Method for dynamic allocation of network slice bandwidth, electronic device and storage medium

Abstract

The application provides a network slice broadband dynamic allocation method, an electronic device and a storage medium, and comprises the following steps: taking the periodic historical traffic of each network slice in the current access network service as the input data of a gated recurrent neural network, so that the gated recurrent neural network outputs the predicted traffic of each network slice in the next bandwidth allocation period after the current time; determining the total demand bandwidth of the access network service and the proportional relationship of the demand bandwidth of each network slice in the total demand bandwidth according to the predicted traffic of each network slice; and allocating the total bandwidth to be allocated in the target network corresponding to the access network service according to the proportional relationship of the demand bandwidth of each network slice in the total demand bandwidth, so as to obtain the allocation bandwidth of each network slice. The application can obtain the predicted traffic value of each network slice more quickly and accurately, can more reasonably complete the resource allocation of the network bandwidth in the dynamic slice, and can meet various service demands of the service forwarding in the Internet.

Description

Technical Field

[0001] This invention relates to the field of access network slicing resource allocation technology, and in particular to a dynamic allocation method, electronic device, and storage medium for network slice bandwidth. Background Technology

[0002] The widespread use of the Internet has led to increasing demands for network slicing. To address the diverse application scenarios of the Internet and meet the traffic patterns of different types of services, network slicing supported by Software-Defined Networking (SDN) and Network Functions Virtualization (NFV) is commonly used. This involves building multiple dedicated logical networks on a single physical platform to meet the different network capability needs of various customers, enabling flexible service customization and improving network resource utilization. However, due to the inherent burstiness of Internet traffic, early network slice resource allocation was mostly focused on static deployment, neglecting the subsequent dynamic deployment and expansion requirements. This resulted in poor flexibility in network slice resource allocation, leading to network slice resources allocated to different users that do not meet their specific needs and, to some extent, wasting physical resources. Summary of the Invention

[0003] In view of this, embodiments of the present invention provide a method for dynamic allocation of network slice bandwidth, an electronic device, and a storage medium to eliminate or improve one or more defects existing in the prior art.

[0004] One aspect of the present invention provides a method for dynamic allocation of network slice bandwidth, comprising: statistically analyzing the periodic historical traffic of each network slice in the current access network service according to a pre-set bandwidth allocation period; using the periodic historical traffic of each network slice as input data of a gated recurrent neural network, so that the gated recurrent neural network outputs the predicted traffic of each network slice in the access network service in the next bandwidth allocation period after the current time; determining the total required bandwidth of the access network service and the proportion of the required bandwidth of each network slice in the total required bandwidth according to the predicted traffic of each network slice; allocating the total bandwidth to be allocated in the target network corresponding to the access network service according to the proportion of the required bandwidth of each network slice in the total required bandwidth, thereby obtaining the allocated bandwidth of each network slice in the access network service.

[0005] In some embodiments of the present invention, determining the total required bandwidth of the access network service based on the predicted traffic of each network slice, and the proportional relationship of the required bandwidth of each network slice in the total required bandwidth, includes:

[0006] The required bandwidth of each network slice is determined by using the predicted traffic of each network slice and combining the priority coefficient of the network slice; the total required bandwidth of the access network service is determined based on the required bandwidth of each network slice; and the proportion of the required bandwidth of each network slice in the total required bandwidth is obtained.

[0007] In some embodiments of the present invention, the step of allocating the total bandwidth to be allocated in the network according to the proportional relationship to obtain the allocated bandwidth of each network slice in the access network service includes: allocating the remaining total bandwidth to be allocated in the target network after removing the minimum overhead bandwidth of each network slice obtained in advance, according to the proportional relationship of the required bandwidth of each network slice in the total required bandwidth of the access network service, to obtain the additional required bandwidth allocated to each network slice.

[0008] The allocated bandwidth for each network slice is obtained by combining the minimum overhead bandwidth and the required bandwidth of each network slice.

[0009] In some embodiments of the present invention, the method further includes: based on the historical traffic of each network slice, pre-setting a lower limit threshold for the ratio of the guaranteed overhead bandwidth of each network slice to the total bandwidth to be allocated in the target network, so that the minimum overhead bandwidth of each reserved network slice meets the service forwarding needs during the off-peak period of network service forwarding.

[0010] In some embodiments of the present invention, after determining the proportion of the required bandwidth of each network slice in the total required bandwidth of the access network service based on the predicted traffic of each network slice, the method further includes: if the proportion of the required bandwidth of a network slice in the total required bandwidth of the access network service is greater than a pre-set upper limit threshold for the proportion of the guaranteed overhead bandwidth of the network slice to the total bandwidth to be allocated in the network, then the upper limit threshold of the network slice is corrected to the proportion of the required bandwidth of the network slice in the total required bandwidth of the access network service.

[0011] In some embodiments of the present invention, the service types of the access network services include low-latency type and high-speed type;

[0012] For the low-latency access network service, the required bandwidth of each network slice is determined by using the predicted traffic of each network slice and combining the priority coefficient of the network slice, including: using the predicted traffic of each network slice and combining the priority coefficient of the network slice and the network latency coefficient of the access network service to determine the required bandwidth of each network slice.

[0013] For the high-speed access network service, the required bandwidth of each network slice is determined by using the predicted traffic of each network slice and combining the priority coefficient of the network slice. This includes: using the predicted traffic of each network slice and combining the priority coefficient of the network slice with the network rate coefficient of the access network service to determine the required bandwidth of each network slice.

[0014] In some embodiments of the present invention, the gated recurrent neural network uses stochastic gradient descent as the optimization algorithm and hyperbolic tangent function as the activation function.

[0015] In some embodiments of the present invention, the method further includes: training the gated recurrent neural network using historical traffic from each network slice in each bandwidth allocation period, and obtaining the training effect value of the gated recurrent neural network through the mean absolute percentage error; the mean absolute percentage error is:

[0016]

[0017] Where N represents the total number of network slices, This represents the actual traffic value of the nth network slice. This represents the predicted traffic value for the nth network slice.

[0018] Another aspect of the present invention provides an electronic device including a processor and a memory, characterized in that the memory stores computer instructions, and the processor is configured to execute the computer instructions stored in the memory, wherein when the computer instructions are executed by the processor, the electronic device implements the steps of the network slice bandwidth dynamic allocation method described above.

[0019] Another aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, characterized in that, when executed by a processor, the program implements the steps of the dynamic allocation method for network slice bandwidth as described above.

[0020] The network slicing bandwidth dynamic allocation method, electronic device, and storage medium of the present invention can obtain the predicted traffic value of each network slice more quickly and accurately through a gated recurrent neural network; based on this, the network bandwidth of the target network is allocated, which can more rationally complete the bandwidth resource allocation in the dynamic slice and meet the various service requirements of service forwarding in the industrial Internet.

[0021] Additional advantages, objects, and features of the invention will be set forth in part in the description which follows, and will also become apparent in part to those skilled in the art upon studying the description, or may be learned by practice of the invention. The objects and other advantages of the invention can be realized and obtained by means of the structures specifically pointed out in the description and drawings.

[0022] Those skilled in the art will understand that the objectives and advantages achievable with the present invention are not limited to those specifically described above, and that the above and other objectives achievable with the present invention will become clearer from the following detailed description. Attached Figure Description

[0023] The accompanying drawings, which are provided to further illustrate the invention and form part of this application, are not intended to limit the scope of the invention.

[0024] Figure 1 A flowchart for the dynamic allocation of bandwidth for network slices.

[0025] Figure 2 This is a structural diagram of a gated recurrent neural network model in one embodiment of the present invention.

[0026] Figure 3 This is a flowchart illustrating the training process of a gated recurrent neural network according to one embodiment of the present invention.

[0027] Figure 4 This is a structural diagram of an access network system according to an embodiment of the present invention.

[0028] Figure 5 This is a network slicing diagram of access network services in one embodiment of the present invention.

[0029] Figure 6 This is a flowchart illustrating the dynamic allocation of network slice bandwidth in one embodiment of the present invention. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the embodiments and accompanying drawings. Here, the illustrative embodiments and descriptions of this invention are used to explain the invention, but are not intended to limit the invention.

[0031] It should also be noted that, in order to avoid obscuring the invention with unnecessary details, only the structures and / or processing steps closely related to the solution according to the invention are shown in the accompanying drawings, while other details that are not closely related to the invention are omitted.

[0032] It should be emphasized that the term "including / comprises" as used herein refers to the presence of a feature, element, step, or component, but does not exclude the presence or addition of one or more other features, elements, steps, or components.

[0033] It should also be noted that, unless otherwise specified, the term "connection" in this article can refer not only to a direct connection, but also to an indirect connection involving an intermediary.

[0034] In the following description, embodiments of the invention will be illustrated with reference to the accompanying drawings. In the drawings, the same reference numerals represent the same or similar parts, or the same or similar steps.

[0035] The background technology neglects the subsequent deployment and expansion requirements of network slices, resulting in poor resource allocation flexibility and unsuitable resource allocation to different users, leading to a waste of physical resources. Currently, traditional traffic prediction algorithms are commonly used to calculate the predicted traffic for each network slice; however, these algorithms are overly complex and require significant computational resources. Long Short-Term Memory (LSTM) neural networks can also be used for traffic prediction, but their complex model structure and high training cost further complicate the process. This application employs a simpler model structure, the Gated Recurrent Neural Network (GRU), which significantly reduces training costs and provides faster and more accurate predicted traffic values.

[0036] This application provides a method for dynamically allocating bandwidth in network slicing, such as... Figure 1 As shown, steps S110-S130 are included:

[0037] Step S110: According to the preset bandwidth allocation period, the periodic historical traffic of each network slice in the current access network service is statistically analyzed, and the periodic historical traffic of each network slice is used as the input data of the gated recurrent neural network, so that the gated recurrent neural network outputs the predicted traffic of each network slice in the access network service in the next bandwidth allocation period after the current time.

[0038] The model structure of the gated recurrent neural network is as follows: Figure 2 As shown, where, The input information for the gated neural network represents the actual traffic data of the network slice at the current moment; The hidden state of the gated recurrent neural network at the previous time step serves as the neural network memory of the gated recurrent neural network, containing information about the flow data from the previous time step. This indicates a hidden state that will be passed on to the next time step. This represents the candidate hidden state at the current moment; This is the reset gate of the gated recurrent neural network, used to combine the actual traffic data of the network slice input at the current time with the historical traffic data input to the gated recurrent neural network at previous times; This is the update gate of the gated recurrent neural network, used to control the extent to which hidden information from the previous time step is brought into the current gated recurrent neural network; : indicates the use of an S-shaped growth curve (tanh function), which can transform the input data into values ​​in the range of 0-1; tanh: indicates the hyperbolic tangent function (tanh function), which can transform the input data into values ​​in the range of [-1,1].

[0039] The training of the gated recurrent neural network is as follows: Figure 3 The method includes: using the bandwidth allocation period as the sampling period of the gated recurrent neural network (RNN), collecting historical traffic from each network slice in the access network service, and forming a sampling dataset of the periodic historical traffic for each network slice of the RNN according to the sampling period. The periodic historical traffic in the sampling dataset of each network slice is divided into a training set, a validation set, and a test set according to an appropriate ratio. The periodic historical traffic in the training set of each network slice is preprocessed using a normalization method. The preprocessed periodic historical traffic is then used for training the RNN, and the training effect value of the RNN is obtained through the mean absolute percentage error (MASE). The MASE is:

[0040]

[0041] Where N represents the total number of network slices, This represents the actual traffic value of the nth network slice. This represents the predicted traffic value of the nth network slice. When the mean absolute percentage error obtained by the gated recurrent neural network is less than a preset training standard threshold, the traffic prediction results of the trained gated recurrent neural network are verified and tested using the validation set and test set until the mean absolute percentage error obtained by the gated recurrent neural network is less than the preset training standard threshold.

[0042] In one or more embodiments of this application, the historical traffic of each network slice in the access network service is statistically analyzed according to a pre-set bandwidth allocation period. The periodic historical traffic of each network slice in the current bandwidth allocation period is used as the input data of the trained gated recurrent neural network. In the gated recurrent neural network, under the influence of the hidden state after the prediction is completed in the previous bandwidth allocation period, the predicted traffic of the network slice in the next bandwidth allocation period is output, while the hidden state of the current bandwidth allocation period is retained.

[0043] The gated recurrent neural network used in this application has a simple structure and can obtain the predicted traffic for the next bandwidth allocation cycle more quickly and accurately.

[0044] Step S120: Determine the total bandwidth requirement of the access network service and the proportion of the bandwidth requirement of each network slice in the total bandwidth requirement based on the predicted traffic of each network slice.

[0045] In this application, based on the historical traffic of each network slice in the access network service of the target network, in order to ensure that each network slice can at least meet the service forwarding needs during off-peak hours in the target network, a lower limit threshold is pre-set for the ratio of the guaranteed overhead bandwidth of each network slice to the total bandwidth to be allocated in the target network. This ensures that during bandwidth allocation, a minimum guaranteed overhead bandwidth is reserved for each network slice to meet service forwarding needs during off-peak network traffic. Simultaneously, to guarantee that the guaranteed overhead bandwidth reserved for each network slice does not hinder service forwarding in other network slices, an upper limit threshold is pre-set for the ratio of the guaranteed overhead bandwidth of each network slice to the total bandwidth to be allocated in the target network. If the proportion of the bandwidth demanded by a network slice in the total bandwidth demanded by the access network service is greater than the upper limit threshold of the proportion of the guaranteed overhead bandwidth of the network slice to the total bandwidth to be allocated in the network, then the upper limit threshold of the network slice needs to be adjusted to the proportion of the bandwidth demanded by the network slice in the total bandwidth demanded by the access network service.

[0046] In one or more embodiments of this application, determining the total bandwidth requirement of the access network service based on the predicted traffic of each network slice includes the following steps S121-S123:

[0047] Step S121: Utilize the predicted traffic of each network slice and combine it with the priority coefficient of that network slice to determine the required bandwidth of each network slice in the next bandwidth allocation cycle; specifically, based on the predicted traffic of each network slice obtained from the gated recurrent neural network in the next bandwidth allocation cycle, combine it with the priority coefficient of each network slice to determine the required bandwidth of each network slice in the next bandwidth allocation cycle. ,in, This represents the predicted traffic for the nth network slice in the next bandwidth allocation cycle. This represents the priority coefficient of the nth network slice;

[0048] Step S122: Determine the total required bandwidth of the access network service based on the required bandwidth of each network slice; the total required bandwidth of the access network service is: Where n represents the network slice number in the access network service, n=1,2,3,…,N, and N represents the total number of network slices in the access network service.

[0049] Step S123: Obtain the proportion of each network slice's required bandwidth in the total required bandwidth; specifically, this includes: after obtaining the required bandwidth of all network slices in the next bandwidth allocation cycle, determining the proportion of each network slice's required bandwidth in the total required bandwidth:

[0050]

[0051] in, This represents the relative ratio of the bandwidth demand of the nth network slice in the next bandwidth allocation cycle.

[0052] This includes the relative ratio of the bandwidth demand of each network slice in the next bandwidth allocation cycle. Compared with a preset upper limit threshold, if Then the upper limit threshold of the network slice. Revised to Otherwise, there is no need to modify the pre-set online thresholds for each network slice. .

[0053] In one or more embodiments of this application, based on the predicted traffic of each network slice in the next bandwidth allocation cycle obtained from the gated recurrent neural network, a QoS (Quality of Service) coefficient is introduced to ensure that the bandwidth allocated to each network slice can meet its respective quality of service. For example, for common access network services, the service types include low-latency and high-speed types. For low-latency access network services, the required bandwidth of each network slice is determined by combining the predicted traffic of each network slice, the priority coefficient of the network slice, and the network latency coefficient of the access network service. For high-speed access network services, the required bandwidth of each network slice is determined by combining the predicted traffic of each network slice, the priority coefficient of the network slice, and the network speed coefficient of the access network service. The formula for calculating the network speed coefficient s of the access network service is: ,in, , This represents the actual historical traffic of the nth network slice. This represents the threshold for evaluating the network speed of the access network service. If the network speed of the network slice transmission is normal, it is considered excellent; otherwise, it is considered deficient. The formula for calculating the network delay coefficient td of the access network service is: ,in, , This represents the threshold for evaluating the network latency of the access network service. If the network latency of the network slice transmission is normal, it is considered excellent; otherwise, it is considered deficient.

[0054] In the above embodiments, for low-latency access network services, the bandwidth requirement of each network slice in the next bandwidth allocation cycle is: In the high-speed access network services described above, the bandwidth requirement for each network slice in the next bandwidth allocation cycle is: .

[0055] Step S130: Allocate the total bandwidth to be allocated in the target network corresponding to the access network service according to the proportion of the required bandwidth of each network slice in the total required bandwidth, and obtain the allocated bandwidth of each network slice in the access network service.

[0056] The above step S130 specifically includes: based on a lower limit threshold of the ratio between the guaranteed overhead bandwidth of each network slice and the total bandwidth to be allocated in the target network, as preset. Determine the minimum guaranteed overhead bandwidth for each network slice; the minimum guaranteed overhead bandwidth is... ,in, This represents the total bandwidth to be allocated in the target network corresponding to the access network service.

[0057] Then, based on the proportion of the required bandwidth of each network slice in the total required bandwidth of the access network service, the remaining total bandwidth to be allocated in the target network after removing the minimum guaranteed overhead bandwidth of each network slice obtained in advance is allocated to obtain the additional required bandwidth allocated to each network slice; combining the minimum overhead bandwidth and required bandwidth of each network slice, the allocated bandwidth of each network slice is obtained. The allocated bandwidth of each network slice is:

[0058]

[0059] in, This represents the total bandwidth allocated to the nth network slice. This represents the total bandwidth remaining to be allocated in the target network after removing the minimum guaranteed overhead bandwidth of each pre-acquired network slice.

[0060] This application first allocates a minimum guaranteed overhead bandwidth to each network slice to ensure that each network slice can complete service forwarding during the off-peak hours of the target network, thereby effectively reducing or even avoiding packet loss caused by uneven bandwidth allocation. Then, according to the proportion of the bandwidth demand of each network slice in the total bandwidth demand of the access network services, the remaining total bandwidth to be allocated in the target network after removing the pre-acquired minimum guaranteed overhead bandwidth of each network slice is allocated, so that the transmission bandwidth in the target network is evenly distributed to each network slice according to its own needs, which can effectively avoid the waste of bandwidth resources in the target network.

[0061] In one or more embodiments of this application, the access network structure of the dynamic allocation method for network slicing bandwidth is as follows: Figure 4As shown, in the control plane of a software-defined network (SDN), the SDN controller classifies access network services in the target network, including factory services and office services; factory services are low-latency types, and office services are high-speed types. Based on the traffic structure of the factory and office services, network slices for each access network service are determined, such as... Figure 5 As shown. In this embodiment, a gated recurrent neural network with stochastic gradient descent as the optimization algorithm and hyperbolic tangent function as the activation function is used to obtain the predicted traffic of each network slice in each access network service; wherein, during the training process of the gated recurrent neural network, the historical traffic in the historical traffic set is divided into training set, validation set and test set according to the proportions of 80%, 10% and 10% to train the gated recurrent neural network.

[0062] In this embodiment, the flowchart for the dynamic bandwidth allocation of each network slice in each access network service of the target network is as follows: Figure 6 As shown:

[0063] For each network slice in each access network service of the target network, the periodic historical traffic of the network slice in the current bandwidth allocation period is statistically analyzed, and the periodic historical traffic in the current bandwidth allocation period is input into the trained gated recurrent neural network, which outputs the predicted traffic for the next bandwidth allocation period.

[0064] For each network slice in the factory operations, the priority coefficient of that network slice is used. The predicted traffic of this network slice in the next bandwidth allocation cycle. And the network latency coefficient of the project's services A combined approach is used to obtain the bandwidth requirements of each network slice. Based on this, minimum guaranteed overhead bandwidth is allocated to each network slice in the factory operations; for each network slice in the office building operations, the bandwidth is allocated according to the priority coefficient of that network slice. The predicted traffic of this network slice in the next bandwidth allocation cycle. By combining the network latency coefficient and network rate coefficient 's' of the office building's services, the required bandwidth of each network slice can be obtained. Based on this, minimum guaranteed overhead bandwidth is allocated to each network slice in the office building business.

[0065] Then, the remaining total bandwidth to be allocated in the target network, after deducting the minimum guaranteed overhead bandwidth of each pre-acquired network slice, is allocated to obtain the allocated bandwidth for each network slice. Simultaneously, by comparing the proportion of each network slice's required bandwidth in the total required bandwidth of the access network service corresponding to that network slice with a pre-set upper limit threshold for the proportion of the guaranteed overhead bandwidth of that network slice to the total bandwidth to be allocated in the network, the upper limit threshold for each network slice in the access network service is updated.

[0066] Corresponding to the above method, the present invention also provides an electronic device, which includes a computer device, the computer device including a processor and a memory, the memory storing computer instructions, the processor executing the computer instructions stored in the memory, and when the computer instructions are executed by the processor, the electronic device implements the steps of the dynamic allocation method for network slice bandwidth as described above.

[0067] This invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the aforementioned dynamic allocation method for network slicing bandwidth. The computer-readable storage medium can be a tangible storage medium, such as random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, floppy disks, hard disks, removable storage disks, CD-ROMs, or any other form of storage medium known in the art.

[0068] Those skilled in the art will understand that the exemplary components, systems, and methods described in conjunction with the embodiments disclosed herein can be implemented in hardware, software, or a combination of both. Whether implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this invention. When implemented in hardware, it can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this invention are programs or code segments used to perform the desired tasks. The programs or code segments can be stored in a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried in a carrier wave.

[0069] It should be clarified that the present invention is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of the present invention.

[0070] In this invention, features described and / or illustrated for one embodiment may be used in the same or similar manner in one or more other embodiments, and / or combined with or in place of features of other embodiments.

[0071] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, various modifications and variations of the embodiments of the present invention are possible. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for dynamically allocating bandwidth in network slicing, characterized in that, include: According to the preset bandwidth allocation period, the periodic historical traffic of each network slice in the current access network service is statistically analyzed, and the periodic historical traffic of each network slice is used as the input data of the gated recurrent neural network so that the gated recurrent neural network outputs the predicted traffic of each network slice in the access network service in the next bandwidth allocation period after the current time. The total bandwidth requirement of the access network service is determined based on the predicted traffic of each network slice, and the proportion of the bandwidth requirement of each network slice in the total bandwidth requirement. Based on the proportion of the required bandwidth of each network slice in the total required bandwidth, the total bandwidth to be allocated in the target network corresponding to the access network service is allocated, thereby obtaining the allocated bandwidth of each network slice in the access network service. The total bandwidth requirement for the access network service is determined based on the predicted traffic of each network slice, and the proportion of each network slice's bandwidth requirement in the total bandwidth requirement includes: Based on the predicted traffic of each network slice in the next bandwidth allocation cycle obtained from the gated recurrent neural network, and combined with the priority coefficient of each network slice, the required bandwidth of each network slice in the next bandwidth allocation cycle is determined, as follows: ,in, This represents the priority coefficient of the nth network slice. This represents the predicted traffic for the nth network slice in the next bandwidth allocation cycle; The total bandwidth requirement for the access network service is determined based on the bandwidth requirements of each network slice; the total bandwidth requirement for the access network service is: Where n represents the network slice number in the access network service, n=1,2,3,…,N, and N represents the total number of network slices in the access network service; Once the bandwidth requirements of all network slices in the next bandwidth allocation cycle are obtained, the proportion of each network slice's bandwidth requirement in the total bandwidth requirement is determined using the following formula: ； in, This represents the relative bandwidth demand of the nth network slice in the next bandwidth allocation cycle.

2. The method according to claim 1, characterized in that, Based on the aforementioned proportional relationship, the total bandwidth to be allocated in the network is allocated to obtain the allocated bandwidth for each network slice in the access network service, including: Based on the proportion of the required bandwidth of each network slice in the total required bandwidth of the access network service, the remaining total bandwidth to be allocated in the target network after removing the minimum overhead bandwidth of each network slice obtained in advance is allocated to obtain the additional required bandwidth allocated to each network slice. The allocated bandwidth for each network slice is obtained by combining the minimum overhead bandwidth and the required bandwidth of each network slice.

3. The method according to claim 1, characterized in that, Also includes: Based on the historical traffic of each network slice, a lower limit threshold is pre-set for the ratio of the guaranteed overhead bandwidth of each network slice to the total bandwidth to be allocated in the target network, so that the minimum overhead bandwidth of each reserved network slice can meet the service forwarding needs during the off-peak period of network service forwarding.

4. The method according to claim 1, characterized in that, After determining the proportion of the required bandwidth of each network slice in the total required bandwidth of the access network service based on the predicted traffic of each network slice, the process further includes: If the ratio of the bandwidth demanded by a network slice to the total bandwidth demanded by the access network service is greater than the upper limit threshold of the ratio of the guaranteed overhead bandwidth of the network slice to the total bandwidth to be allocated in the network, then the upper limit threshold of the network slice is corrected to the ratio of the bandwidth demanded by the network slice to the total bandwidth demanded by the access network service.

5. The method according to claim 1, characterized in that, The service types of the access network services include low-latency and high-speed types; For the low-latency access network service, the required bandwidth of each network slice is determined by using the predicted traffic of each network slice and combining the priority coefficient of the network slice, including: using the predicted traffic of each network slice and combining the priority coefficient of the network slice and the network latency coefficient of the access network service to determine the required bandwidth of each network slice. For the high-speed access network service, the required bandwidth of each network slice is determined by using the predicted traffic of each network slice and combining the priority coefficient of the network slice. This includes: using the predicted traffic of each network slice and combining the priority coefficient of the network slice with the network rate coefficient of the access network service to determine the required bandwidth of each network slice.

6. The method according to claim 1, characterized in that, The gated recurrent neural network uses stochastic gradient descent as the optimization algorithm and hyperbolic tangent function as the activation function.

7. The method according to claim 1, characterized in that, Also includes: The gated recurrent neural network is trained using historical traffic from each network slice in each bandwidth allocation cycle, and the training performance value of the gated recurrent neural network is obtained through the mean absolute percentage error; the mean absolute percentage error is: Where N represents the total number of network slices, This represents the actual traffic value of the nth network slice. This represents the predicted traffic value for the nth network slice.

8. An electronic device comprising a processor and a memory, characterized in that, The memory stores computer instructions, and the processor executes the computer instructions stored in the memory. When the computer instructions are executed by the processor, the electronic device implements the steps of the dynamic allocation method for network slice bandwidth as described in any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the dynamic allocation method for network slice bandwidth as described in any one of claims 1 to 7.

Images