A method and system for cross-network element multicast

By deploying multicast functionality software and custom protocols, a topology map is constructed, and a spectral clustering algorithm is used to divide the network into sub-regions. Roles are dynamically assigned, and multicast data transmission is optimized. This solves the problem of insufficient resource status awareness in traditional multicast protocols and achieves more efficient and secure multicast transmission.

CN120768707BActive Publication Date: 2026-11-17深圳市飞思通信技术有限公司
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202510900600.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-07-01
Publication Date
2026-11-17
Estimated Expiration
2045-07-01

AI Technical Summary

Technical Problem

Traditional multicast protocols lack the ability to perceive the status of network element resources in real time, which leads to the failure to fully consider factors such as link bandwidth and node load when selecting forwarding paths. This can easily cause local congestion and resource idleness. Moreover, existing methods mostly adopt static methods in role allocation and topology division, which are difficult to adapt to dynamic changes in network status and affect overall transmission efficiency.

Method used

Deploy multicast functionality software, use a custom protocol to query the resource status of each network element, construct a topology map, define a partitioning objective function, use a spectral clustering algorithm to divide the network into sub-regions, define a convex quadratic optimization cost function for dynamic role allocation, generate unicast forwarding instructions, encapsulate and synchronously encrypt multicast data packets, check data integrity, analyze multicast operation performance indicators, generate a multicast group status table, and construct a visualization interface.

Benefits of technology

It improves the accuracy of network information acquisition, enhances the flexibility of role allocation and the security of data transmission, and strengthens the system's dynamic response capability and overall transmission efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120768707B_ABST
    Figure CN120768707B_ABST
Patent Text Reader

Abstract

This invention discloses a cross-network element multicast method and system, relating to the field of network communication technology. The method includes: deploying multicast function software; using a custom protocol to query the resource status of each network element; constructing a topology map; defining a partitioning objective function; using a spectral clustering algorithm to divide the network into sub-regions; defining a convex quadratic optimization cost function; performing dynamic role allocation; defining constraints; using a corrected search direction to solve the problem; generating unicast forwarding instructions; encapsulating multicast data packets; selecting the path with the minimum cost ratio; performing data transmission and synchronization encryption; and checking data integrity. This invention improves the rationality of the topology map by partitioning the network based on a spectral clustering algorithm and constructing sub-regions that conform to the current network state. It also enhances the robustness of multicast services under sudden traffic surges and network failure scenarios by defining a convex quadratic optimization cost function, introducing constraints for dynamic role allocation, and using a corrected search direction to solve the problem.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of network communication technology, and in particular to a method and system for cross-network element multicast. Background Technology

[0002] With the rapid development of network communication technology, multicast technology has become a research hotspot because it can effectively reduce network load and improve transmission efficiency. With the development of new network architectures such as software-defined networking and network function virtualization, multicast technology has also gradually evolved from static configuration to dynamic scheduling, supporting more flexible member management and resource allocation strategies. Currently, mainstream multicast protocols such as PIM-SM and IGMP Snooping have been widely deployed in IP networks.

[0003] Existing technologies have initially achieved multicast functionality, but still have many shortcomings. Traditional multicast protocols lack the ability to perceive the status of network element resources in real time, which leads to the failure to fully consider factors such as link bandwidth and node load when selecting forwarding paths, which can easily cause local congestion and resource idleness. Secondly, existing methods mostly adopt static methods in role allocation and topology division, which are difficult to adapt to dynamic changes in network status and affect overall transmission efficiency. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a cross-network element multicast method and system to solve the problem that traditional multicast protocols lack the ability to perceive the real-time status of network element resources, resulting in the failure to fully consider factors such as link bandwidth and node load when selecting forwarding paths, which easily leads to local congestion and resource idleness. Secondly, existing methods mostly adopt static methods in role allocation and topology division, which are difficult to adapt to dynamic changes in network status and affect the overall transmission efficiency.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a cross-network element multicast method, which includes,

[0008] Deploy multicast software, use a custom protocol to query the resource status of each network element, construct a topology map, define a partitioning objective function, and use a spectral clustering algorithm to divide the network into sub-regions;

[0009] Define a convex quadratic optimization cost function, perform dynamic role allocation, define constraints, use a modified search direction to solve the problem, and generate unicast forwarding instructions;

[0010] The multicast data packets are encapsulated, the path with the lowest cost ratio is selected, data transmission and synchronization are encrypted, and data integrity is checked.

[0011] Analyze the performance metrics of multicast operation and optimize multicast operation;

[0012] Audit the status of multicast groups, generate a multicast group status table, build a visual interface to display and store the multicast operation status data.

[0013] As a preferred embodiment of the cross-network element multicast method described in this invention, the deployment of multicast function software includes: using a custom protocol to query the resource status of each network element, constructing a topology map, defining a partitioning objective function, and using a spectral clustering algorithm to divide the network into sub-regions, including:

[0014] Install dedicated multicast management software on the server and lightweight multicast client software on the client. The software communicates via the TCP / IP protocol stack.

[0015] The server uses a custom protocol to query the resource status of each network element, calculates the comprehensive resource score, and establishes a resource database with each network element ID, available bandwidth, multicast group address, computing power, and comprehensive resource score data.

[0016] Using SNMP to scan the network topology, network elements are treated as nodes, links as edges, and the delay and bandwidth between nodes are used as edge weights to obtain a topology graph G = (V, E, W). The standard deviations of the delay, bandwidth, and comprehensive resource score between nodes are calculated, and the similarity of each pair of nodes is calculated. The network is divided into K sub-regions using a spectral clustering algorithm. Based on the similarity of each pair of nodes, the degree matrix is ​​calculated, and the Laplace matrix is ​​calculated using the phase difference method. The eigenvectors corresponding to the first K smallest non-zero eigenvalues ​​of the Laplace matrix are extracted to form a feature matrix.

[0017] Initialize the clustering assignment matrix H, define the partitioning objective function, treat H as a continuous variable, and optimize using gradient descent to obtain the optimal sub-region partitioning with the goal of minimizing the partitioning objective function.

[0018] As a preferred embodiment of the cross-network element multicast method described in this invention, the following steps are included: defining a convex quadratic optimization cost function, performing dynamic role allocation, defining constraints, solving using a corrected search direction, and generating unicast forwarding instructions:

[0019] Based on the topology graph G, define decision variables, initialize cost function parameters, calculate dynamic factors, calculate the score of each node, sort the scores of each node in descending order, and select candidate forwarding nodes.

[0020] Construct a cost function, construct a flow conservation constraint allocation matrix and a bandwidth constraint allocation matrix, traverse each node of the topology graph G, identify outgoing and incoming links, define flow conservation and bandwidth constraints, and fill in the flow conservation constraint allocation matrix and the bandwidth constraint allocation matrix.

[0021] Based on dynamic factors, adaptive obstacle parameters are calculated;

[0022] Based on the cost function and bandwidth constraints, the initial decision variables are set as a neutral starting point. The corrected search direction is calculated, the decision variables are updated, and the solution is iterated. After each iteration, the bandwidth constraints are verified. If the constraints are not met, the obstacle update parameters are halved and the solution is recalculated. Otherwise, the iteration continues until the maximum number of iterations is reached. Then, the iteration stops, and the optimal decision variables are output as a multicast tree.

[0023] The server traverses the multicast tree, extracts the path, and outputs a unicast forwarding instruction.

[0024] As a preferred embodiment of the cross-network element multicast method of the present invention, the steps of encapsulating multicast data packets, selecting the path with the lowest cost ratio, performing data transmission and synchronization encryption, and checking data integrity include:

[0025] The server receives business data from the data source and encapsulates the business data, multicast group address, priority label, timestamp, and sequence number into a multicast data packet;

[0026] Traverse the multicast tree, generate a path metadata table, use SNMP to query device status, obtain resource information of intermediate devices and links, and filter unusable paths;

[0027] For each multicast data packet, select the optimal transmission path, prioritizing the needs of VIP users. Calculate the consumption ratio of each path and select the path with the lowest consumption ratio as the optimal path.

[0028] Check the network elements in the optimal path, use high-bandwidth links for network elements within an optimal sub-region, and use low-latency links across regions;

[0029] The server uses IPsec to encrypt the service data in multicast packets, generates encrypted packets, and sends the encrypted packets to multicast group members using the optimal path.

[0030] Based on the priority label of multicast packets, the DSCP field in the packet header is set for VIP and ordinary users. The server sends QoS configuration instructions to each intermediate device through a custom protocol and configures queue parameters on the intermediate devices.

[0031] The intermediate device receives encrypted data packets and unicast forwarding instructions sent by the server, parses the forwarding instructions, extracts the next-hop IP address from the optimal path, and forwards the data packets. The last-hop device forwards the data packets to the client.

[0032] The client uses IPsec to decrypt data packets, obtains business data, sequence number, timestamp and priority label, verifies the continuity of sequence number, marks it as a potential packet loss if the sequence number is not continuous, records the sequence number difference, and sends a packet loss notification to the server through a custom protocol.

[0033] The decrypted business data is processed using the CRC-32 algorithm to generate a checksum, which is then sent back to the server. This checksum is compared with the checksum sent by the server. If they match, the transmission is successful; otherwise, the packet is lost.

[0034] Data marked as packet loss or for which packet loss notifications were received will be replayed;

[0035] The difference between the client's received timestamp and the server's sent timestamp is calculated using the phase difference method as the feedback delay.

[0036] As a preferred embodiment of the cross-network element multicast method of the present invention, the step of analyzing the performance indicators of multicast operation and optimizing multicast operation includes:

[0037] Performance metrics, including latency, packet loss rate, and bandwidth utilization, are collected and estimated based on historical data through regression. After normalization, a comprehensive network performance score is calculated. Thresholds for the comprehensive performance scores of VIP and regular users are set separately. If the comprehensive performance score of a path exceeds the corresponding threshold, the path is marked as a "path that needs optimization" and an optimization process is initiated. If VIP users are involved, optimization is triggered immediately, while regular users wait in a queue.

[0038] As a preferred embodiment of the cross-network element multicast method of the present invention, wherein: the auditing of multicast group status and the generation of a multicast group status table include:

[0039] In each sub-region, the server calculates the node score, selects the network element with the highest node score as the audit node, generates an audit task package from each client subset, RSA public key, and heartbeat message template in the sub-region, and sends the task package to the audit node using the gRPC protocol;

[0040] The audit node sends heartbeat messages to the assigned subset of clients through a custom protocol, calculates an activity score, sets an activity score threshold based on historical regression estimation, and filters out clients with activity scores less than the activity score threshold as offline clients, otherwise they are considered active clients;

[0041] The audit node sends authentication requests to active clients, requesting the return of an RSA signature. Active clients that do not return an RSA signature are marked as illegitimate clients.

[0042] The audit node reports offline and unauthorized clients to the server via gRPC. The server removes the offline and unauthorized clients, records the reason for removal and timestamp, adds the IP of the unauthorized client to the server and intermediate devices to prevent subsequent joining requests, and generates a multicast group status table.

[0043] As a preferred embodiment of the cross-network element multicast method of the present invention, the step of constructing a visual interface to display and store multicast operation status data includes:

[0044] The Grafana framework is used to generate interactive dashboards that display network topology diagrams, performance curve line graphs, and multicast group status tables. InfluxDB is used to store the network topology, performance curve line graphs, and multicast group status tables by timestamp.

[0045] Secondly, the present invention provides a cross-network element multicast system, comprising,

[0046] The deployment and partitioning module is used to deploy multicast function software, use a custom protocol to query the resource status of each network element, construct a topology map, define the partitioning objective function, and use a spectral clustering algorithm to partition the network into sub-regions;

[0047] Define the solution module, which is used to define the convex quadratic optimization cost function, perform dynamic role allocation, define constraints, use the corrected search direction to solve the problem, and generate unicast forwarding instructions;

[0048] The encapsulation and transmission module is used to encapsulate multicast data packets, select the path with the lowest cost ratio, perform data transmission and synchronization encryption, and check data integrity.

[0049] The analysis and optimization module is used to analyze the performance metrics of multicast operation and optimize multicast operation.

[0050] The audit module is used to audit the status of multicast groups and generate a multicast group status table.

[0051] The visualization and storage module is used to build a visual interface to display and store multicast operation status data.

[0052] Thirdly, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein the computer program, when executed by the processor, implements any step of the cross-network element multicast method as described in the first aspect of the present invention.

[0053] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the cross-network element multicast method as described in the first aspect of the present invention.

[0054] The beneficial effects of this invention are as follows: By deploying multicast function software, using a custom protocol to query the resource status of each network element, constructing a topology map, defining a partitioning objective function, and using a spectral clustering algorithm to divide the network into sub-regions; defining a convex quadratic optimization cost function, performing dynamic role allocation, defining constraints, using a corrected search direction for solution, and generating unicast forwarding instructions; encapsulating multicast data packets, selecting the path with the minimum cost ratio, performing data transmission and synchronization encryption, and checking data integrity; analyzing multicast operation performance indicators and optimizing multicast operation; auditing multicast group status, generating a multicast group status table, constructing a visual interface to display and store multicast operation status data, improving the accuracy of network information acquisition, increasing the flexibility of role allocation, and enhancing the security of data transmission. Attached Figure Description

[0055] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0056] Figure 1 This is a flowchart of the cross-network element multicast method in Example 1.

[0057] Figure 2 This is a schematic diagram of the cross-network element multicast system in Example 1.

[0058] Figure 3 This is a schematic diagram of the audit multicast group status in Example 1.

[0059] Figure 4 This is a flowchart of the multicast data transmission and verification process in Example 1. Detailed Implementation

[0060] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0061] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0062] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0063] Example 1, referring to Figures 1 to 4 This is the first embodiment of the present invention, which provides a cross-network element multicast method, including the following steps:

[0064] S1. Deploy multicast function software, use a custom protocol to query the resource status of each network element, construct a topology map, define a partitioning objective function, and use a spectral clustering algorithm to divide the network into sub-regions;

[0065] Preferably, dedicated multicast management software is installed on the server (multicast server) to be responsible for topology management, resource allocation and multicast group maintenance;

[0066] Install lightweight multicast client software on the client side (terminal device, such as 5G UE, IoT sensor) to support custom protocol interaction;

[0067] The software communicates via the TCP / IP protocol stack, and the message format is a custom protocol (including identifier, priority, and timestamp), ensuring that intermediate devices do not need to support multicast protocols.

[0068] The server uses a custom protocol to query the resource status of each network element and calculates the comprehensive resource score. The formula is:

[0069]

[0070] Among them, R i The comprehensive resource score for the i-th network element. The total bandwidth of the i-th network element is σ represents the used bandwidth of the i-th network element. B The standard deviation of the available bandwidth for all network elements. The used CPU resources of the i-th network element. The total CPU resources of the i-th network element. Let be the available bandwidth of the i-th network element. The computing power of the i-th network element;

[0071] A resource database is established using each network element ID, available bandwidth, multicast group address, computing power, and comprehensive resource score data.

[0072] Using SNMP to scan the network topology, network elements are treated as nodes, links as edges, and the latency and bandwidth between nodes are used as edge weights, resulting in a topology graph G = (V, E, W), where V is the set of network elements, E is the set of links, and W = {d ij ,b ij Let )|(i,j)∈E} be the set of link weights, and d ij Let b be the one-way delay from node i to node j. ij The available bandwidth of the link between node i and node j;

[0073] Based on the topology graph G, calculate the standard deviation of latency, bandwidth, and overall resource score between nodes, and calculate the similarity of each pair of nodes using the following formula:

[0074]

[0075] Among them, S ij For the similarity between node i and node j, σ d σ b and σ R d respectively ij b ij and R i +R j Standard deviation;

[0076] Based on the similarity of each pair of nodes, a spectral clustering algorithm is used to divide the network and generate K sub-regions.

[0077] Based on the similarity of each pair of nodes, the degree matrix is ​​calculated, the Laplacian matrix is ​​calculated using the phase difference method, and the eigenvectors corresponding to the first K smallest non-zero eigenvalues ​​of the Laplacian matrix are extracted to form the feature matrix.

[0078] Initialize the clustering assignment matrix H (representing the probability of each node being assigned to a region in the matrix), where H has the same number of rows and columns as the feature matrix (a node in the feature matrix is ​​marked as 1 if it is assigned to a region in the matrix, and 0 otherwise). Define the partitioning objective function, as follows:

[0079]

[0080] Where P is the partition objective function value, k is the sub-region index, and all sub-regions are traversed. For the set of subregions, D k σ is the average delay of subregion k. D B is the delayed standard deviation. k R is the minimum bandwidth of subregion k. k The average comprehensive resource score for sub-region k;

[0081] Treating H as a continuous variable, and aiming to minimize the partitioning objective function, we use gradient descent to optimize and obtain the optimal sub-region partitioning.

[0082] By using a custom protocol to query network element resource status and calculate a comprehensive resource score, a standardized scoring mechanism is introduced, improving the scientific nature and fairness of resource scheduling. This helps to make more reasonable path selection and role allocation decisions when network load changes. By calculating the similarity between nodes, a spectral clustering algorithm is introduced for network partitioning, which helps to reduce the performance loss caused by cross-region forwarding and improve overall transmission efficiency. By defining a partitioning objective function and using gradient descent to optimize sub-region partitioning, the objective function comprehensively considers latency fluctuations, minimum bandwidth bottlenecks, and the average resource score within the sub-region, achieving a balance between performance, stability, and resource utilization in the partitioning results. The use of gradient descent to solve the problem ensures the convergence and computational efficiency of the algorithm, improving the system's processing capacity in large-scale networks.

[0083] S2. Define a convex quadratic optimization cost function, perform dynamic role allocation, define constraints, use a modified search direction to solve the problem, and generate unicast forwarding instructions.

[0084] Preferably, based on the network topology G, a decision variable is defined to indicate whether a link is used for a multicast tree (1 for use, 0 for no use);

[0085] Initialize the cost function parameters, including setting the bandwidth penalty coefficient α and the load balancing coefficient β using an experimental method;

[0086] Calculate the dynamic factor using the following formula:

[0087]

[0088] Where δ is the dynamic factor, reflecting the frequency of member changes, Δ|R t | represents the number of member changes (joins / leaves) at time t, retrieved from the Redis record, T. w The time window is provided by the server-side timer, where ΔT is the total duration of the window and |R| is the current number of receivers.

[0089] Based on the dynamic factor, the score for each node is calculated using the following formula:

[0090]

[0091] Among them, score i For the score of node i, w i Let B be the node priority weight (1 for VIP-related nodes, 0.5 for ordinary nodes), derived from the member table. i C i and D iThese represent node bandwidth, computing power, and average latency, respectively. max C max and D max For B v C v and D v The maximum value;

[0092] The scores of each node are sorted in descending order, and the top 50% of the node scores are selected as candidate forwarding nodes.

[0093] Construct the cost function, formula:

[0094]

[0095] Where R is the set of client nodes, P i Let E be the set of links from the server to node i, e be the index of the link, E′ be the set of links connected to candidate forwarding nodes, and D be the set of links connected to the server to node i. e B e and f i These are latency, bandwidth capacity, and traffic requirements, respectively. e As decision variables, To normalize the weighted delay, VIP users are given priority. To normalize bandwidth consumption and reduce resource usage, To optimize load balancing, the load ratio is calculated using the squared load factor.

[0096] Initialize the sparse matrix, construct the flow conservation constraint assignment matrix with dimension |V|×|E|, and construct the bandwidth constraint assignment matrix with dimension |E|×|E|.

[0097] Traverse each node of the topology graph G, identify outgoing and incoming links, and define flow conservation and bandwidth constraints, formulas:

[0098]

[0099] Where out(i) and in(i) are the sets of links originating from and entering from node i, s is the server node index, and V′ is the candidate forwarding node index;

[0100] Formula meaning: The server s outputs |R| streams, each stream reaches a receiver i∈R, non-candidate nodes do not participate in forwarding, and forwarding nodes i∈V′ maintain stream balance;

[0101] In the flow conservation constraint allocation matrix, for each node, if the link e∈out(i), the corresponding element is set to +1, and if e∈in(v), the corresponding element is set to -1. The right-hand vector of the flow conservation constraint allocation matrix is ​​set as |R| for the server, -1 for the receiver, 0 for the candidate forwarding node, and 0 for the others.

[0102] Traverse the links and fill the bandwidth constraints, formula:

[0103]

[0104] Among them, f i For the traffic requirements of client i;

[0105] In the bandwidth constraint allocation matrix, for each e∈E, the corresponding row element is set to f. i Set the right-hand vector of the bandwidth constraint allocation matrix to B. e ;

[0106] Based on the dynamic factor, the adaptive barrier parameter is calculated using the following formula:

[0107]

[0108] Where θ is the obstacle update parameter, controlling the iteration step size, and ε is the upper limit of θ, a fixed value set using an empirical method. Network size factor;

[0109] Based on the cost function and bandwidth constraints, the initial decision variable is set as a neutral starting point, and the corrected search direction is calculated using the following formula:

[0110]

[0111] Where, Δx e To correct the search direction, adjust x e The iteration steps;

[0112] The decision variables are updated using the following formula:

[0113] η=min(1,θ / ||Δx e ||),

[0114] x′ e ←x e +ηΔx e ,

[0115] Where η is the update step size, x′ e For the updated decision variables;

[0116] The solution is obtained through iterative steps. After each iteration, the bandwidth constraint is verified. If the constraint is not met, the obstacle update parameter is halved and the solution is recalculated. Otherwise, the iteration continues until the maximum number of iterations is reached. Then, the iteration stops and the optimal decision variable is output as a multicast tree.

[0117] The server traverses the multicast tree, extracts the path, and outputs a unicast forwarding command (supporting transparent forwarding).

[0118] By defining 0-1 decision variables to represent whether a link is selected into a multicast tree path, a binary state model of available links in the network topology is achieved, thus transforming the multicast routing problem into a mathematically optimizable form. This provides the basic structural support for the subsequent construction of cost functions and constraints. By using experimental methods to set key parameters in the cost function, quantitative control of resource consumption and load distribution during multicast path selection is achieved, ensuring the applicability and adaptability of the optimization objective in real-world network environments. By calculating dynamic factors based on the number of member changes, time window length, and total number of receivers, real-time evaluation of the frequency of network member changes is achieved, reflecting the activity level and uncertainty of multicast sessions. Furthermore, by constructing a cost function that includes normalized weighted delay, bandwidth consumption, and squared load ratio... This function achieves joint optimization of multicast paths across multiple performance dimensions, balancing service quality and resource utilization. By constructing a sparse matrix representation of flow conservation and bandwidth constraints, it transforms the multicast routing problem into an integer programming problem with linear constraints, ensuring that path selection conforms to network physical limitations. By calculating adaptive obstacle parameters based on dynamic factors and network size factors, it achieves dynamic adjustment of convergence speed and accuracy during iterative solution, improving solution efficiency. By calculating and correcting the search direction and updating decision variables by step size, it achieves progressive approximation optimization of multicast paths, continuously approaching the global optimum while satisfying constraints. This improves the accuracy and efficiency of path generation, enabling the multicast tree to adapt to network state changes more quickly and enhancing the system's dynamic response capability.

[0119] S3. Encapsulate multicast data packets, select the path with the lowest cost ratio, perform data transmission and synchronization encryption, and check data integrity.

[0120] Preferably, the server receives business data from the data source and encapsulates the business data, multicast group address (obtained from the resource database), priority label, timestamp, and sequence number into a multicast data packet;

[0121] The business data refers to real-time data from external devices connected to the server and internally generated data (such as file data provided by the enterprise);

[0122] Traverse the multicast tree to generate a path metadata table, including the start point, end point, intermediate devices, and link identifiers for each path. Use SNMP to query device status, obtain resource information for intermediate devices and links, and filter unusable paths (such as disconnected links and offline devices).

[0123] The resource information includes the path's available bandwidth, energy consumption, end-to-end latency, and throughput.

[0124] For each multicast data packet, select the optimal transmission path, prioritizing the needs of VIP users, and calculate the cost ratio C for each path. c,formula:

[0125]

[0126] Among them, D c B c E c and Y c These represent the end-to-end delay, available bandwidth, and E for path c, respectively. c and throughput, For transmission efficiency, throughput is divided by latency, which reflects path performance;

[0127] Extract the consumption ratio of all available paths from the path database, and select the path with the lowest consumption ratio as the optimal path;

[0128] Check the network elements in the optimal path, use high-bandwidth links for network elements within an optimal sub-region, and use low-latency links across regions;

[0129] The server uses IPsec (ESP mode, AES-256 encryption algorithm) to encrypt the service data of multicast packets, retains the packet header (multicast group address, priority label, sequence number, timestamp), encrypts only the service data, generates encrypted packets, and sends the encrypted packets to multicast group members using the optimal path;

[0130] Based on the priority label of multicast packets, set the DSCP field in the packet header for VIP and ordinary users. Set the DSCP value to EF for VIP users and BE for ordinary users.

[0131] The server sends QoS configuration instructions to each intermediate device through a custom protocol. The instructions include EF and BE.

[0132] Configure queue parameters on intermediate devices. EF uses strict priority scheduling to ensure that it is processed before BE. BE uses a weighted fair queue to allocate the remaining bandwidth.

[0133] The intermediate device receives encrypted data packets and unicast forwarding instructions sent by the server, parses the forwarding instructions, extracts the next-hop IP address from the optimal path, and forwards the data packets. The last-hop device forwards the data packets to the client (multicast group member).

[0134] The client uses IPsec to decrypt data packets, obtains business data, sequence number, timestamp and priority label, verifies the continuity of sequence number, marks it as a potential packet loss if the sequence number is not continuous, records the sequence number difference (indicating the number of lost data packets), and sends a packet loss notification to the server through a custom protocol.

[0135] The decrypted business data is processed using the CRC-32 algorithm to generate a checksum, which is then sent back to the server. This checksum is compared with the checksum sent by the server. If they match, the transmission is successful; otherwise, the packet is lost.

[0136] Data marked as packet loss or for which packet loss notifications were received will be replayed;

[0137] The difference between the client's received timestamp and the server's sent timestamp is calculated using the phase difference method as the feedback delay.

[0138] The server acquires real-time business data from external devices or internally generated data sources and encapsulates it into multicast data packets along with multicast group addresses, priority tags, timestamps, and sequence numbers. This achieves structured organization and metadata appending of the raw data. By traversing the multicast forwarding tree structure, a path metadata table containing the start and end points, intermediate devices, and link identifiers of each path is generated. The SNMP protocol is used to query the status of each intermediate device to obtain resource information including available bandwidth, energy consumption, latency, and throughput. Unavailable paths are filtered, and a dynamically updated network topology view and resource status database are established to eliminate faulty paths and ensure path availability. By calculating the path consumption ratio, the optimal path is selected. A quantitative evaluation model is introduced to comprehensively consider performance and energy efficiency factors to achieve intelligent path decision-making. By setting the DSCP field based on priority tags and configuring QoS parameters for intermediate devices, differentiated quality of service control is implemented at the network layer, prioritizing the experience of high-value users. By parsing forwarding instructions and executing unicast forwarding through intermediate devices, multicast data is forwarded in a "unicast path concatenation" manner, enhancing forwarding flexibility and path control capabilities. By replaying lost packets, a closed-loop feedback mechanism is established to ensure the integrity of data transmission.

[0139] S4. Analyze the performance indicators of multicast operation and optimize multicast operation;

[0140] Preferably, performance metrics, including latency, packet loss rate, and bandwidth utilization, are collected and estimated based on historical data through regression. These metrics are then normalized, and the overall network performance score P′ is calculated using the formula:

[0141]

[0142] Where SD, SL and SB are the normalized values ​​of feedback delay, packet loss rate and bandwidth utilization, respectively.

[0143] Set comprehensive performance score thresholds for VIP and regular users respectively. If the comprehensive performance score of a path is greater than the corresponding comprehensive performance score threshold, mark the path as a "path that needs optimization" and initiate the optimization process. If VIP users are involved, the optimization will be triggered immediately, while regular users will wait in the queue.

[0144] The optimization process includes re-running convex optimization on paths marked as "needing optimization", generating a multicast tree, selecting a new path using the minimum cost ratio, and encapsulating the new path into a unicast forwarding instruction, which is then resent to the intermediate device via a custom protocol.

[0145] By continuously collecting end-to-end performance parameters during multicast transmission, data support is provided for subsequent intelligent decision-making. By establishing a regression model based on historical performance data for estimation and normalizing each indicator, standardized comparison between indicators of different dimensions is achieved. Through a set comprehensive performance score function, the three normalized indicators are weighted and fused to generate a numerical indicator reflecting the overall path quality, realizing a quantitative evaluation of the quality of multicast paths. By setting comprehensive performance score thresholds for VIP users and ordinary users respectively, marked paths are classified and managed, realizing differentiated service level control. By re-running the convex optimization algorithm on the paths that need optimization, a new multicast forwarding tree structure is generated, and the path with the minimum consumption ratio is selected, realizing dynamic reconstruction and resource reallocation of forwarding paths. By encapsulating the optimized path information into unicast forwarding instructions and sending them to intermediate devices with the help of a custom protocol, efficient collaboration between the control plane and the data plane is achieved.

[0146] S5. Audit the multicast group status, generate a multicast group status table, build a visual interface to display and store the multicast operation status data;

[0147] Preferably, the server calculates the node score within each sub-region using the following formula:

[0148]

[0149] Among them, SA l The score for node l within the sub-region. The CPU utilization of node l. The network element bandwidth utilization rate of node l;

[0150] Select the network element with the highest node score as the audit node, generate an audit task package from each client subset (ID, IP, and priority), RSA public key, and heartbeat message template (custom protocol, including timestamp and sequence number) of the sub-region, and send the task package to the audit node using the gRPC protocol (prioritizing distribution to sub-regions containing VIP users).

[0151] The audit node sends heartbeat messages to a subset of assigned clients via a custom protocol and calculates the activity score AC using the formula:

[0152]

[0153] Among them, T resT represents the client response time. se T is the heartbeat sending time. max The maximum allowed response time is set.

[0154] Based on historical regression estimation, an activity score threshold is set, and clients with AC less than the activity score threshold are considered offline clients; otherwise, they are considered active clients.

[0155] The audit node sends authentication requests to active clients, requesting the return of an RSA signature. Active clients that do not return an RSA signature are marked as illegitimate clients.

[0156] The audit node reports offline and unauthorized clients to the server via gRPC. The server removes the offline and unauthorized clients, records the reason for removal (offline or unauthorized) and timestamp, adds the IP of the unauthorized client to the server and intermediate devices to prevent subsequent joining requests, and generates a multicast group status table.

[0157] By dynamically calculating node scores based on CPU utilization and bandwidth utilization within each sub-region, a quantitative assessment of the resource load status of each node within the sub-region is achieved. By selecting the node with the highest score as the audit node and encapsulating client subset information, RSA public keys, and heartbeat templates into audit task packages, targeted distribution and parameter pre-setting of audit tasks are achieved. By using the gRPC protocol for task package transmission and prioritizing task distribution to sub-regions containing VIP users, rapid response and key protection for high-priority areas are achieved. By periodically sending heartbeat packets through a custom protocol and calculating activity scores based on response time differences, dynamic perception of client online status is achieved. By building a regression model based on historical activity score data and dynamically setting activity thresholds, adaptive adjustment of client status judgment standards is achieved. By initiating RSA signature verification requests to active clients, secondary confirmation of client identity is achieved, preventing spoofing attacks and unauthorized access. By summarizing the status of all clients and generating a multicast group status table, a comprehensive visualization of the current multicast group member composition is achieved.

[0158] Furthermore, the Grafana framework is used to generate interactive dashboards that display network topology diagrams, performance curve line graphs, and multicast group status tables. InfluxDB is then used to store the network topology, performance curve line graphs, and multicast group status tables by timestamp.

[0159] Interactive dashboards lower the barrier to operation and maintenance and improve response efficiency. Multicast group status tables enable centralized management of member status. InfluxDB’s time-series storage mechanism establishes a complete operational trajectory database.

[0160] This embodiment also provides a cross-network element multicast system, including:

[0161] The deployment and partitioning module is used to deploy multicast function software, use a custom protocol to query the resource status of each network element, construct a topology map, define the partitioning objective function, and use a spectral clustering algorithm to partition the network into sub-regions;

[0162] Define the solution module, which is used to define the convex quadratic optimization cost function, perform dynamic role allocation, define constraints, use the corrected search direction to solve the problem, and generate unicast forwarding instructions;

[0163] The encapsulation and transmission module is used to encapsulate multicast data packets, select the path with the lowest cost ratio, perform data transmission and synchronization encryption, and check data integrity.

[0164] The analysis and optimization module is used to analyze the performance metrics of multicast operation and optimize multicast operation.

[0165] The audit module is used to audit the status of multicast groups and generate a multicast group status table.

[0166] The visualization and storage module is used to build a visual interface to display and store multicast operation status data.

[0167] This embodiment also provides a computer device applicable to the cross-network element multicast method, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the cross-network element multicast method proposed in the above embodiment.

[0168] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0169] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the cross-network element multicast method proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0170] In summary, this invention improves the accuracy of network information acquisition, enhances the flexibility of role allocation, and strengthens data transmission security by: deploying multicast software, using a custom protocol to query the resource status of each network element, constructing a topology map, defining a partitioning objective function, and using a spectral clustering algorithm to divide the network into sub-regions; defining a convex quadratic optimization cost function, performing dynamic role allocation, defining constraints, solving the problem using a corrected search direction, and generating unicast forwarding instructions; encapsulating multicast data packets, selecting the path with the minimum cost ratio, performing data transmission and synchronization encryption, and checking data integrity; analyzing multicast operation performance indicators and optimizing multicast operation; auditing multicast group status, generating a multicast group status table, constructing a visual interface to display and store multicast operation status data, and so on.

[0171] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A cross-network element multicast method, characterized in that: include, Deploy multicast software, use a custom protocol to query the resource status of each network element, construct a topology graph G, define a partitioning objective function, and use a spectral clustering algorithm to divide the network into sub-regions; Define a convex quadratic optimization cost function, perform dynamic role allocation, define constraints, use a modified search direction to solve the problem, and generate unicast forwarding instructions; The multicast data packets are encapsulated, the path with the lowest cost ratio is selected, data transmission and synchronization are encrypted, and data integrity is checked. Analyze the performance metrics of multicast operation and optimize multicast operation; Audit the status of multicast groups, generate a multicast group status table, build a visual interface to display and store multicast operation status data; The process includes defining a convex quadratic optimization cost function, performing dynamic role allocation, defining constraints, solving using a corrected search direction, and generating unicast forwarding instructions, including: Based on the topology graph G, define decision variables, initialize cost function parameters, calculate dynamic factors, calculate the score of each node, sort the scores of each node in descending order, and select candidate forwarding nodes. Construct a cost function, construct a flow conservation constraint allocation matrix and a bandwidth constraint allocation matrix, traverse each node of the topology graph G, identify outgoing and incoming links, define flow conservation and bandwidth constraints, and fill in the flow conservation constraint allocation matrix and the bandwidth constraint allocation matrix. Based on dynamic factors, adaptive obstacle parameters are calculated; Based on the cost function and bandwidth constraints, the initial decision variables are set as a neutral starting point. The corrected search direction is calculated, the decision variables are updated, and the solution is iterated. After each iteration, the bandwidth constraints are verified. If the constraints are not met, the obstacle update parameters are halved and the solution is recalculated. Otherwise, the iteration continues until the maximum number of iterations is reached. Then, the iteration stops, and the optimal decision variables are output as a multicast tree. The server traverses the multicast tree, extracts the path, and outputs a unicast forwarding instruction.

2. The cross-network element multicast method of claim 1, wherein: The multicast deployment software uses a custom protocol to query the resource status of each network element, constructs a topology graph G, defines a partitioning objective function, and uses a spectral clustering algorithm to divide the network into sub-regions, including: Install dedicated multicast management software on the server and lightweight multicast client software on the client. The software communicates via the TCP / IP protocol stack. The server uses a custom protocol to query the resource status of each network element, calculates the comprehensive resource score, and establishes a resource database with each network element ID, available bandwidth, multicast group address, computing power, and comprehensive resource score data. Using SNMP to scan the network topology, network elements are treated as nodes, links as edges, and the delay and bandwidth between nodes are used as edge weights to obtain a topology graph G = (V, E, W). The standard deviations of the delay, bandwidth, and comprehensive resource score between nodes are calculated, and the similarity of each pair of nodes is calculated. The network is divided into K sub-regions using a spectral clustering algorithm. Based on the similarity of each pair of nodes, the degree matrix is ​​calculated, and the Laplace matrix is ​​calculated using the phase difference method. The eigenvectors corresponding to the first K smallest non-zero eigenvalues ​​of the Laplace matrix are extracted to form a feature matrix. Initialize the clustering assignment matrix H, define the partitioning objective function, treat H as a continuous variable, and optimize using gradient descent to obtain the optimal sub-region partitioning with the goal of minimizing the partitioning objective function.

3. The cross-network element multicast method of claim 2, wherein: The encapsulation of multicast data packets, selection of the path with the lowest cost ratio, data transmission and synchronization encryption, and data integrity checks include: The server receives business data from the data source and encapsulates the business data, multicast group address, priority label, timestamp, and sequence number into a multicast data packet; Traverse the multicast tree, generate a path metadata table, use SNMP to query device status, obtain resource information of intermediate devices and links, and filter unusable paths; For each multicast data packet, select the optimal transmission path, prioritize the needs of VIP users, calculate the consumption ratio of each path, and select the path with the lowest consumption ratio as the optimal path. Check the network elements in the optimal path, use high-bandwidth links for network elements within an optimal sub-region, and use low-latency links across regions; The server uses IPsec to encrypt the service data in multicast packets, generates encrypted packets, and sends the encrypted packets to multicast group members using the optimal path. Based on the priority label of multicast packets, the DSCP field in the packet header is set for VIP and ordinary users. The server sends QoS configuration instructions to each intermediate device through a custom protocol and configures queue parameters on the intermediate devices. The intermediate device receives encrypted data packets and unicast forwarding instructions sent by the server, parses the forwarding instructions, extracts the next-hop IP address from the optimal path, and forwards the data packets. The last-hop device forwards the data packets to the client. The client uses IPsec to decrypt data packets, obtains business data, sequence number, timestamp and priority label, verifies the continuity of sequence number, marks it as a potential packet loss if the sequence number is not continuous, records the sequence number difference, and sends a packet loss notification to the server through a custom protocol. The decrypted business data is processed using the CRC-32 algorithm to generate a checksum, which is then sent back to the server. This checksum is compared with the checksum sent by the server. If they match, the transmission is successful; otherwise, the packet is lost. Data marked as packet loss or for which packet loss notifications were received will be replayed; The difference between the client's received timestamp and the server's sent timestamp is calculated using the phase difference method as the feedback delay.

4. The cross-network element multicast method of claim 3, wherein: The analysis of multicast operation performance metrics and optimization of multicast operation include: Performance metrics, including latency, packet loss rate, and bandwidth utilization, are collected and estimated based on historical data through regression. After normalization, the network's overall performance score is calculated. Thresholds for the overall performance scores of VIP and regular users are set separately. If the overall performance score of a path exceeds the corresponding threshold, the path is marked as a "path that needs optimization" and an optimization process is initiated. If VIP users are involved, optimization is triggered immediately, while regular users wait in a queue.

5. The cross- network element multicast method of claim 4, wherein: The audit of multicast group status generates a multicast group status table, including: In each sub-region, the server calculates the node score, selects the network element with the highest node score as the audit node, generates an audit task package from each client subset, RSA public key, and heartbeat message template in the sub-region, and sends the task package to the audit node using the gRPC protocol; The audit node sends heartbeat messages to the assigned subset of clients through a custom protocol, calculates an activity score, sets an activity score threshold based on historical regression estimation, and filters out clients with activity scores less than the activity score threshold as offline clients, otherwise they are considered active clients; The audit node sends authentication requests to active clients, requesting the return of an RSA signature. Active clients that do not return an RSA signature are marked as illegitimate clients. The audit node reports offline and unauthorized clients to the server via gRPC. The server removes the offline and unauthorized clients, records the reason for removal and timestamp, adds the IP of the unauthorized client to the server and intermediate devices to prevent subsequent joining requests, and generates a multicast group status table.

6. The cross- network element multicast method of claim 5, wherein: The construction of a visual interface to display and store multicast operation status data includes: The Grafana framework is used to generate interactive dashboards that display network topology diagrams, performance curve line graphs, and multicast group status tables. InfluxDB is used to store the network topology, performance curve line graphs, and multicast group status tables by timestamp.

7. A cross-network element multicast system, based on the cross-network element multicast method according to any one of claims 1 to 6, characterized in that: include, The deployment partitioning module is used to deploy multicast function software, use a custom protocol to query the resource status of each network element, construct a topology graph G, define a partitioning objective function, and use a spectral clustering algorithm to partition the network into sub-regions; A solution module is defined to define the convex quadratic optimization cost function, perform dynamic role allocation, define constraints, solve using a corrected search direction, and generate unicast forwarding instructions. Specifically, defining the convex quadratic optimization cost function, performing dynamic role allocation, defining constraints, solving using a corrected search direction, and generating unicast forwarding instructions includes: based on the topology graph G, defining decision variables, initializing cost function parameters, calculating dynamic factors, calculating the score of each node, sorting the scores of each node in descending order, and selecting candidate forwarding nodes; constructing the cost function, constructing the flow conservation constraint allocation matrix and the bandwidth constraint allocation matrix, and traversing each node of the topology graph G. The system identifies outgoing and incoming links, defines flow conservation and bandwidth constraints, and fills the flow conservation constraint allocation matrix and bandwidth constraint allocation matrix. Based on the dynamic factor, it calculates adaptive obstacle parameters. Based on the cost function and bandwidth constraints, it sets the initial decision variables as a neutral starting point, calculates the corrected search direction, updates the decision variables, and iterates to solve the problem. After each iteration, it verifies the bandwidth constraints. If the constraints are not met, the obstacle update parameters are halved and the problem is solved again. Otherwise, it continues to iterate until the maximum number of iterations is reached, at which point it stops iterating and outputs the optimal decision variables as a multicast tree. The server traverses the multicast tree, extracts paths, and outputs unicast forwarding instructions. The encapsulation and transmission module is used to encapsulate multicast data packets, select the path with the lowest cost ratio, perform data transmission and synchronization encryption, and check data integrity. The analysis and optimization module is used to analyze the performance metrics of multicast operation and optimize multicast operation. The audit module is used to audit the status of multicast groups and generate a multicast group status table. The visualization and storage module is used to build a visual interface to display and store multicast operation status data.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the cross-network element multicast method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the cross-network element multicast method according to any one of claims 1 to 6.

Citation Information

Patent Citations

  • Energy-saving routing spectrum allocation method based on anycast and multicast

    CN112203168A

  • NFV network multicast service function chain embedding method and system

    CN117439939A