Routing congestion rate prediction-based unfixed-length burst package allocation method and related equipment

Abstract

The invention belongs to the field of communication networks, and discloses a variable-length burst package allocation method based on route congestion rate prediction and related equipment, the method adopts a pre-trained congestion rate prediction network model to accurately predict the congestion rate of each optical network core node at the next moment, and further combines a plurality of pre-calculated shortest paths to accurately predict the congestion rate of each optical network core node at the next moment. The optimal path is determined by comprehensively considering the congestion state, and then the uncertain-length burst packet is dynamically generated based on the average congestion rate of the optimal path so as to adapt to the actual congestion condition of the network. By adopting the method, the problem of congestion aggravation caused by only depending on the shortest path in the traditional method is effectively avoided, and meanwhile, the problem that the resource utilization rate is low or the conflict probability is increased possibly due to the fact that the size of a fixed-length burst packet cannot be adjusted according to real-time congestion is also solved, so that the service quality and the overall performance of the network are remarkably improved; the multi-objective optimization of improving the network capacity, reducing the network delay, improving the resource utilization rate and reducing the burst packet conflict probability is realized.

Description

Technical Field

[0001] The invention belongs to the technical field of communication networks, and in particular relates to a method for allocating indefinite-length burst packets based on routing congestion rate prediction and related equipment. Background Art

[0002] With the rapid development of Internet technology, network data traffic has shown explosive growth. As an efficient optical communication technology, optical burst switching (OBS) network, referred to as optical network, has broad application prospects in the next generation of high-speed communication networks because it can effectively utilize the high bandwidth characteristics of optical fiber. However, in OBS networks, due to the burstiness and uncertainty of business traffic, network congestion has become one of the key factors limiting its performance improvement. When the business traffic in the network exceeds the carrying capacity of the link, it will cause problems such as packet loss and increased transmission delay, which seriously affects the service quality of the network.

[0003] At present, there are some solutions to the congestion problem in OBS networks. For example, traditional routing algorithms are usually based on the shortest path priority principle, such as the Dijkstra algorithm, which only considers the length of the path and ignores the real-time congestion status of the network. When the network is congested, still choosing the shortest path may lead to increased congestion and fail to achieve effective utilization of network resources. In addition, some methods use fixed-length burst packets to transmit data, but fixed-length burst packets cannot be adjusted according to the actual congestion of the network, which may increase the probability of burst packet collisions in congested links or low resource utilization in idle links.

[0004] It can be seen that the traditional routing transmission method lacks consideration of real-time congestion status, which increases the probability of burst packet collisions and reduces the service quality and overall performance of the network. Summary of the invention

[0005] The present invention provides a method for assembling indefinite-length burst packets based on routing congestion rate prediction and related equipment to solve the technical problem that the traditional routing transmission method lacks consideration of real-time congestion status, resulting in an increased probability of burst packet collisions and reduced network service quality and overall performance.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: In a first aspect, the present invention provides a method for allocating variable-length burst packets based on routing congestion rate prediction, comprising: The congestion rate prediction network model based on pre-training is used to predict the congestion rate of each optical network core node at the next moment; Determine the optimal path based on multiple pre-calculated shortest paths and the congestion rate of each optical network core node at the next moment; A burst packet of variable length corresponding to the new service is generated based on the average congestion rate of the optimal path; the burst packet of variable length is used to enter a sending queue and wait for the system to send it.

[0007] A further improvement of the present invention is that before using the pre-trained congestion rate prediction network model to predict the congestion rate of each optical network core node at the next moment, the congestion rate prediction network model is trained, and the specific steps are as follows: Construct a BP neural network model; wherein the input layer of the BP neural network model is used to receive historical congestion rates, current service traffic, congestion status of adjacent nodes, time information and service type information, the hidden layer is used to process and extract features, and the output layer is used to output predicted congestion rates; The BP neural network model is trained using real node congestion data to obtain a congestion rate prediction network model.

[0008] The present invention is further improved in that the use of a pre-trained congestion rate prediction network model to predict the congestion rate of each optical network core node at the next moment includes: Obtain current congestion data of each optical network core node; The acquired current congestion data are input into the congestion rate prediction network model, and the congestion rate of each optical network core node at the next moment is output.

[0009] The present invention is further improved in that, before determining the optimal path based on the multiple shortest paths obtained by pre-calculation and the congestion rate of each optical network core node at the next moment, multiple shortest paths are obtained by calculation, and the specific steps are as follows: Obtain network topology and existing service information of optical network; Based on the network topology and existing service information of the optical network, the communication cost between adjacent nodes is determined and a communication cost matrix is ​​generated; When new services arrive at the edge node, the k-shortest paths algorithm is used to calculate based on the communication cost matrix to obtain multiple shortest paths corresponding to the new services.

[0010] The present invention is further improved in that the method of determining the optimal path based on the pre-calculated multiple shortest paths and the congestion rate of each optical network core node at the next moment includes: The average congestion rate of each shortest path is calculated using the congestion rate of each optical network core node at the next moment; Sort the average congestion rate of each shortest path from low to high, and select the shortest path with the lowest average congestion rate as the optimal path.

[0011] A further improvement of the present invention is that, if the average congestion rates of multiple shortest paths are the same, the shortest path with the smallest number of hops is selected as the optimal path.

[0012] The present invention is further improved in that the method of generating a burst packet of indefinite length corresponding to a new service based on the average congestion rate of the optimal path comprises: Cache new services received by edge nodes to obtain service cache data; Calculate the burst packet length based on the average congestion rate corresponding to the optimal path; Based on the calculated burst data packet length, service data of corresponding length is read from the service buffer data to generate a burst data packet; A burst control packet is generated based on the burst data packet length; wherein the burst data packet and the burst control packet constitute the variable-length burst packet; after being generated, the burst data packet and the burst control packet enter a sending queue and wait for the system to send them.

[0013] In a second aspect, the present invention provides a variable-length burst packet allocation system based on routing congestion rate prediction, comprising: A congestion rate prediction module, used to predict the congestion rate of each optical network core node at the next moment using a pre-trained congestion rate prediction network model; An optimal path determination module, used to determine the optimal path based on multiple shortest paths obtained in advance and the congestion rate of each optical network core node at the next moment; The variable length burst packet generation module is used to generate variable length burst packets corresponding to new services based on the average congestion rate of the optimal path; the variable length burst packets are used to enter the sending queue and wait for the system to send them.

[0014] In a third aspect, the present invention provides a device, comprising: Memory for storing computer programs; The processor is used to implement the steps of the above-mentioned variable-length burst packet allocation method based on routing congestion rate prediction when executing the computer program.

[0015] In a fourth aspect, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, is used to implement the steps of the above-mentioned method for assembling variable-length burst packets based on routing congestion rate prediction.

[0016] Compared with the prior art, the present invention has the following beneficial effects: The present invention provides a method for assembling variable-length burst packets based on route congestion rate prediction. The method uses a pre-trained congestion rate prediction network model to accurately predict the congestion rate of each optical network core node at the next moment, and then combines multiple pre-calculated shortest paths to comprehensively consider the congestion state to determine the optimal path, and then dynamically generates variable-length burst packets based on the average congestion rate of the optimal path to adapt to the actual congestion situation of the network. The method effectively avoids the problem of aggravated congestion caused by relying only on the shortest path in traditional methods, and also solves the problem that fixed-length burst packets cannot be resized according to real-time congestion, which may lead to low resource utilization or increased conflict probability, thereby significantly improving the service quality and overall performance of the network; and realizing multi-objective optimization of increasing network capacity, reducing network delay, improving resource utilization and reducing the probability of burst packet conflicts.

[0017] Preferably, in the present invention, the congestion rate prediction network model is trained to ensure that the model can accurately predict the future congestion rate, which improves the accuracy of the prediction and further enhances the reliability of the optimal path selection and burst packet allocation based on the prediction results.

[0018] Preferably, in the present invention, by acquiring current congestion data and inputting it into the prediction model, the network congestion status is updated in real time, providing timely and accurate information support for subsequent path selection and burst packet allocation, which helps to dynamically adapt to network changes and optimize data transmission.

[0019] Preferably, in the present invention, multiple shortest paths are pre-calculated, providing abundant options for subsequent optimal path selection, which helps to quickly and accurately find the optimal transmission path when facing different network conditions and business requirements.

[0020] Preferably, in the present invention, the optimal path is determined based on the predicted future congestion rate to ensure smooth data transmission path and low latency, which helps to improve network transmission efficiency and user experience.

[0021] Preferably, in the present invention, when the average congestion rate is the same, the path with the smallest number of hops is selected as the optimal path, which helps to further reduce the delay and cost of data transmission, thereby optimizing the utilization of network resources and improving the overall network performance.

[0022] Preferably, in the present invention, variable-length burst packets are generated based on the average congestion rate of the optimal path, ensuring that the burst packets can adapt to the network conditions and reduce congestion and delay during transmission. At the same time, the generation of burst control packets helps the system to more effectively manage and control the transmission of burst packets, thereby improving the reliability and efficiency of data transmission. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1A flowchart of node congestion rate prediction provided by an embodiment of the present invention; Figure 2 An optimal path selection flow chart provided for an embodiment of the present invention; Figure 3 A flow chart of generating a burst packet of indefinite length provided by an embodiment of the present invention; Figure 4 A flow chart of a method for allocating variable-length burst packets based on routing congestion rate prediction provided by an embodiment of the present invention; Figure 5 A schematic diagram of the structure of a variable-length burst packet distribution system based on routing congestion rate prediction provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0024] In order to further understand the content of the present invention, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the embodiments are only for explaining the present invention and are not intended to limit it.

[0025] The technical terms involved in the present invention are explained as follows: BP: Back-Propagation, or back-propagation algorithm. In neural network training, the BP algorithm is used to calculate the gradient of the loss function with respect to the network weights, and back-propagate the gradient to update the weights, so that the output of the neural network is closer to the expected value.

[0026] BCP: Burst Control Packet. In an optical burst switching network, BCP is a burst control packet that contains at least basic information such as the source address, destination address, service priority, transmission path, offset time, and burst packet length. It is used to perform operations such as resource reservation and path control before the data burst packet (BDP) is transmitted.

[0027] BDP: Burst Data Packet. In an optical burst switching system, it carries service data and its length is dynamically adjusted according to the congestion rate of the path in the network. It is an integer multiple of the basic length.

[0028] OBS network: Optical Burst Switching, optical burst switching network. Optical burst switching is an optical communication technology that can effectively utilize the high bandwidth characteristics of optical fiber and has broad application prospects in the next generation of high-speed communication networks.

[0029] QoS: Quality of Service. It is used to measure the quality of services provided by the network to users, including indicators such as network delay, packet loss rate, bandwidth, etc.

[0030] MSE: Mean Squared Error. It is a commonly used loss function. In neural network training, it is used to measure the difference between the predicted value and the true value, and optimize the performance of the neural network by minimizing MSE.

[0031] k-shortest paths: K shortest paths algorithm.

[0032] As described in the background technology, at present, the traditional routing transmission algorithm only considers the length of the path, but ignores the real-time congestion status of the network. When the network is congested, still choosing the shortest path may lead to aggravated congestion and fail to achieve effective utilization of network resources. In addition, some methods transmit data by setting burst packets of fixed length, but the burst packets of fixed length cannot be adjusted according to the actual congestion of the network, which may lead to an increase in the probability of burst packet collision in congested links, or low resource utilization in idle links.

[0033] Example 1 In order to solve the above problems, this embodiment provides a variable-length burst packet allocation method based on routing congestion rate prediction. This method accurately predicts the congestion rate of each node in the network, selects the optimal communication path, and dynamically adjusts the length of the burst packet according to the congestion situation of the path, thereby effectively improving network resource utilization, reducing the probability of burst packet conflict, and improving the service quality and overall performance of the network.

[0034] This embodiment provides a method for allocating variable-length burst packets based on routing congestion rate prediction, which is characterized by comprising: The congestion rate prediction network model based on pre-training is used to predict the congestion rate of each optical network core node at the next moment; Determine the optimal path based on multiple pre-calculated shortest paths and the congestion rate of each optical network core node at the next moment; A burst packet of variable length corresponding to the new service is generated based on the average congestion rate of the optimal path; the burst packet of variable length is used to enter a sending queue and wait for the system to send it.

[0035] In this embodiment, the operating environment is an optical burst switching network, which includes 10 core nodes and 5 edge nodes, and the network topology is a mesh structure. Multiple services are running in the network, including voice services, video services and data services, and different services have different service priorities and bandwidth requirements.

[0036] The following further explains the variable-length burst packet assembly method provided by this embodiment with reference to the accompanying drawings: This embodiment provides a method for allocating variable-length burst packets based on route congestion rate prediction, which is as follows: like Figure 1As shown, step 1, in the optical network, predicting the congestion rate of each core node at the next moment, includes: Neural network construction: A BP neural network with one hidden layer is constructed. The number of input layer nodes is 5, which receives the historical congestion rate of the core node, the current service flow, the congestion status of the adjacent nodes, time information and service type information. The number of hidden layer nodes is 10, and the Sigmoid activation function is used. The number of output layer nodes is 1, which outputs the predicted congestion rate.

[0037] It should be noted that the BP neural network can be configured with one or more hidden layers. Considering the real-time performance of the communication network, one hidden layer is used in this embodiment.

[0038] Neural network training: Collect the historical congestion data of the core nodes in the optical network in the past week as training data, a total of 1000 sets of data. The training data is divided into a training set (800 sets) and a test set (200 sets). Use the training set to train the BP neural network, set the learning rate to 0.01, and the number of training times to 1000. During the training process, calculate the mean square error (MSE). When the MSE is less than 0.01, the training is considered to have converged, and the congestion rate prediction network model is output.

[0039] Obtain node information: Real-time monitoring of the number of BCPs arriving at the core node per unit time as current congestion data. Record the acquired congestion data every 10 minutes for BP neural network training and real-time congestion status monitoring.

[0040] Congestion rate prediction: Use the trained congestion rate prediction network model to predict the congestion rate of the core node at the next moment. For example, at a certain moment, input the current node information, that is, the current congestion data, including the current service flow of the core node, the congestion status of the adjacent nodes, time information and service type information; the congestion rate prediction network model outputs the predicted congestion rate of 0.6 (indicating that the number of BCP arrivals per unit time is relatively large, and the network has a certain degree of congestion).

[0041] like Figure 2 As shown, step 2, determining the optimal path, includes: Get network topology information: Get the network topology of the optical network and the existing service information in the optical network, including service priority, service required bandwidth, and service path. For example, the service priority of voice service is high, and the bandwidth requirement is 64kbps; the service priority of video service is medium, and the bandwidth requirement is 1Mbps; the service priority of data service is low, and the bandwidth requirement is 512kbps.

[0042] Calculate the communication cost matrix W: According to the service priority and bandwidth requirements in the optical network, calculate the communication cost matrix W between adjacent nodes; the size of the communication cost matrix W is the same as the number of nodes in the network topology, which represents the cost required for communication between adjacent nodes. For services with high service priority, the communication cost is low; for services with large bandwidth requirements, the communication cost is appropriately increased. For example, the communication cost from node A to node B for voice services is 1, the communication cost for video services is 3, and the communication cost for data services is 5.

[0043] Calculate K shortest paths: When a new service arrives at an edge node, assuming it is a video service, the system calculates K = 3 shortest paths for the service according to the k-shortest paths algorithm and the communication cost matrix W. The three shortest paths obtained are path 1: edge node 1-core node 2-core node 5-destination node; path 2: edge node 1-core node 3-core node 6-destination node; path 3: edge node 1-core node 4-core node 7-destination node.

[0044] Calculate the average congestion rate N of the K shortest paths: After obtaining the K shortest paths, calculate the average congestion rate N of each shortest path. N is equal to the ratio of the sum of the predicted congestion rates of each node in the shortest path to the number of nodes, that is, N=(G1+G2+...+Gn) / n, where G1, G2,..., Gn are the predicted congestion rates of each node in the path, and n is the number of nodes in the path. For example: the predicted congestion rates of the nodes in path 1 are 0.3, 0.4, and 0.5 respectively, and the average congestion rate N1=(0.3+0.4+0.5) / 3=0.4; the predicted congestion rates of the nodes in path 2 are 0.2, 0.3, and 0.6 respectively, and the average congestion rate N2=(0.2+0.3+0.6) / 3=0.37; the predicted congestion rates of the nodes in path 3 are 0.5, 0.6, and 0.7 respectively, and the average congestion rate N3=(0.5+0.6 +0.7) / 3=0.6.

[0045] Sorting and selecting the optimal path M: Sort the K shortest paths from low to high according to the average congestion rate N, and select the path M with the lowest average congestion rate as the optimal path for this communication. If there are multiple shortest paths with the same congestion rate, select the shortest path with the smallest number of hops as the optimal path for this communication. Sort the above three paths from low to high according to the average congestion rate. It can be seen that the average congestion rate of path 2 is the lowest, so path 2 is selected as the optimal path for this communication.

[0046] like Figure 3 As shown, step 3, generating a burst packet of indefinite length, includes: Waiting for new services: The edge node is in a waiting state, waiting for new IP services to arrive.

[0047] Data caching: After receiving a service, the edge node needs to cache the service data to ensure that the generation of BDP (burst data packet) can obtain the corresponding data; in this embodiment, when the edge node receives a video service data, it caches it to obtain service cache data.

[0048] Obtaining the optimal path M: If the data cache sent to a certain node is not empty, further calculate the optimal path M of the service; according to the above optimal path selection method, determine that the optimal path of the new service is path 2.

[0049] Calculate the BDP packet length: Calculate the burst data packet length (BDP packet length) based on the average congestion rate of the optimal path M. The principle is that the higher the average congestion rate, the shorter the packet length. The shortest L min = l, the lower the congestion rate, the longer the packet length, the longest is L max =nxl, where l is the basic length set by the optical burst switching system and n is a positive integer. The BDP packet length is an integer multiple of the basic length l.

[0050] In this embodiment, it is known that the basic length l=1000 bytes set by the optical burst switching system, and the BDP packet length is calculated according to the average congestion rate of 0.37 of path 2. Since the congestion rate is low, a longer packet length is selected, and the calculated burst data packet length is 3000 bytes.

[0051] It should be noted that: in this embodiment, the linear proportional method can be used to solve the BDP packet length L; according to network characteristics and experience, L=l×(k1-k2×C), where k1 and k2 are network coefficients, and C represents the average congestion rate of the optimal path.

[0052] Of course, in this embodiment, the piecewise function method can also be used to solve the BDP packet length L; define the piecewise function: divide different intervals according to the range of congestion rate, corresponding to different packet length calculation methods. Assume that when 0≤C<0.4, L=lx(3+0.5x(0.4-C)); when 0.4≤C<0.7, L=l; when C>0.7, L=lx0.5; judge the interval and calculate: because the average congestion rate of path 2 C=0.37 is in the interval of 0≤C<0.4, substitute l=1000 bytes and C=0.37 into the corresponding formula, and get L≈3000 bytes, so the burst data packet length is taken as 3000 bytes.

[0053] In addition, the solution method for the BDP packet length L can also adopt the neural network prediction method, using the current average congestion rate as the input of the packet length prediction neural network, and finally outputting the burst data packet length; wherein, the packet length prediction neural network is obtained through training using the historical average congestion rate and the historical burst data packet length as training data.

[0054] BDP generation: According to the calculated BDP packet length, business data of corresponding length is read from the cache of business cache data. In this embodiment, a resource allocation graph algorithm or a banker's algorithm can be used to ensure that deadlock does not occur during data reading.

[0055] BCP generation: Generates a BCP, which contains basic information such as source address, destination address, service priority (medium), transmission path (path 2), offset time (set to 5ms according to the network) and burst packet length.

[0056] Sending queue: After BCP and BDP are generated, they enter the sending queue and wait for the system to send them, thus realizing the assembly of burst packets of variable length.

[0057] It can be seen that this method can effectively predict the network congestion rate, select the optimal path, and generate variable-length burst packets of appropriate length, thereby improving network performance and meeting the needs of different services.

[0058] Compared with the existing routing transmission mode, this embodiment provides a variable-length burst packet allocation method based on routing congestion rate prediction, which has the following advantages: First, improve network resource utilization: by predicting the congestion rate of the core node, selecting the optimal path and adjusting the burst packet length according to the congestion rate, network resources can be allocated more reasonably. Sending longer burst packets on the path with low congestion rate can make full use of the link bandwidth; sending shorter burst packets on the path with high congestion rate can avoid excessive link congestion, thereby improving the utilization of network resources.

[0059] Second, reduce the probability of burst packet collision: dynamically adjust the burst packet length according to the path congestion rate, reducing the possibility of burst packet collision in congested links. Reasonable burst packet length and path selection enable data packets to be transmitted more orderly, reducing the probability of collision and improving the reliability of data transmission.

[0060] Third, improve network service quality: accurate congestion rate prediction and optimal path selection effectively reduce network delay and packet loss. At the same time, communication cost calculation and path selection based on business priority and required bandwidth can better meet the QoS requirements of different businesses and improve the overall service quality of the network.

[0061] Fourth, enhance network adaptability: This method can perceive the network congestion status in real time and dynamically adjust the transmission strategy according to the network status. It has strong adaptability to the suddenness and uncertainty of network traffic and can maintain good performance in different network environments.

[0062] Example 2 like Figure 4As shown, this embodiment provides a method for assembling variable-length burst packets based on routing congestion rate prediction, comprising the following steps: The congestion rate prediction network model based on pre-training is used to predict the congestion rate of each optical network core node at the next moment; Determine the optimal path based on multiple pre-calculated shortest paths and the congestion rate of each optical network core node at the next moment; A burst packet of variable length corresponding to the new service is generated based on the average congestion rate of the optimal path; the burst packet of variable length is used to enter a sending queue and wait for the system to send it.

[0063] Before using the pre-trained congestion rate prediction network model to predict the congestion rate of each optical network core node at the next moment, the congestion rate prediction network model is trained, and the specific steps are as follows: Construct a BP neural network model; wherein the input layer of the BP neural network model is used to receive historical congestion rates, current service traffic, congestion status of adjacent nodes, time information and service type information, the hidden layer is used to process and extract features, and the output layer is used to output predicted congestion rates; The BP neural network model is trained using real node congestion data to obtain a congestion rate prediction network model.

[0064] Specifically, the method of using a pre-trained congestion rate prediction network model to predict the congestion rate of each optical network core node at the next moment includes: Obtain current congestion data of each optical network core node; The acquired current congestion data are input into the congestion rate prediction network model, and the congestion rate of each optical network core node at the next moment is output.

[0065] In this embodiment, before determining the optimal path based on the pre-calculated multiple shortest paths and the congestion rate of each optical network core node at the next moment, multiple shortest paths are obtained by calculation, and the specific steps are as follows: Obtain network topology and existing service information of optical network; Based on the network topology and existing service information of the optical network, the communication cost between adjacent nodes is determined and a communication cost matrix is ​​generated; When new services arrive at the edge node, the k-shortest paths algorithm is used to calculate based on the communication cost matrix to obtain multiple shortest paths corresponding to the new services.

[0066] Here, the determining of the optimal path based on the pre-calculated multiple shortest paths and the congestion rate of each optical network core node at the next moment includes: The average congestion rate of each shortest path is calculated using the congestion rate of each optical network core node at the next moment; Sort the average congestion rate of each shortest path from low to high, and select the shortest path with the lowest average congestion rate as the optimal path.

[0067] In particular, if the average congestion rates of multiple shortest paths are the same, the shortest path with the smallest number of hops is selected as the optimal path.

[0068] Here, the generating of the variable-length burst packet corresponding to the new service based on the average congestion rate of the optimal path includes: Cache new services received by edge nodes to obtain service cache data; Calculate the burst packet length based on the average congestion rate corresponding to the optimal path; Based on the calculated burst data packet length, service data of corresponding length is read from the service buffer data to generate a burst data packet; A burst control packet is generated based on the burst data packet length; wherein the burst data packet and the burst control packet constitute the variable-length burst packet; after being generated, the burst data packet and the burst control packet enter a sending queue and wait for the system to send them.

[0069] like Figure 5 As shown, this embodiment also provides a variable-length burst packet allocation system based on routing congestion rate prediction, including: a congestion rate prediction module, used to use a pre-trained congestion rate prediction network model to predict the congestion rate of each optical network core node at the next moment; an optimal path determination module, used to determine the optimal path based on multiple shortest paths obtained in advance and the congestion rate of each optical network core node at the next moment; a variable-length burst packet generation module, used to generate a variable-length burst packet corresponding to a new service based on the average congestion rate of the optimal path; the variable-length burst packet is used to enter a sending queue and wait for the system to send it.

[0070] The present invention also provides a device, comprising: a memory for storing a computer program; a processor for implementing the steps of the indefinite-length burst packet assembly method based on routing congestion rate prediction when executing the computer program.

[0071] When the processor executes the computer program, the above-mentioned steps of variable-length burst packet assembly based on routing congestion rate prediction are implemented, for example: using a pre-trained congestion rate prediction network model to predict the congestion rate of each optical network core node at the next moment; determining the optimal path based on multiple pre-calculated shortest paths and the congestion rate of each optical network core node at the next moment; generating a variable-length burst packet corresponding to a new service based on the average congestion rate of the optimal path; the variable-length burst packet is used to enter a sending queue and wait for the system to send it.

[0072] Alternatively, the processor implements the functions of each module in the above-mentioned system when executing the computer program, for example: a congestion rate prediction module, which is used to use a pre-trained congestion rate prediction network model to predict the congestion rate of each optical network core node at the next moment; an optimal path determination module, which is used to determine the optimal path based on multiple shortest paths obtained in advance and the congestion rate of each optical network core node at the next moment; a variable-length burst packet generation module, which is used to generate a variable-length burst packet corresponding to a new service based on the average congestion rate of the optimal path; the variable-length burst packet is used to enter a sending queue and wait for the system to send it.

[0073] Exemplarily, the computer program can be divided into one or more modules / units, and the one or more modules / units are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units can be a series of computer program instruction segments that can complete preset functions, and the instruction segments are used to describe the execution process of the computer program in the variable-length burst packet distribution device based on routing congestion rate prediction. For example, the computer program can be divided into a congestion rate prediction module, an optimal path determination module, and a variable-length burst packet generation module; the specific functions of each module are as follows: a congestion rate prediction module, which is used to use a pre-trained congestion rate prediction network model to predict the congestion rate of each optical network core node at the next moment; an optimal path determination module, which is used to determine the optimal path based on multiple pre-calculated shortest paths and the congestion rate of each optical network core node at the next moment; a variable-length burst packet generation module, which is used to generate a variable-length burst packet corresponding to a new service based on the average congestion rate of the optimal path; the variable-length burst packet is used to enter the sending queue and wait for the system to send it.

[0074] The indefinite-length burst package distribution device based on the prediction of the routing congestion rate can be a computing device such as a desktop computer, a notebook, a PDA, and a cloud server. The indefinite-length burst package distribution device based on the prediction of the routing congestion rate may include, but is not limited to, a processor and a memory. Those skilled in the art can understand that the above is an example of an indefinite-length burst package distribution device based on the prediction of the routing congestion rate, and does not constitute a limitation on the indefinite-length burst package distribution device based on the prediction of the routing congestion rate, and may include more components than the above, or a combination of certain components, or different components, for example, the indefinite-length burst package distribution device based on the prediction of the routing congestion rate may also include input and output devices, network access devices, buses, etc.

[0075] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc. The processor is the control center of the variable-length burst packet allocation based on routing congestion rate prediction, and uses various interfaces and lines to connect various parts of the variable-length burst packet allocation device based on routing congestion rate prediction.

[0076] The memory can be used to store the computer program and / or module, and the processor implements various functions of the variable-length burst packet distribution device based on routing congestion rate prediction by running or executing the computer program and / or module stored in the memory and calling the data stored in the memory.

[0077] The memory may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0078] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for assembling variable-length burst packets based on routing congestion rate prediction are implemented.

[0079] If the module / unit integrated in the variable-length burst packet distribution system based on routing congestion rate prediction is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium.

[0080] Based on such understanding, the present invention implements all or part of the processes in the above-mentioned method for assembling indefinite-length burst packets based on routing congestion rate prediction, and can also be completed by instructing related hardware through a computer program, and the computer program can be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of the above-mentioned method for assembling indefinite-length burst packets based on routing congestion rate prediction can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or preset intermediate form, etc.

[0081] The computer-readable storage medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium, etc.

[0082] It should be noted that the content contained in the computer-readable storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable storage media do not include electrical carrier signals and telecommunication signals.

[0083] The above embodiment is only one of the implementation methods that can realize the technical solution of the present invention. The scope of protection claimed by the present invention is not limited to this embodiment alone, but also includes changes, replacements and other implementation methods that can be easily thought of by any technician familiar with the technical field within the technical scope disclosed by the present invention. Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for allocating variable-length burst packets based on routing congestion rate prediction, characterized in that: include: The congestion rate prediction network model based on pre-training is used to predict the congestion rate of each optical network core node at the next moment; Determine the optimal path based on multiple pre-calculated shortest paths and the congestion rate of each optical network core node at the next moment; A burst packet of variable length corresponding to the new service is generated based on the average congestion rate of the optimal path; the burst packet of variable length is used to enter a sending queue and wait for the system to send it.

2. The method for allocating variable-length burst packets based on routing congestion rate prediction according to claim 1, characterized in that: Before using the pre-trained congestion rate prediction network model to predict the congestion rate of each optical network core node at the next moment, the congestion rate prediction network model is trained, and the specific steps are as follows: Construct a BP neural network model; wherein the input layer of the BP neural network model is used to receive historical congestion rates, current service traffic, congestion status of adjacent nodes, time information and service type information, the hidden layer is used to process and extract features, and the output layer is used to output predicted congestion rates; The BP neural network model is trained using real node congestion data to obtain a congestion rate prediction network model.

3. The method for assembling variable-length burst packets based on routing congestion rate prediction according to claim 1, characterized in that: The method of using a pre-trained congestion rate prediction network model to predict the congestion rate of each optical network core node at the next moment includes: Obtain current congestion data of each optical network core node; The acquired current congestion data are input into the congestion rate prediction network model, and the congestion rate of each optical network core node at the next moment is output.

4. The method for allocating variable-length burst packets based on routing congestion rate prediction according to claim 1, characterized in that: Before determining the optimal path based on the multiple shortest paths obtained by pre-calculation and the congestion rate of each optical network core node at the next moment, multiple shortest paths are obtained by calculation, and the specific steps are as follows: Obtain network topology and existing service information of optical network; Based on the network topology and existing service information of the optical network, the communication cost between adjacent nodes is determined and a communication cost matrix is ​​generated; When new services arrive at the edge node, the k-shortest paths algorithm is used to calculate based on the communication cost matrix to obtain multiple shortest paths corresponding to the new services.

5. The method for allocating variable-length burst packets based on routing congestion rate prediction according to claim 1, characterized in that: The determining of the optimal path based on the pre-calculated multiple shortest paths and the congestion rate of each optical network core node at the next moment includes: The average congestion rate of each shortest path is calculated using the congestion rate of each optical network core node at the next moment; Sort the average congestion rate of each shortest path from low to high, and select the shortest path with the lowest average congestion rate as the optimal path.

6. The method for assembling variable-length burst packets based on routing congestion rate prediction according to claim 5, characterized in that: If the average congestion rates of multiple shortest paths are the same, the shortest path with the smallest number of hops is selected as the optimal path.

7. The method for allocating variable-length burst packets based on routing congestion rate prediction according to claim 1, characterized in that: The generating of the variable-length burst packets corresponding to the new service based on the average congestion rate of the optimal path includes: Cache new services received by edge nodes to obtain service cache data; Calculate the burst packet length based on the average congestion rate corresponding to the optimal path; Based on the calculated burst data packet length, service data of corresponding length is read from the service buffer data to generate a burst data packet; A burst control packet is generated based on the burst data packet length; wherein the burst data packet and the burst control packet constitute the variable-length burst packet; after being generated, the burst data packet and the burst control packet enter a sending queue and wait for the system to send them.

8. A variable-length burst packet distribution system based on routing congestion rate prediction, characterized in that: include: A congestion rate prediction module, used to predict the congestion rate of each optical network core node at the next moment using a pre-trained congestion rate prediction network model; An optimal path determination module, used to determine the optimal path based on multiple shortest paths obtained in advance and the congestion rate of each optical network core node at the next moment; The variable length burst packet generation module is used to generate variable length burst packets corresponding to new services based on the average congestion rate of the optimal path; the variable length burst packets are used to enter the sending queue and wait for the system to send them.

9. A device, characterized in that: include: Memory for storing computer programs; A processor is used to implement the steps of the variable-length burst packet assembly method based on routing congestion rate prediction as described in any one of claims 1-7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, it is used to implement the steps of the variable-length burst packet assembly method based on routing congestion rate prediction as described in any one of claims 1 to 7.

Images