target ONU.

[0038] In some embodiments, the method further comprises: step S105A, in the case that the processing mode is model training, sending the feature data, the identity information and the processing mode to the computing center, so that the computing center trains a preset machine learning model according to the feature data to obtain a target ONU identification model.

[0039] In some embodiments, step S105A can further comprise: in the case that the processing mode is model training, sending the target signal type of the target upstream signal to the computing center, so that the computing center associates the target ONU identification model with the target signal type and the identity information.

[0040] In this embodiment, the training and application of the ONU identification model can be implemented in the computing center, and the OLT only needs to perform preliminary processing on the upstream signal, send the obtained feature data, identity information, signal type, processing mode and other data to the computing center, and perform the next step of processing by the computing center. When the processing mode is ONU identification, the computing center identifies whether the identity information of the target ONU is usurped, and the OLT only needs to send the data to be identified to the computing center and receive the identification result.

[0041] Embodiment B, the OLT can collect data and perform ONU identification, and the computing center performs model training.

[0042] In some embodiments, after step S104, the method further comprises: step S106B, in the case that the processing mode is ONU identification, inputting the identity information and the feature data into a pre-trained target ONU identification model to obtain an identification result, wherein the identification result is used to indicate the legitimacy of the identity information of the target ONU.

[0043] In some embodiments, step S106B can further include: determining the target ONU identification model from the pre-trained at least one ONU identification model according to a target signal type of the target upstream signal.

[0044] In the present embodiment, the pre-trained at least one ONU identification model is obtained from a computing center. Therefore, the method can further include: step S101B, obtaining the pre-trained at least one ONU identification model from the computing center, wherein each of the ONU identification models corresponds to one of the identity information and one signal type.

[0045] In some embodiments, the method further includes: step S105B, in the case that the processing mode is model training, sending the feature data, the identity information and the processing mode to the computing center, so that the computing center trains a preset machine learning model according to the feature data to obtain a target ONU identification model.

[0046] In some embodiments, step S105B can further include: in the case that the processing mode is model training, sending a target signal type of the target upstream signal to the computing center, so that the computing center associates the target ONU identification model with the target signal type and the identity information.

[0047] In the present embodiment, the training of the ONU identification model is implemented in the computing center, which sends the trained model to the OLT. The OLT needs to perform preliminary processing on the upstream signal and can also perform ONU identification on the feature data. The OLT only needs to send relevant training data to the computing center when the processing mode is model training. When the processing mode is ONU identification, the OLT can identify whether the identity of the ONU is usurped according to the to-be-identified data.

[0048] Embodiment C, the OLT can collect data, train a model and perform ONU identification.

[0049] In some embodiments, after step S104, the method further includes: step S106C, in the case that the processing mode is ONU identification, inputting the identity information and the feature data into a pre-trained target ONU identification model to obtain an identification result, wherein the identification result is used to indicate the legitimacy of the identity information of the target ONU.

[0050] In some embodiments, step S106C can further include: determining a target ONU identification model from the pre-trained at least one ONU identification model according to the identity information and a signal type of the target upstream signal.

[0051] In some embodiments, the pre-trained at least one ONU identification model is trained locally at the OLT.

[0052] In some embodiments, the method further comprises: in the case where the processing mode is model training, training a preset machine learning model according to the feature data to obtain a target ONU identification model, in step S105C.

[0053] In some embodiments, step S105C can further comprise: associating the target ONU identification model with the target signal type of the target uplink signal and the identity information.

[0054] In an exemplary embodiment, the preset machine learning model is a Convolutional Neural Network (CNN) model, wherein the CNN model comprises: an input layer, a convolutional layer, a pooling layer, a ReLU layer, and an output layer.

[0055] In some embodiments, the model training can be an initial training process of the model before application, or a training process of updating the model during application.

[0056] In some embodiments, after step S108A / step S106B / step S106C, the OLT has obtained the identification result, at which time the OLT can process according to the identification result. The identification result includes whether the ONU identity information is legal, whether it is impersonated, etc. The specific processing process includes, but is not limited to, at least one of the following: in the case where the identity information of the target ONU is illegal or impersonated, instructing the target ONU to go offline; in the case where the identity information of the target ONU is legal or not impersonated, instructing the target ONU to remain online.

[0057] In this embodiment, the OLT can kick off-line the ONU with impersonated identity information, and this method can effectively and reliably prevent the ONU identity from being impersonated, thereby solving the problem of normal ONU service being stolen and enjoyed due to the impersonation of the SN of the normal ONU in the related art.

[0058] In some embodiments, the signal type of the target uplink signal is a preset target signal type.

[0059] In some embodiments, before step S102, the method further comprises: step S100, sending a management control message for allocating the uplink bandwidth to the target ONU, wherein the target signal type is carried in the management control message.

[0060] In some embodiments, the target signal type comprises at least one of the following: an ONU Registration message, an ONU Serial Number message, a bit sequence of all 1s, a bit sequence of 0s and 1s alternating, a Pseudo-Random Binary Sequence (PRBS).

[0061] In the present embodiment, the OLT can allocate uplink bandwidths to a plurality of ONUs and specify the signal types that can be carried by the uplink bandwidths. The uplink bandwidths allocated to different ONUs are different, and thus, the OLT can determine the ONU to which an uplink signal belongs according to the uplink bandwidth of the received signal when collecting the uplink signal, i.e., the OLT can determine the identity information of the target ONU according to the uplink bandwidth in step S104.

[0062] In some embodiments, step S102 can comprise the following steps:

[0063] Step S1022, receiving the target uplink signal from the target ONU on the uplink bandwidth;

[0064] Step S1024, performing data sampling on the target uplink signal at a preset sampling rate to obtain sampling data;

[0065] Step S1026, generating the feature data according to the sampling data.

[0066] In some embodiments, after step S1022, the method further comprises step S1023, analyzing the target uplink signal to obtain the target signal type of the target uplink signal. For example, when the uplink signal is a normal signal, the uplink signal can carry signal type, ONU identity information, etc., and by analyzing the signal, the information can be directly obtained.

[0067] In some embodiments, the data sampling in step S1024 comprises, but is not limited to, actual sampling, down-sampling, interpolation sampling, etc., and can also be combined with multiple sampling methods. The specific value of the sampling rate can be set according to the requirements of the ONU identification model for input data, and the present disclosure does not limit this.

[0068] In some embodiments, step S1026 of generating the feature data according to the sampling data can comprise the following steps:

[0069] Step S1026-2, buffering the sampling data and performing signal analysis on the sampling data to determine whether the sampling data satisfies a preset condition;

[0070] Step S1026-4: In a case where the sampling data meets the preset condition, the sampling data or the preprocessed sampling data is taken as the feature data.

[0071] In this embodiment, the sampling data in step S1026-4 is usually a feature sequence, and the feature data can be a feature sequence or other format data obtained by preprocessing the feature sequence, such as image data. For example, the feature data can also include an eye diagram, and correspondingly, the preprocessing of the sampling data can be converting the feature sequence into an eye diagram. The format of the feature data can be set according to the requirement of the ONU identification model for the input data.

[0072] In an example embodiment, the training data / identification data is in a sequence format, and the training model is a CNN convolutional neural network including 23 layers of input layer, convolutional layer, pooling layer, ReLU layer and output layer. The OLT can obtain a sampling rate 8 times the upstream rate through actual sampling or down-sampling or interpolation sampling, and directly input the sampling data sequence into the CNN for training and judgment, and the test accuracy rate is 90.0%.

[0073] In an example embodiment, the training data / identification data is in an image format, and the training model is a CNN convolutional neural network including 12 layers of input layer, convolutional layer, pooling layer, ReLU layer and output layer. The OLT can obtain a sampling rate 2 times the upstream rate through actual sampling or down-sampling or interpolation sampling, and convert the sampling data into an eye diagram, and input the eye diagram into the CNN for training and judgment, and the test accuracy rate is 98.0%.

[0074] In this embodiment, the OLT continuously receives and collects upstream signals on the upstream bandwidth, and caches the sampling data of all signals to the local database. The OLT continuously analyzes the cached data, and when the preset condition is triggered, the corresponding sampling data is taken out from the local database for the next step processing. The preset condition can be a trigger condition of the model training, or a trigger condition of the ONU identification.

[0075] In some embodiments, the preset condition in step S1026-2 includes at least one of the following: preset collection time, preset collection period, preset signal type, preset abnormal position, service data producing error code, signal mutation.

[0076] In an example embodiment, the preset collection time can include: when the ONU is just online, during the ONU activation process, when the ONU has not been authenticated, when the ONU authentication fails, etc.

[0077] In some embodiments, the preset condition can include a trigger condition for ONU identification and a trigger condition for model training, both of which trigger the step of extracting the sampling data from the cache and taking the sampling data or the preprocessed sampling data as the feature data. For example, the trigger condition for model training can be set as successful ONU authentication.

[0078] In an exemplary embodiment, the OLT can capture signal mutations by signal analysis on the cached data, determine the presence of abnormal data, and take the corresponding feature data as the data to be identified. The types of signal mutations include, but are not limited to, power, amplitude, frequency, phase, signal-to-noise ratio (SNR), etc. The OLT can take a snapshot of the cached data when capturing the signal mutations.

[0079] In an exemplary embodiment, the OLT can also trigger ONU identification or training / update of the ONU identification model at a fixed time point or with a fixed period by setting a preset collection time or a preset collection period.

[0080] In an exemplary embodiment, the OLT can also trigger ONU identification according to the forward error correction (FEC) decoding of the upstream signal, such as when decoding fails or error codes are generated in the service data.

[0081] In some embodiments, the processing mode of the feature data in step S104 can include the following steps:

[0082] Step S1042, after the target ONU completes activation or goes online, performing identity authentication on the target ONU.

[0083] Step S1044, in the case of successful identity authentication of the target ONU, determining that the target ONU is working normally, and determining the processing mode as model training.

[0084] Step S1046, in the case that the target ONU does not perform identity authentication, identity authentication fails, or the sampling data of the target upstream signal meets a preset condition, determining the processing mode as ONU identification.

[0085] In some embodiments, the preset condition in step S1046 includes at least one of the following: a preset collection time, a preset collection period, a preset signal type, a preset abnormal position, error codes generated in service data, and signal mutations. For example, the preset collection time can include when the ONU is just online, during ONU activation, when the ONU does not perform identity authentication, when the ONU identity authentication fails, etc.

[0086] In the embodiment, the training data for model training requires feature data of normal working ONUs (i.e. the identity is not impersonated), therefore, when performing initial model training, the ONU identity information needs to be authenticated first, and the related data of the uplink signal sent by the ONU is used for model training when it is determined that the ONU identity is not impersonated.

[0087] In some embodiments, if the processing manner is ONU identification, but the identification result is that the identity information of the target ONU is not impersonated, the corresponding feature data can also be used to update the ONU identification model.

[0088] In the embodiments of the present disclosure, the identity of the ONU can be determined according to the feature data in the uplink signal, which can effectively and reliably prevent the ONU identity from being impersonated, and further solve the problem that the SN of the normal ONU is impersonated in the related art, resulting in the problem that the service of the normal ONU is stolen and enjoyed.

[0089] According to another embodiment of the present disclosure, a passive optical network uplink signal processing method is also provided, which is applied to a computing center. FIG. 2 is a flowchart of the passive optical network uplink signal processing method of the computing center according to the embodiments of the present disclosure, as shown in FIG. 2, the flowchart includes the following steps:

[0090] In step S202, feature data of a target uplink signal, identity information of a target ONU, and a processing manner of the feature data are obtained from an optical line terminal (OLT).

[0091] In the embodiment, the feature data is data obtained by the OLT collecting the signal, and exemplarily, the feature data can be a series of waveform amplitude values at corresponding time points obtained by sampling the time domain waveform of the signal according to a certain sampling rate. Alternatively, the series of data can also be converted to obtain other formats of feature data, such as image data of an eye diagram.

[0092] In the embodiment, the ONU identity information is a unique identity of the ONU, which can include but is not limited to an ONU serial number, an ONU identifier, etc.

[0093] In the embodiments of the present disclosure, the identity legitimacy of the ONU can be determined according to the feature data in the uplink signal, which can effectively and reliably prevent the ONU identity from being impersonated, and further solve the problem that the SN of the normal ONU is impersonated in the related art, resulting in the problem that the service of the normal ONU is stolen and enjoyed.

[0094] In some embodiments, the processing mode includes ONU identification and model training. According to the processing mode, the computing center can use the feature data to train a model. The computing center can also input the feature data into an ONU identification model to locally perform ONU identification, or can also send the trained ONU identification model to the OLT, which can input the feature data into the ONU identification model to perform ONU identification.

[0095] In some embodiments, the functional distribution of the OLT and the computing center includes the following cases: A, the OLT only collects data, and the computing center performs model training and ONU identification; B, the OLT can collect data and perform ONU identification, and the computing center performs model training. The method on the computing center side will be described in detail based on the above functional distribution.

[0096] Embodiment A, the OLT only collects data, and the computing center performs model training and ONU identification.

[0097] In some embodiments, the method further includes: step S204A, in the case of model training, training a preset machine learning model according to the feature data to obtain a target ONU identification model.

[0098] In an exemplary embodiment, the preset machine learning model is a convolutional neural network model, wherein the convolutional neural network model includes an input layer, a convolutional layer, a pooling layer, a ReLU layer, and an output layer.

[0099] In some embodiments, step S202 can also include obtaining a target signal type of the target upstream signal from the OLT.

[0100] In some embodiments, step S204A can also include associating the trained target ONU identification model with the target signal type and the identity information.

[0101] In some embodiments, the method further includes: step S206A, in the case of ONU identification, inputting the identity information and the feature data into the target ONU identification model to obtain an identification result, wherein the identification result is used to indicate the legitimacy of the identity information of the target ONU.

[0102] In some embodiments, step S206A can also include determining a target ONU identification model from at least one pre-trained ONU identification model according to a target signal type of the target upstream signal.

[0103] In some embodiments, after step S206A, the method further includes step S208A, sending the identification result to the OLT, wherein the OLT is used to process the identification result.

[0104] In this embodiment, if the identification result indicates that the target ONU's identity information is invalid or has been impersonated, the OLT will take certain restrictive measures on the ONU after receiving the identification result. For example, the target ONU can be taken offline to protect the ONU's identity information, effectively and reliably prevent the ONU's identity from being impersonated, and thus solve the problem in related technologies where the SN of a normal ONU is counterfeited, resulting in the normal ONU's services being stolen and enjoyed.

[0105] In this embodiment, the training and application of the ONU identification model can both be implemented in the computing center. The OLT only needs to perform preliminary processing on the uplink signal and send the obtained feature data, identity information, signal type, processing method, and other data to the computing center, which then performs corresponding processing according to the processing method. When the processing method is ONU identification, the computing center identifies whether the identity information of the target ONU has been misused and sends the identification result to the OLT.

[0106] In Example B, the OLT can collect data and perform ONU identification, and the computing center can train the model.

[0107] In some embodiments, the method further includes: step S204B, in the case where the processing mode is model training, training a preset machine learning model based on the feature data to obtain a target ONU recognition model.

[0108] In some embodiments, step S202 may further include: obtaining the target signal type of the target uplink signal from the OLT.

[0109] In some embodiments, step S204B may further include: associating the trained target ONU recognition model with the target signal type and the identity information.

[0110] In an exemplary embodiment, the preset machine learning model is a convolutional neural network model, wherein the convolutional neural network model includes: an input layer, a convolutional layer, a pooling layer, a ReLU layer, and an output layer.

[0111] In some embodiments, the method further includes: step S206B, sending at least one trained ONU identification model to the OLT, wherein each ONU identification model corresponds to an identity information and a signal type.

[0112] In the embodiment, the training of the ONU identification model is implemented in the computing center, and the computing center only needs to send the trained model to the OLT, and the OLT uses the trained model to identify the to-be-identified data by itself.

[0113] In some embodiments, the target signal type includes at least one of the following: an ONU registration message, an ONU serial number message, a bit sequence of all 1s, a bit sequence of 0 and 1 alternately, and a pseudo-random binary sequence PRBS. In the embodiments of the present disclosure, the model training and ONU identification are both based on the signal type and ONU identity information. Different ONUs or different signal types can correspond to different ONU identification models.

[0114] According to another embodiment of the present disclosure, a passive optical network architecture is also provided, which includes: an optical line terminal OLT configured to perform the steps in any of the method embodiments applied to the OLT; and a computing center configured to perform the steps in any of the method embodiments applied to the computing center.

[0115] FIG. 3 is a schematic diagram of a passive optical network architecture according to an embodiment of the present disclosure. As shown in FIG. 3, the passive optical network architecture can include: an optical line terminal OLT 10, and a computing center 20.

[0116] In the embodiment, the passive optical network architecture can further include one or more optical network units ONUs. The OLT is connected to the one or more optical network units ONUs through an optical distribution network ODN.

[0117] In the embodiment, a malicious ONU can steal the identity information of a normal ONU after the normal ONU is offline and initiate an activation process to the OLT. Through the embodiments of the present disclosure, the OLT or the computing center can record the identity information of the normal ONU and learn the feature data of the uplink signal of the normal ONU through the ONU identification model, which can effectively and reliably prevent the ONU identity from being impersonated, thereby solving the problem that the SN of the normal ONU is impersonated in the related art, and the service of the normal ONU is stolen and enjoyed.

[0118] In the embodiment, the passive optical network architecture mainly includes the following three functional modules, including: data collection, data training (machine learning), and data decision (ONU identification).

[0119] In the embodiment, the data collection includes: collecting training data and collecting decision data. The training data is generally collected for a specific signal. The OLT collects the signals normally sent by the ONUs, or instructs the ONUs to send specific signals and then the OLT collects them, which are used for data training and forming a decision model. The decision data is generally collected for a target signal and input into a decision model for data decision.

[0120] In the embodiment, the data obtained by collecting the signal is generally a series of waveform amplitude values at corresponding time points obtained by sampling the time-domain waveform of the signal at a certain sampling rate. Of course, based on the series of data, other format data can also be converted, such as eye diagram image data.

[0121] In the passive optical network architecture, the distribution of the three functional modules of data collection, data training and data decision can include the following three implementation manners:

[0122] Manner 1: The data collection module, the model training module and the data decision module are all arranged in the OLT, and the data training and the data decision can be implemented in the computing center local to the OLT. Exemplarily, the computing center can be a digital signal processor (DSP) built in the OLT or an OLT software system.

[0123] Manner 2: The data collection module is in the OLT, and the model training module and the data decision module are in the computing center outside the OLT to perform data training and data decision. Exemplarily, the computing center can be a network management system, a computing single board in the same machine frame, etc., and the OLT needs to send the training data and the data to be decided and related information to the computing center.

[0124] Manner 3: The data collection module and the data decision module are in the OLT, and the model training module is in the computing center outside the OLT. The OLT sends the training data to the computing center, and the computing center configures the decision model (i.e., the ONU identification model) and related parameters back to the OLT after the training is completed.

[0125] In the embodiment of the present disclosure, when an ONU is initially online, the OLT collects uplink signal data, and establishes a corresponding relationship between the uplink signal data and the identity information through training in the OLT or the computing center. When the ONU works subsequently, including continuously staying online, power-on restart, reactivation, etc., the OLT also needs to collect the uplink signal of the ONU and identify the matching relationship between the uplink signal and the identity information in the OLT or the computing center to judge the legitimacy of the ONU. When the ONU identity is not legitimate or the ONU identity is usurped, certain restriction measures can be taken on the ONU to protect the identity information of the ONU, effectively and reliably prevent the ONU identity from being usurped, and thus solve the problem that the SN of the normal ONU is imitated in the related art, causing the business of the normal ONU to be stolen and enjoyed.

[0126] FIG. 4 is a schematic diagram of the functional distribution of the passive optical network architecture according to an embodiment of the present disclosure. As shown in FIG. 4, taking the implementation manner 2 as an example, the data collection functional module is built in the OLT, and the model training and data decision functional modules are arranged in the computing center outside the OLT.

[0127] In the embodiment, the OLT built-in DSP module samples the upstream signal from the ONU and sends it to the computing center for training and decision-making. In order to capture data, the OLT DSP also buffers and analyzes the upstream signal data, and if an anomaly is found, the data is snapped and sent to the computing center. The OLT DSP can also snap the buffered data when a preset condition is triggered, which can be set according to requirements, such as periodic triggering or triggering when the service data error code is triggered. The OLT sends the snapshot data to the computing center as training data to train the machine learning model, or as decision-making data to input the decision-making model for decision-making. The OLT can send the snapshot data directly to the computing center, or can preprocess the snapshot data before sending it to the computing center.

[0128] In the embodiment, the computing center contains a model training module and a data decision-making module (trained decision-making model), and the computing center can directly train the OLT DSP output signal data, or can preprocess the output signal data before training.

[0129] In the embodiment, when the OLT sends the snapshot data to the computing center, it also sends the data related information such as processing method, signal type, ONU identity information, snapshot time, etc. to the computing center. The processing method can indicate whether the data is training data or decision-making data, or can indicate the category of the decision-making model, such as ONU identification model or other processing model.

[0130] In the embodiment, the data acquisition function can acquire the characteristic data of the target upstream signal by sampling the optical signal, or by converting the optical signal to an electrical signal and then sampling the point signal. The characteristic data can be a sequence of signal waveform amplitude values corresponding to multiple sampling points. The sampling time interval of the sampling points is determined by the sampling rate, which can be set to 2 times or higher to better acquire signal characteristic data, but the present disclosure is not limited thereto, and the sampling rate can also be lower than 2 times.

[0131] In the embodiment, the required buffer size of the sampling data can be calculated according to the required time for service processing or signal analysis. For example, if FEC decoding produces an error code, the size of one FEC code block in 50G-PON is 17280 bits, then the buffer size should be set to at least 17280* sampling rate* sampling bit number. The sampling bit number is the number of bits used to represent one sampling point. The larger the bit number, the more accurate the value represented by the sampling point.

[0132] In the embodiment, the OLT performs continuous signal analysis on the buffered sampling data and captures signal mutations. Exemplarily, the signal mutations can include mutations of power, amplitude, frequency, phase, SNR, etc. When a signal mutation is captured or a preset condition is triggered, a snapshot of the buffered data is taken and the sampling data in the buffer is outputted.

[0133] In the embodiment, when the snapshot is triggered, the related information of the sampling data is outputted in addition to the previously buffered sampling data. Exemplarily, the related information of the sampling data can include processing mode (model training, ONU identification or other data processing), signal type (data code type, such as normal message, all 1, 01 alternation, PRBS, etc.), abnormal position (error bit position, signal mutation position), snapshot time, etc.

[0134] In the embodiment, the snapshot data can be triggered and captured by external demand, such as periodically, FEC decoding error code, or triggered and captured when a signal mutation is captured by signal analysis.

[0135] In the embodiment, the data length of the data to be decided and the training data is usually set as a fixed data length or time length. Exemplarily, it can be set as 200 ns sampling data, but the disclosure is not limited thereto. The training data can be periodically collected and the decision model is updated.

[0136] In some embodiments, the length of the training data can also be set according to the length of the acquired data to be decided, and the decision model is retrained.

[0137] In the embodiment of the disclosure, the main risk of ONU impersonation is that after the normal ONU is offline, the impersonated ONU detects the offline of the normal ONU and initiates the impersonation process. Therefore, the uplink signal of the ONU can be captured and decided during the ONU activation process. The OLT can capture the uplink burst corresponding to the Serial_Number ONU message or the Registration message normally sent by the ONU, or capture the signal of a specific code type sent by the ONU specified by the OLT during the activation process or after the activation is completed, and sample the data to be decided.

[0138] In the embodiment, the data decision module uses the trained decision model (machine learning model) to make inference decision for the data to be decided. Exemplarily, the decision model is an ONU identification model, which can give a decision result of whether the data to be decided is from the ONU corresponding to the specified identity information (such as SN).

[0139] In the embodiment, the data decision module classifies the data to be decided according to relevant data before inputting the data to be decided into the decision model, for example, classifies according to processing mode, ONU identity information, signal type, etc., and determines the corresponding decision model (such as the target ONU identification model). Similarly, during model training, the training data is also classified according to relevant data, and the corresponding decision model is trained according to the training data of different categories.

[0140] According to the embodiments of the present disclosure, the identity of the ONU can be decided according to the characteristic data in the uplink signal, which can effectively and reliably prevent the identity of the ONU from being impersonated, thereby solving the problem that the SN of the normal ONU is impersonated in the related art, and the service of the normal ONU is stolen and enjoyed.

[0141] In an embodiment of the present disclosure, in order to obtain the characteristic data of a single ONU as training data, the characteristic data of the ONU can be obtained by the following manner: data collection is performed on the uplink service signal normally sent by the ONU; data collection is performed on the idle service signal sent by the ONU.

[0142] In some embodiments, the OLT can configure the signal type sent by the ONU, instructing the ONU to send data of a special code type. The signal type can include but is not limited to: normal service data, a bit sequence of all 1s, a bit sequence of 0 and 1 alternately, a pseudo-random binary sequence (PRBS). The normal service data can include the uplink burst corresponding to the ONU registration (Registration) message and the ONU serial number (Serial_Number ONU) message.

[0143] In some embodiments, the OLT can send a management control message to the ONU, allocate uplink bandwidth to the ONU, and / or allocate a specified signal type to the ONU.

[0144] In an exemplary embodiment, the management control message can be a burst profile (Burst_Profile) message in a physical layer operation, administration, and maintenance (PLOAM) message, which can indicate the signal type through the 5th byte and the 6th byte.

[0145] Table 1 shows the composition structure of the burst profile message (Burst_Profile message) in the PLOAM message.

[0146] Table 1:

[0147] As shown in Table 1, in the burst configuration message, the 8 bits of the 5th byte can be expressed as "VVVR BBPP", wherein the 2 bits of PP are usually used to indicate the index of the burst configuration, and the signal type can be indicated through the 2 bits in the 5th byte. The 8 bits of the 6th byte can be expressed as "NNMM RRCF", wherein the 2 bits of RR are reserved bits, and the signal type can be indicated by setting the values of the 2 bits in the 6th byte.

[0148] In an example embodiment, the correspondence between the value of RR and the signal type is as follows: 00 indicates all normal data, 01 indicates all 1 bits, 10 indicates 01 alternating bits, and 11 indicates a PRBS code type, but the present disclosure is not limited thereto.

[0149] In an example embodiment, the OLT can configure the signal type sent by the ONU to the ONU in advance through the Burst_Profile PLOAM message in Table 1. When the OLT allocates bandwidth to the ONU, the burst profile in the Bandwidth Map (BWmap) entry can be specified as PP in the 5th byte in the Burst_Profile PLOAM message, and then the ONU will send the data code type (signal type) corresponding to the value of RR specified by PP in the uplink bandwidth specified by the BWmap entry.

[0150] In the present embodiment, the OLT allocates uplink bandwidth to the ONU through the BWmap message, and the BWmap message can include multiple 8-byte allocation structures, each of which contains a 2-bit burst profile field.

[0151] In the present embodiment, the OLT can configure the ONU-related resources, allocate uplink bandwidth, and allocate signal types. When the OLT receives the uplink signal of the target ONU, it can preliminarily filter out the uplink signal from the target ONU according to the relevant resources associated with the target ONU such as the uplink bandwidth and the signal type, determine the machine learning model for identifying the identity of the target ONU, and then make a decision on whether the ONU is being impersonated, thereby solving the problem of the SN of the normal ONU being impersonated in the related art, resulting in the problem of the normal ONU's service being stolen and enjoyed.

[0152] FIG. 5 is a schematic diagram of the functional distribution of the passive optical network architecture according to another embodiment of the present disclosure. As shown in FIG. 5, in the mode 3, the data collection module and the data decision module are in the OLT, the model training module is in the computing center outside the OLT, the OLT sends the training data to the computing center, and the computing center configures the decision model (i.e., the ONU identification model) and its related parameters back to the OLT after the training is completed.

[0153] In the embodiment, compared with the mode 2, only the model training module is integrated in the computing center outside the OLT in the mode 3, and the data decision module and the data collection module are integrated in the OLT, and the computing center further needs to configure the parameters of the decision model obtained by the model training back to the OLT.

[0154] In another embodiment, the function distribution of the passive optical network architecture can also be implemented in the mode 1. Compared with the mode 2, the model training module and the data decision module are integrated in the OLT in the mode 1, and the functions and interfaces of the modules are similar, which will not be described herein.

[0155] According to the embodiments of the present disclosure, the optical line terminal OLT collects a target upstream signal from a target ONU on the upstream bandwidth pre-allocated to the target ONU, obtains feature data of the target upstream signal, determines identity information of the target ONU according to the upstream bandwidth, and determines a processing mode of the feature data, so that the identity of the ONU can be decided according to the feature data in the upstream signal, the identity of the ONU can be effectively and reliably prevented from being impersonated, and the problem that the SN of the normal ONU is impersonated in the related art, causing the service of the normal ONU to be stolen and enjoyed is solved.

[0156] The embodiments of the present disclosure further provide a computer readable storage medium, which stores a computer program, and the computer program is run by a processor to perform the steps in any of the method embodiments.

[0157] In an example embodiment, the computer readable storage medium can include, but is not limited to, a U disk, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and various media that can store computer programs.

[0158] The embodiments of the present disclosure further provide an electronic device, which includes a memory storing a computer program and a processor configured to run the computer program to perform the steps in any of the method embodiments.

[0159] In an example embodiment, the electronic device can further include a transmission device connected to the processor and an input / output device connected to the processor.

[0160] The embodiments of the present disclosure further provide a computer program product, which includes a computer program, and the computer program is run by a processor to perform the steps in any of the method embodiments.

[0161] The specific examples in the present embodiment can refer to the examples described in the above embodiments and exemplary embodiments, which will not be repeated here.

[0162] Obviously, those skilled in the art should understand that the modules or steps of the present disclosure described above can be realized by general computing devices, which can be centralized on a single computing device or distributed on a network composed of multiple computing devices, which can be realized by program codes executable by the computing devices, so that they can be stored in storage devices and executed by the computing devices, and in some cases, the steps shown or described can be executed in different order, or they can be respectively manufactured into individual integrated circuit modules, or multiple modules or steps among them can be manufactured into a single integrated circuit module to realize. Thus, the present disclosure is not limited to any specific combination of hardware and software.

[0163] The above only describes the preferred embodiments of the present disclosure and is not intended to limit the present disclosure. For those skilled in the art, the present disclosure can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. within the principles of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for processing an upstream signal in a passive optical network, applied to an optical line terminal (OLT), the method comprising: collecting a target upstream signal from a target optical network unit (ONU) on an upstream bandwidth pre-assigned to the target ONU, to obtain characteristic data of the target upstream signal; determining identity information of the target ONU according to the upstream bandwidth, and determining a processing mode of the characteristic data. The method further comprises: in the case that the processing mode is ONU identification, sending the characteristic data, the identity information and the processing mode to a computing center, so that the computing center identifies the legitimacy of the identity information of the target ONU. The method further comprises: in the case that the processing mode is ONU identification, sending a signal type of the target upstream signal to the computing center.

2. The method of claim 1, wherein, The method further comprises: receiving an identification result from the computing center, wherein the identification result is used to indicate the legitimacy of the identity information of the target ONU. The method further comprises: in the case that the processing mode is ONU identification, inputting the identity information and the characteristic data into a pre-trained target ONU identification model to obtain an identification result, wherein the identification result is used to indicate the legitimacy of the identity information of the target ONU.

3. The method of claim 2, wherein, The method further comprises: determining the target ONU identification model from at least one pre-trained ONU identification model according to the signal type of the target upstream signal.

4. The method of claim 2, wherein, The pre-trained target ONU identification model is obtained from a computing center. The pre-trained target ONU identification model is trained locally in the OLT.

5. The method of claim 1, wherein, The method further comprises: processing according to the identification result, including at least one of the following: in the case that the identity information of the target ONU is not legitimate or is impersonated, instructing the target ONU to go offline; 6. The method of claim 5, wherein, in the case that the identity information of the target ONU is legitimate or is not impersonated, instructing the target ONU to remain online. The signal type of the target upstream signal is a preset target signal type.

7. The method of claim 5, wherein, The method further comprises: sending a management control message for assigning the upstream bandwidth to the target ONU, wherein the target signal type is carried in the management control message.

8. The method of claim 5, wherein, The target signal type includes at least one of the following: an ONU registration message, an ONU serial number message, a bit sequence of all 1s, a bit sequence of 0 and 1 alternately, a pseudo-random binary sequence (PRBS).

9. The method of claim 4 or 5, wherein, The method of collecting a target upstream signal from a target ONU on an upstream bandwidth pre-assigned to the target ONU to obtain characteristic data of the target upstream signal comprises: receiving the target upstream signal from the target ONU on the upstream bandwidth; sampling data of the target upstream signal at a preset sampling rate to obtain sampling data; generating the characteristic data according to the sampling data.

10. The method of claim 1, wherein, The method of generating the characteristic data according to the sampling data comprises:

11. The method of claim 10, wherein, buffering the sampling data and performing signal analysis on the sampling data to determine whether the sampling data meets a preset condition. ​ 12. The method of claim 11, wherein, ​ 13. The method of claim 1, wherein, ​ ​ ​ ​ 14. The method of claim 13, wherein, ​ ​ In a case where the sampling data meets the preset condition, the sampling data or the preprocessed sampling data is taken as the feature data.

15. The method of claim 14, wherein, The preset condition comprises at least one of a preset collection time, a preset collection period, a preset signal type, a preset abnormal position, a service data error code, and a signal mutation.

16. The method of claim 1, wherein, The processing mode of the feature data is determined, comprising: After the target ONU completes activation or goes online, identity authentication is performed on the target ONU. In a case where the target ONU identity authentication succeeds, it is determined that the target ONU works normally, and the processing mode is determined as model training. In a case where the target ONU does not perform identity authentication, identity authentication fails, or the sampling data of the target uplink signal meets a preset condition, the processing mode is determined as ONU identification.

17. The method of claim 1, wherein, The method further comprises: In a case where the processing mode is model training, the feature data, the identity information, and the processing mode are sent to a computing center, so that the computing center trains a preset machine learning model according to the feature data to obtain a target ONU identification model.

18. The method of claim 1, wherein, The method further comprises: In a case where the processing mode is model training, a target signal type of the target uplink signal is sent to the computing center, so that the computing center associates the target ONU identification model with the target signal type and the identity information.

19. A passive optical network uplink signal processing method applied to a computing center, the method comprising: obtaining, from an optical line terminal (OLT), feature data of a target uplink signal, identity information of a target ONU, and a processing mode of the feature data.

20. The method of claim 19, wherein, The method further comprises: In a case where the processing mode is model training, a preset machine learning model is trained according to the feature data to obtain a target ONU identification model.

21. The method of claim 20, wherein, The preset machine learning model is a convolutional neural network model, wherein the convolutional neural network model comprises an input layer, a convolutional layer, a pooling layer, a ReLU layer, and an output layer.

22. The method of claim 19, wherein, The method further comprises: obtaining, from the OLT, a target signal type of the target uplink signal.

23. The method of claim 22, wherein, The method further comprises: associating the trained target ONU identification model with the target signal type and the identity information.

24. The method of claim 22, wherein, The target signal type comprises at least one of an ONU registration message, an ONU serial number message, a bit sequence of all 1s, a bit sequence of 0 and 1 alternately, and a pseudo-random binary sequence (PRBS).

25. The method of claim 19, wherein, The method further comprises: In a case where the processing mode is ONU identification, the identity information and the feature data are input into a pre-trained target ONU identification model to obtain an identification result, wherein the identification result is used to indicate the legitimacy of the identity information of the target ONU.

26. The method of claim 25, wherein, The method further comprises: determining the target ONU identification model from at least one pre-trained ONU identification model according to a target signal type of the target uplink signal.

27. The method of claim 25, wherein, The method further comprises: sending the identification result to the OLT, wherein the OLT is used to perform processing according to the identification result.

28. The method of claim 19, wherein, The method further comprises: sending the trained at least one ONU identification model to the OLT, wherein each of the ONU identification models corresponds to one of the identity information and one of the signal types.

29. A passive optical network architecture, the architecture comprising: an optical line termination (OLT) configured to perform the method of any of claims 1 to 18; a computing center configured to perform the method of any of claims 19 to 28.

30. A computer readable storage medium having stored therein a computer program, wherein, The computer program, which performs the method of any of claims 1 to 28 when executed by a processor.

31. An electronic device comprising a memory and a processor, the memory having stored therein a computer program, the processor being configured to execute the computer program to perform the method of any of claims 1 to 28.

32. A computer program product comprising a computer program which, when executed by a processor, carries out the steps of the method of any of claims 1 to 28.

Images