Intelligent IPv6 address allocation model and optimization method thereof

Abstract

The invention belongs to the technical field of IPv6 address allocation, and relates to an intelligent IPv6 address allocation model and an optimization method thereof, and the method comprises the steps: obtaining and preprocessing equipment use records, address allocation records and flow information, determining an IPv6 address pool range according to network planning, and dividing address blocks according to different regions; predicting a future address demand by using the preprocessed data and adopting a prediction model, setting an allocation priority rule according to equipment and user types, dynamically allocating an IPv6 address, and when the demand quantity exceeds the available address quantity, allocating according to a priority sequence; regularly counting the state of the address pool, analyzing the fragmentation degree, and identifying and reporting the fragmentation condition through a clustering algorithm; according to the fragmented analysis report and a preset threshold value, the address blocks are divided again; intelligent prediction, priority allocation and fragmentation optimization technologies are comprehensively applied, the flexibility and efficiency of IPv6 address allocation are effectively improved, and diversified and dynamic address requirements in a modern network environment are met.

Description

Technical Field

[0001] This application belongs to the technical field of IPv6 address allocation. More specifically, it relates to an intelligent IPv6 address allocation model and its optimization method. Background Art

[0002] With the rapid development of the Internet, the number of global network devices has shown an explosive growth. The traditional IPv4 address space is gradually running out, and the IPv6 protocol has emerged to provide a wider address space. However, although the IPv6 address space is huge, there are still many challenges in how to allocate and manage addresses efficiently and reasonably. The current IPv6 address allocation methods mainly focus on static allocation and simple dynamic allocation, and these methods have many deficiencies when facing complex and changing network environments.

[0003] The static allocation method is often used in early network planning. Its characteristic is to pre-allocate fixed address blocks to different network regions or devices. The main problems of this method include:

[0004] Low utilization rate: Since the allocated address blocks remain unchanged for a long time, there may be a large difference between the actual usage and the pre-planning, resulting in some address blocks being idle for a long time, while there may be a shortage of addresses in other regions.

[0005] Poor flexibility: Static allocation cannot quickly respond to changes in network topology and the number of devices, and it is necessary to manually adjust the allocation scheme frequently, with high costs and low efficiency. Summary of the Invention

[0006] The present invention provides an intelligent IPv6 address allocation model and its optimization method, aiming to solve the technical problems of low utilization rate and poor flexibility of current static allocation.

[0007] The intelligent IPv6 address allocation model and its optimization method include the following steps:

[0008] Step 1: Obtain device usage records, address allocation records, and traffic information, and preprocess the obtained data;

[0009] Step 2: According to the network plan, determine the range of the IPv6 address pool, output the available range of the IPv6 address pool, and divide the address blocks according to different regions according to the available range of the IPv6 address pool;

[0010] Step 3: Based on the preprocessed device usage records, address allocation records, and traffic information, use a prediction model to predict the address demand in the next period of time;

[0011] Step 4: Specify the allocation priority rules according to the device type and user type, and dynamically allocate IPv6 addresses based on the predicted address requirements, priority rules, and divided address blocks. If the demand exceeds the current available address quantity, allocate them in the order of priority;

[0012] Step 5: Based on the address allocation records and the status of the address pool, regularly count the number of unused and used address blocks, analyze the degree of fragmentation, use the clustering algorithm to analyze the usage of address blocks, and obtain a fragmentation analysis report;

[0013] Step 6: Re-divide the address blocks according to the fragmentation analysis report and the preset threshold.

[0014] The present invention ensures the accuracy and consistency of data by acquiring and preprocessing device usage records, address allocation records, and traffic information. Then, determine the scope of the IPv6 address pool according to the network plan and divide the address blocks by different regions. Utilize the preprocessed data and adopt a prediction model to predict future address requirements to ensure the forward-looking nature of the allocation scheme. Set the allocation priority rules according to the device and user types and dynamically allocate IPv6 addresses to ensure that key devices and users obtain resources first. When the demand exceeds the available address quantity, allocate them in the order of priority to avoid waste of resources. Regularly count the status of the address pool, analyze the degree of fragmentation, and identify and report the fragmentation situation through the clustering algorithm. According to the fragmentation analysis report and the preset threshold, re-divide the address blocks to reduce fragmentation and optimize the address utilization rate. This method comprehensively applies intelligent prediction, priority allocation, and fragmentation optimization technologies, effectively improving the flexibility and efficiency of IPv6 address allocation and meeting the diverse and dynamic address requirements in the modern network environment.

[0015] Preferably, step 2 includes the following steps:

[0016] Determine the region category: According to the business requirements and network plan, divide the network region into a core region, a general region, and an edge region;

[0017] Set the initial allocation of address blocks: Set the initial allocation size of address blocks for different types of regions, where the address block size of the core region is larger than that of the general region, and the address block size of the general region is larger than that of the edge region;

[0018] Dynamic adjustment: Dynamically adjust the address block allocation according to the actual usage situation. If the actual usage quantity is less than the initial allocation quantity, recycle some unused address blocks; if the actual demand of a certain region exceeds the initial allocation quantity, allocate additional address blocks from the remaining address pool.

[0019] Preferably, the prediction model adopts an LSTM model, and the specific structure is as follows:

[0020] Input layer: used for data input;

[0021] The first LSTM layer: includes 50 LSTM units and returns sequences as True;

[0022] The second LSTM layer: includes 50 LSTM units and returns sequences as False;

[0023] Fully connected layer: includes one fully connected unit for outputting a predicted value;

[0024] Output layer: processes the predicted value output by the fully connected layer using a linear activation function and then outputs it.

[0025] Preferably, before the data is input into the prediction model, feature processing is performed on the device usage records, address allocation records, and traffic information, including the following steps:

[0026] Device usage records:

[0027] Extract time features: extract periodic features from the timestamp;

[0028] Device features: extract device ID, device type, and user type features;

[0029] Usage statistical features: extract the number of sessions and session duration;

[0030] Address allocation records:

[0031] Time features: extract periodic features from the timestamp;

[0032] Allocation features: extract the number of addresses allocated each time and the allocation frequency;

[0033] Historical allocation features: extract the moving average and cumulative sum of the allocation quantity over a past period of time;

[0034] Traffic information:

[0035] Time features: extract periodic features from the timestamp;

[0036] Traffic features: extract traffic size and traffic type features;

[0037] Historical traffic features: the moving average and cumulative sum of the traffic size over a past period of time;

[0038] Normalize all the extracted features; finally, convert the data format into a three-dimensional array [samples, time_steps, features], where samples represents the number of samples; time_steps represents the number of time steps for each sample; features represents the number of features for each time step.

[0039] Preferably, the loss function of the prediction model is as follows:

[0040]

[0041] In the formula: represents the value of the loss function; N represents the number of samples; y i represents the true value of the i-th sample; represents the predicted value of the i-th sample; θ represents the model parameter vector; M represents the number of model parameters; θ j represents the j-th model parameter; λ represents the weight of the regularization term; w i represents the sample weight, weighted according to feature importance.

[0042] Preferably, step 4 includes the following steps:

[0043] Determine the priorities of device and user types: Divide devices into high-priority, medium-priority, and low-priority based on device type; divide users into high-priority, medium-priority, and low-priority based on user type;

[0044] Calculate the comprehensive priority of each request:

[0045]

[0046] In the formula: P i represents the comprehensive priority of the i-th request; represents the device type priority of the i-th request; represents the user type priority of the i-th request; α and β represent weight coefficients;

[0047] Dynamically allocate IPv6 addresses: Obtain the address requirements of each request i in the future for a period of time according to the prediction model; sort each request in descending order of comprehensive priority; process each request sequentially and allocate address blocks; if the current available address quantity meets the requirements, directly allocate, otherwise allocate according to the priority order.

[0048] Preferably, step 5 includes the following steps:

[0049] Obtain the status of the address pool: Obtain the status of the current IPv6 address pool, including the used or unused situation of each address block; count the number of unused address blocks and used address blocks;

[0050] Analyze the fragmentation degree: Use the fragmentation index F to quantify the fragmentation degree:

[0051]

[0052] Where: F represents the fragmentation index; N fragments represents the number of unused address blocks; N total represents the number of address blocks in the address pool;

[0053] Data preparation: Construct the feature vector x of the address block usage i :

[0054] x i =(s i , d i );

[0055] Where: s i represents the size of the address block; d i represents the usage duration of the address block;

[0056] Clustering process: Apply the k-mean clustering algorithm to cluster the feature vector to obtain the centers and distributions of each cluster:

[0057]

[0058] Where: C represents the cluster assignment vector, indicating the cluster to which each address block belongs. Each element c i corresponds to the cluster number of the i-th address block; μ j represents the center point of the j-th cluster, which is the average value of the feature vectors of all address blocks within the cluster; k represents the number of clusters preset in the cluster analysis;

[0059] Analysis result: Analyze the characteristics of each cluster based on the clustering result, and identify the high-fragmentation area and low-fragmentation area based on a preset threshold;

[0060] Generate a fragmentation analysis report: including the total number of address blocks, the number of used and unused address blocks, the fragmentation index, and the detailed situation of each cluster. The detailed situation of each cluster includes the cluster center, the number of address blocks within the cluster, the high-fragmentation area, and the low-fragmentation area.

[0061] The beneficial effects of the present invention include:

[0062] The present invention ensures the accuracy and consistency of data by obtaining and preprocessing device usage records, address allocation records, and traffic information. Then, it determines the range of the IPv6 address pool according to network planning and divides the address blocks by different regions. Using the preprocessed data, it predicts future address requirements with a prediction model to ensure the forward-looking nature of the allocation plan. It sets allocation priority rules according to device and user types and dynamically allocates IPv6 addresses to ensure that key devices and users obtain resources first. When the demand exceeds the available address quantity, it allocates addresses in the order of priority to avoid resource waste. It regularly counts the status of the address pool, analyzes the fragmentation degree, and identifies and reports fragmentation situations through a clustering algorithm. According to the fragmentation analysis report and a preset threshold, it re-divides the address blocks to reduce fragmentation and optimize address utilization. This method comprehensively applies intelligent prediction, priority allocation, and fragmentation optimization technologies, effectively improving the flexibility and efficiency of IPv6 address allocation and meeting the diverse and dynamic address requirements in the modern network environment. Brief Description of the Drawings

[0063] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0064] Figure 1 It is the overall step block diagram provided by the embodiment of the present invention. Detailed Embodiments

[0065] In order to make the technical problems, technical solutions, and beneficial effects to be solved by the present application clearer, the following further details the present application in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0066] It should be noted that when an element is referred to as "fixed to" or "disposed on" another element, it can be directly on the other element or indirectly on the other element. When an element is referred to as "connected to" another element, it can be directly connected to the other element or indirectly connected to the other element.

[0067] It should be understood that the orientation or positional relationship indicated by terms such as "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation to the present application.

[0068] In addition, the terms "first" and "second" are only used for descriptive purposes and should not be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present application, "a plurality of" means two or more unless otherwise specifically defined.

[0069] See Figure 1 As shown, a further description is made of the optimal embodiment of the present invention;

[0070] An intelligent IPv6 address allocation model and its optimization method, comprising the following steps:

[0071] Step 1: Obtain device usage records, address allocation records, and traffic information, and preprocess the obtained data;

[0072] The preprocessing includes:

[0073] Data cleaning: Remove duplicate records, correct error data, and fill in missing values.

[0074] Data formatting: Uniformly format the data and convert it into a structured data table.

[0075] Data storage: Store the preprocessed data in a database

[0076] Step 2: Determine the scope of the IPv6 address pool according to the network plan, output the available IPv6 address pool scope, and divide the address blocks according to different regions based on the available IPv6 address pool scope;

[0077] Step 3: Based on the preprocessed device usage records, address allocation records, and traffic information, use a prediction model to predict the address demand in the next period of time;

[0078] Step 4: Specify the allocation priority rules according to the device type and user type, and dynamically allocate IPv6 addresses based on the predicted address demand, priority rules, and divided address blocks. If the demand exceeds the current available address quantity, allocate according to the priority order;

[0079] Step 5: Based on the address allocation records and the status of the address pool, regularly count the number of unused and used address blocks, analyze the degree of fragmentation, use clustering algorithms to analyze the usage of address blocks, and obtain a fragmentation analysis report;

[0080] Step 6: Re-divide the address blocks according to the fragmentation analysis report and a preset threshold.

[0081] In the present invention, by obtaining and preprocessing device usage records, address allocation records, and traffic information, the accuracy and consistency of data are ensured. Then, the IPv6 address pool range is determined according to network planning, and address blocks are divided by different regions. Using the preprocessed data, a prediction model is adopted to predict future address requirements, ensuring the forward-looking nature of the allocation scheme. Allocation priority rules are set according to device and user types, and IPv6 addresses are dynamically allocated to ensure that key devices and users obtain resources first. When the demand exceeds the available address quantity, allocation is carried out in the order of priority to avoid resource waste. Regularly count the status of the address pool, analyze the degree of fragmentation, identify and report the fragmentation situation through clustering algorithms. According to the fragmentation analysis report and a preset threshold, re-divide the address blocks to reduce fragmentation and optimize the address utilization rate. This method comprehensively applies intelligent prediction, priority allocation, and fragmentation optimization technologies, effectively improving the flexibility and efficiency of IPv6 address allocation and meeting the diverse and dynamic address requirements in the modern network environment.

[0082] As a possible implementation manner of this embodiment, step 2 includes the following steps:

[0083] Determine the region category: According to business requirements and network planning, divide the network region into a core region, a general region, and an edge region;

[0084] Set the initial address block allocation: Set the size of the initially allocated address blocks for different types of regions, where the size of the address blocks in the core region is larger than that in the general region, and the size of the address blocks in the general region is larger than that in the edge region;

[0085] Dynamic adjustment: Dynamically adjust the address block allocation according to the actual usage situation. If the actual usage quantity is less than the initial allocation quantity, recycle some unused address blocks; if the actual demand in a certain region exceeds the initial allocation quantity, allocate additional address blocks from the remaining address pool.

[0086] In this embodiment, by dividing the network region into a core region, a general region, and an edge region and setting different address block allocation strategies according to actual business requirements, it can be ensured that the core region has sufficient address resources to meet the requirements of high traffic and high priority, while avoiding waste of resources in the general region and the edge region. This regional management method improves the overall resource utilization efficiency and reduces the idle and waste of address resources.

[0087] A method for dynamically adjusting address block allocation can flexibly recycle and reallocate address blocks according to actual usage, avoiding resource waste and resource shortage problems caused by unreasonable initial allocation. For example, when the actual demand in a certain area exceeds the expectation, address blocks can be quickly allocated to ensure network stability and service quality.

[0088] As a possible implementation of this embodiment, the prediction model adopts an LSTM model, and the specific structure is as follows:

[0089] Input layer: used for data input;

[0090] The first LSTM layer: includes 50 LSTM units, and returns the sequence True;

[0091] The second LSTM layer: includes 50 LSTM units, and returns the sequence False;

[0092] Fully connected layer: includes a fully connected unit for outputting a prediction value;

[0093] Output layer: processes the prediction value output by the fully connected layer using a linear activation function and then outputs it.

[0094] As a possible implementation of this embodiment, before the data is input into the prediction model, feature processing is performed on the device usage record, address allocation record, and traffic information, including the following steps:

[0095] Device usage record:

[0096] Extract time features: extract periodic features from the timestamp;

[0097] Device features: extract device ID, device type, and user type features;

[0098] Usage statistical features: extract the number of sessions and session duration;

[0099] Address allocation record:

[0100] Time features: extract periodic features from the timestamp;

[0101] Allocation features: extract the number of addresses allocated each time and the allocation frequency;

[0102] Historical allocation features: extract the moving average and cumulative sum of the allocation quantity in the past period of time;

[0103] Traffic information:

[0104] Time features: extract periodic features from the timestamp;

[0105] Flow characteristics: Extract the flow size and flow type characteristics;

[0106] Historical flow characteristics: The moving average and cumulative sum of the flow size over a past period;

[0107] Normalize all the extracted characteristics; Finally, convert the data format into a three-dimensional array [samples, time_steps, features], where samples represents the number of samples; time_steps represents the number of time steps for each sample; features represents the number of features for each time step.

[0108] As a possible implementation of this embodiment, the loss function of the prediction model is as follows:

[0109]

[0110] In the formula: represents the value of the loss function; N represents the number of samples; y i represents the true value of the i-th sample; represents the predicted value of the i-th sample; θ represents the model parameter vector; M represents the number of model parameters; θ j represents the j-th model parameter; λ represents the weight of the regularization term; w i represents the sample weight, weighted according to the feature importance.

[0111] In this embodiment, the loss function not only considers the error between the predicted value and the true value, but also introduces a regularization term to prevent the model from overfitting. This design can ensure the prediction accuracy while avoiding the model from being too complex, thus improving the generalization ability of the model. By introducing the sample weight and weighting according to the feature importance, personalized optimization can be carried out for the importance of different features, improving the performance of the model in practical applications. For example, assigning higher weights to samples of important features can enhance the sensitivity of the model to key features.

[0112] As a possible implementation of this embodiment, step 4 includes the following steps:

[0113] Determine the priorities of device and user types: Based on the device type, divide the devices into high-priority, medium-priority, and low-priority; Based on the user type, divide the users into high-priority, medium-priority, and low-priority;

[0114] Calculate the comprehensive priority of each request:

[0115]

[0116] In the formula: P i represents the comprehensive priority of the i-th request; Indicates the device type priority of the i-th request; Indicates the user type priority of the i-th request; α and β represent weight coefficients;

[0117] Dynamically allocate IPv6 addresses: Obtain the address requirements of each request i in the future period according to the prediction model; Sort according to the comprehensive priority of each request from high to low; Process each request in order and allocate address blocks; If the current available address quantity meets the requirements, allocate directly, otherwise allocate according to the priority order.

[0118] In this embodiment, the priority is set according to the device type and user type, and the comprehensive priority is calculated to ensure that the needs of high-priority devices and users are preferentially met in the case of limited address resources.

[0119] As a possible implementation manner of this embodiment, step 5 includes the following steps:

[0120] Obtain the address pool status: Obtain the status of the current IPv6 address pool, including the used or unused situation of each address block; Count the number of unused address blocks and used address blocks;

[0121] Analyze the fragmentation degree: Use the fragmentation index F to quantify the fragmentation degree:

[0122]

[0123] In the formula: F represents the fragmentation index; N fregments represents the number of unused address blocks; N total represents the number of address blocks in the address pool;

[0124] Data preparation: Construct the feature vector x of the address block usage situation i :

[0125] x i =(s i , d i );

[0126] In the formula: s i represents the size of the address block; d i represents the usage duration of the address block;

[0127] Clustering process: Apply the k-mean clustering algorithm to cluster the feature vectors to obtain the centers and distributions of each cluster:

[0128]

[0129] In the formula: C represents the cluster assignment vector, indicating the cluster to which each address block belongs, and each element c iThe cluster number corresponding to the i-th address block; μ j Represents the center point of the j-th cluster, which is the average value of the feature vectors of all address blocks within the cluster; k represents the preset number of clusters in the clustering analysis;

[0130] Analysis result: Analyze the characteristics of each cluster based on the clustering result, and identify high-fragmentation regions and low-fragmentation regions based on the preset threshold;

[0131] Generate a fragmentation analysis report: including the total number of address blocks, the number of used and unused address blocks, the fragmentation index, and the detailed information of each cluster. The detailed information of each cluster includes the cluster center, the number of address blocks within the cluster, high-fragmentation regions, and low-fragmentation regions.

[0132] In this embodiment, by regularly counting and analyzing the usage of the address pool to quantify the fragmentation degree, the fragmentation problem of address resources can be discovered and solved in a timely manner, and resource management can be optimized. For example, by identifying high-fragmentation regions through the clustering algorithm, targeted resource integration and reallocation can be carried out to improve the utilization efficiency of the address pool.

[0133] Based on the clustering analysis and the fragmentation analysis report, provide detailed decision support information to help network managers make scientific and reasonable resource allocation decisions. For example, by analyzing the cluster center and the number of address blocks within the cluster in the analysis report, the resource usage in different regions can be deeply understood, and the resource allocation strategy can be optimized.

[0134] As a possible implementation manner of this embodiment, step 6 includes the following steps:

[0135] Obtain the fragmentation analysis report:

[0136] The fragmentation analysis report obtained from step 5 includes: the total number of address blocks, the number of used and unused address blocks, the fragmentation index F, and the detailed information of each cluster (cluster center, number of address blocks within the cluster, high-fragmentation regions, and low-fragmentation regions);

[0137] Set the threshold for re-partitioning: According to the network planning and service quality requirements, set the threshold Fthreshold of the fragmentation index F to determine whether it is necessary to re-partition the address blocks.

[0138] Judge whether re-partitioning is needed: If the fragmentation index F is greater than the set threshold Fthreshold, re-partition the address blocks, otherwise maintain the existing allocation.

[0139] Identify high-fragmentation regions: Based on the cluster characteristics in the fragmentation analysis report, identify high-fragmentation regions. These regions usually have a relatively high number of unused address blocks and a relatively high fragmentation index.

[0140] Aggregate address blocks in highly fragmented areas: Merge unused address blocks in highly fragmented areas to form continuous address blocks. Use the following algorithm:

[0141] Merging algorithm:

[0142] 1. Traverse the unused address blocks in the highly fragmented area.

[0143] 2. Merge adjacent unused address blocks into a larger continuous address block.

[0144] 3. Repeat step 2 until no further merging is possible.

[0145] Reallocate address blocks:

[0146] Based on the merged continuous address blocks, reallocate address blocks to each area (core area, general area, and edge area).

[0147] Dynamically adjust the address block allocation according to the actual usage and demand:

[0148] Adjustment algorithm:

[0149] Count the actual usage of each area.

[0150] If the actual usage of a certain area is less than the initial allocation, reclaim some unused address blocks.

[0151] If the actual demand of a certain area exceeds the initial allocation, allocate additional address blocks from the merged continuous address blocks.

[0152] Optimized address block allocation:

[0153] Ensure that the address block allocation after re-partitioning meets the following conditions:

[0154] The size of the address blocks in the core area is larger than that in the general area.

[0155] The size of the address blocks in the general area is larger than that in the edge area.

[0156] Ensure that the address requirements of each area are met while reducing fragmentation.

[0157] Update the address pool status:

[0158] Update the status of the address pool, including the quantity and location of used and unused address blocks.

[0159] Record the address block allocation information after re-partitioning.

[0160] Monitoring and feedback:

[0161] Continuously monitor the usage and fragmentation degree of the address block.

[0162] Regularly generate a new fragmentation analysis report to evaluate the effect of re-partitioning.

[0163] If the fragmentation index exceeds the threshold Fthreshold again, re-partition the above steps again.

[0164] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. An intelligent IPv6 address allocation model and an optimization method thereof, characterized in that: The following steps are involved: Step 1: Obtain device usage records, address allocation records and traffic information, and pre-process the acquired data; Step 2: According to the network planning, determine the range of the IPv6 address pool, output the available IPv6 address pool range, and divide the address blocks into different areas according to the available IPv6 address pool range; Step 3: Based on the pre-processed equipment usage records, address allocation records and traffic information, a prediction model is used to predict the address demand in the future period; Step 4: Specify allocation priority rules based on device type and user type, and dynamically allocate IPv6 addresses based on predicted address demand, priority rules, and divided address blocks. If the demand exceeds the current available address quantity, allocate them in order of priority. Step 5: Based on the address allocation records and address pool status, regularly count the number of unused and used address blocks, analyze the degree of fragmentation, use a clustering algorithm to analyze the usage of address blocks, and obtain a fragmentation analysis report; Step 6: Re-divide the address blocks according to the fragmentation analysis report and the preset threshold.

2. The intelligent IPv6 address allocation model and optimization method thereof according to claim 1, characterized in that: The step 2 comprises the following steps: Determine the area category: According to business needs and network planning, the network area is divided into core area, general area and edge area; Set initial allocation of address blocks: Set the size of the initially allocated address blocks for different types of areas. The address block size of the core area is larger than that of the common area, and the address block size of the common area is larger than that of the edge area. Dynamic adjustment: Dynamically adjust address block allocation based on actual usage. If the actual usage is less than the initial allocation, some unused address blocks are reclaimed. If the actual demand in a certain area exceeds the initial allocation, additional address blocks are allocated from the remaining address pool.

3. The intelligent IPv6 address allocation model and optimization method thereof according to claim 1, characterized in that: The prediction model adopts the LSTM model, and the specific structure is as follows: Input layer: used for data input; The first LSTM layer: includes 50 LSTM units and returns the sequence True; The second LSTM layer: includes 50 LSTM units and returns the sequence False; Fully connected layer: includes a fully connected unit to output a predicted value; Output layer: The predicted value output by the fully connected layer is processed using a linear activation function and then output.

4. The intelligent IPv6 address allocation model and optimization method thereof according to claim 3 is characterized in that: Before data is input into the prediction model, feature processing is performed on the device usage records, address allocation records, and traffic information, including the following steps: Equipment usage records: Extract time features: extract periodic features from timestamps; Device characteristics: extract device ID, device type, and user type characteristics; Use statistical features: extract the number of sessions and session duration; Address allocation record: Time features: extract periodic features from timestamps; Allocation characteristics: extract the number of addresses allocated each time and the allocation frequency; Historical allocation features: extract the moving average and cumulative sum of the allocation quantity over the past period of time; Traffic information: Time features: extract periodic features from timestamps; Traffic characteristics: extract traffic size and traffic type characteristics; Historical traffic characteristics: moving average and cumulative sum of traffic volume over a period of time; Normalize all the extracted features; finally convert the data format into a three-dimensional array [samples, time_steps, features], where samples represents the number of samples; time_steps represents the number of time steps for each sample; and features represents the number of features for each time step.

5. The intelligent IPv6 address allocation model and optimization method thereof according to claim 3, characterized in that: The loss function of the prediction model is as follows: Where: represents the value of the loss function; N represents the number of samples; y i Represents the true value of the i-th sample; represents the predicted value of the i-th sample; θ represents the model parameter vector; M represents the number of model parameters; θ j represents the jth model parameter; λ represents the weight of the regularization term; w i Represents the sample weight, which is weighted according to the feature importance.

6. The intelligent IPv6 address allocation model and optimization method thereof according to claim 1, characterized in that: The step 4 comprises the following steps: Prioritize devices and user types: Classify devices into high priority, medium priority, and low priority based on device type; classify users into high priority, medium priority, and low priority based on user type; Calculate the combined priority of each request: Where: P i Indicates the comprehensive priority of the i-th request; Indicates the device type priority of the i-th request; represents the user type priority of the i-th request; α and β represent weight coefficients; Dynamic allocation of IPv6 addresses: Obtain the address demand of each request i in the future based on the prediction model; Sort each request from high to low according to its comprehensive priority; Process each request in sequence and allocate address blocks; If the current available address quantity meets the demand, allocate it directly; if not, allocate it according to the priority order.

7. The intelligent IPv6 address allocation model and optimization method thereof according to claim 1, characterized in that: The step 5 comprises the following steps: Get address pool status: Get the status of the current IPv6 address pool, including whether each address block is used or unused; count the number of unused address blocks and used address blocks; Analyze the degree of fragmentation: Use the fragmentation index F to quantify the degree of fragmentation: Where: F represents the fragmentation index; N fragments Indicates the number of unused address blocks; N total Indicates the number of address blocks in the address pool; Data preparation: Constructing feature vector x of address block usage i : x i =(s i ,d i ); Where: s i Indicates the size of the address block; d i Indicates the usage time of the address block; Clustering process: Apply the k-mean clustering algorithm to cluster the feature vectors to obtain the center and distribution of each cluster: Where: C represents the cluster allocation vector, which represents the cluster described by each address block, and each element c i The cluster number corresponding to the i-th address block; μ j represents the center point of the jth cluster, which is the average value of the feature vectors of all address blocks in the cluster; k represents the number of clusters preset in the cluster analysis; Analysis results: Based on the clustering results, the characteristics of each cluster are analyzed, and high-fragmentation areas and low-fragmentation areas are identified based on preset thresholds; Generate a fragmentation analysis report: including the total number of address blocks, the number of used and unused address blocks, fragmentation indicators, and details of each cluster, wherein the details of each cluster include the cluster center, the number of address blocks in the cluster, high fragmentation areas, and low fragmentation areas.

Images