Method, device and equipment for remote disaster recovery transmission of IPTV multicast services

Abstract

The invention relates to the technical field of remote disaster recovery transmission, and discloses a remote disaster recovery transmission method, device and equipment for IPTV multicast service. The method comprises: respectively using a multicast convergence point and a multicast transcoding server to build a networking architecture for a first area and a second area to obtain an initial dual-node multicast network; dynamically detecting the multicast transcoding server to obtain a multicast service routing mechanism; creating a VPN instance to obtain a service isolation environment; transcoding the multicast stream received in the first area and synchronizing it to the second area to obtain multicast service data; inputting the multicast service data into a double-layer proxy model, performing deep reinforcement learning on service scheduling, and obtaining an intelligent scheduling strategy; performing load sharing calculation and establishing a dual-active backup mechanism for the first area and the second area to generate a remote disaster recovery transmission scheme. The invention realizes dual-active deployment and dynamic load sharing of the service, and significantly improves the overall service quality of the IPTV live broadcast service.

Description

Technical Field

[0001] The present invention relates to the technical field of remote disaster recovery transmission, and in particular to a remote disaster recovery transmission method, device and equipment for IPTV multicast services. Background Art

[0002] At present, IPTV live broadcast services are mainly implemented through multicast technology, which requires the configuration of multicast convergence points (RPs) to complete the registration and distribution of multicast streams. The IPTV platform sets up a multicast transcoding server to register the multicast video stream to the RP, and the terminal watches the live program through the RP, forming a complete live broadcast service transmission system.

[0003] However, the existing IPTV multicast services have obvious technical shortcomings. The single-point deployment method lacks a disaster recovery mechanism. Once the multicast transcoding server or RP fails, it will affect the service experience of all users in the entire network, and the continuity and reliability of the service cannot be guaranteed. In addition, due to the lack of a cross-regional load sharing mechanism, the multicast transcoding server and RP in a single region need to bear the service access pressure of all users. When the user scale grows rapidly, performance bottlenecks are likely to occur. At the same time, the existing multicast service architecture also lacks intelligent service scheduling capabilities, and cannot be dynamically optimized according to network status and service load, making it difficult to meet the high reliability requirements of large-scale IPTV live services. Summary of the invention

[0004] The present invention provides a method, device and equipment for disaster recovery transmission of IPTV multicast services in different locations. The present invention realizes active-active deployment and dynamic load sharing of services, and significantly improves the overall service quality of IPTV live broadcast services.

[0005] In a first aspect, the present invention provides a method for transmitting IPTV multicast services in different locations in a disaster-tolerant manner, the method comprising:

[0006] A multicast convergence point and a multicast transcoding server are respectively used to build a networking architecture for the first area and the second area, and the multicast transcoding servers in the first area and the second area are configured with the same IP address to obtain an initial two-node multicast network;

[0007] According to the initial two-node multicast network, the multicast transcoding server is dynamically detected by ICMP, and a routing target strategy is established through IGP and BGP protocols to obtain a multicast service routing mechanism;

[0008] Creating a VPN instance for the multicast service routing mechanism, and binding the CDN server and the multicast transcoding server to the corresponding VPN instance to obtain a service isolation environment;

[0009] Based on the service isolation environment, transcode the multicast stream received in the first area and synchronize it to the second area, and realize the transmission of the multicast stream between the two areas through the MSDP peer mechanism to obtain multicast service data;

[0010] Inputting the multicast service data into a double-layer proxy model, performing deep reinforcement learning on service scheduling through an upper-layer policy network and a lower-layer execution network in the double-layer proxy model, and obtaining an intelligent scheduling strategy;

[0011] According to the intelligent scheduling strategy, load sharing calculation and active-active backup mechanism are performed on the first area and the second area to generate a remote disaster recovery transmission solution.

[0012] In a second aspect, the present invention provides an IPTV multicast service remote disaster recovery transmission device, the IPTV multicast service remote disaster recovery transmission device comprising:

[0013] A building module, used to respectively use a multicast convergence point and a multicast transcoding server to build a networking architecture for the first area and the second area, configure the multicast transcoding servers in the first area and the second area with the same IP address, and obtain an initial two-node multicast network;

[0014] A detection module is used to dynamically detect the multicast transcoding server in an ICMP manner according to the initial two-node multicast network, establish a routing target strategy through IGP and BGP protocols, and obtain a multicast service routing mechanism;

[0015] A binding module is used to create a VPN instance for the multicast service routing mechanism and bind the CDN server and the multicast transcoding server to the corresponding VPN instance to obtain a service isolation environment;

[0016] A transmission module, configured to transcode the multicast stream received in the first area and synchronize it to the second area based on the service isolation environment, and realize the transmission of the multicast stream between the two areas through the MSDP peer mechanism to obtain multicast service data;

[0017] A scheduling module, used for inputting the multicast service data into a double-layer proxy model, performing deep reinforcement learning on service scheduling through an upper-layer policy network and a lower-layer execution network in the double-layer proxy model, and obtaining an intelligent scheduling strategy;

[0018] A generation module is used to perform load sharing calculation and establish a dual-active backup mechanism for the first area and the second area according to the intelligent scheduling strategy, and generate a remote disaster recovery transmission plan.

[0019] The third aspect of the present invention provides a computer device, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the computer device executes the above-mentioned IPTV multicast service remote disaster recovery transmission method.

[0020] A fourth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the above-mentioned IPTV multicast service remote disaster recovery transmission method.

[0021] In the technical solution provided by the present invention, by deploying multicast transcoding servers and RP nodes in different locations and adopting the same IP address configuration, the single point failure problem is solved, the system reliability is significantly improved, and the interruption of the whole network service caused by equipment failure is avoided; the NQA function and the IGP / BGP hybrid routing mechanism are introduced to realize accurate server status monitoring and intelligent routing switching, and the timeliness and accuracy of service switching are guaranteed; the VPN isolation technology is adopted to logically isolate the CDN server from the multicast transcoding server, effectively solve the continuity problem of CDN recording service, and ensure the reliability of value-added services such as playback time shift; a two-layer proxy model is adopted for service scheduling, and intelligent load balancing and fault recovery are realized through deep reinforcement learning, which greatly improves the automation level of the system; a complete remote disaster recovery solution is constructed, the active-active deployment and dynamic load sharing of the service are realized, and the overall service quality of the IPTV live broadcast service is significantly improved; the present invention does not require additional dedicated switching equipment, and remote disaster recovery can be realized purely through network configuration, which reduces the deployment cost and simplifies the difficulty of operation and maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0023] Figure 1 A schematic diagram of the steps of a method for remote disaster recovery transmission of IPTV multicast services in an embodiment of the present invention;

[0024] Figure 2 It is a schematic diagram of the structure of a remote disaster recovery transmission device for IPTV multicast services in an embodiment of the present invention;

[0025] Figure 3 It is a schematic block diagram of the structure of a computer device in an embodiment of the present invention. DETAILED DESCRIPTION

[0026] Embodiments of the present invention provide a method, device and equipment for disaster recovery transmission of IPTV multicast services in different locations. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described here can be implemented in an order other than that illustrated or described here. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0027] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 , an embodiment of the IPTV multicast service remote disaster recovery transmission method in the embodiment of the present invention includes:

[0028] Step S1: Use a multicast convergence point and a multicast transcoding server to build a networking architecture for the first area and the second area respectively, configure the multicast transcoding servers in the first area and the second area with the same IP address, and obtain an initial two-node multicast network;

[0029] It is understandable that the execution subject of the present invention may be an IPTV multicast service remote disaster recovery transmission device, or a terminal or a server, which is not specifically limited here. The embodiment of the present invention is described by taking a server as the execution subject as an example.

[0030] Specifically, physical ports are configured for the multicast convergence points of the first area and the second area, so that the multicast convergence points of each area have the ability to connect to the network of other areas. Through the physical port configuration, an independent transmission network connection channel is established for the multicast convergence points of each area, so as to realize the interconnection structure between the areas. Based on the regional interconnection structure, the MSDP protocol is configured between the multicast convergence points of the first area and the second area, and a multicast synchronization mechanism is formed by establishing a peer connection. MSDP (Multicast Source Discovery Protocol), as a cross-domain multicast source discovery protocol, can ensure that multicast source information is shared between multiple areas, thereby realizing synchronous transmission of multicast streams, thereby ensuring that the multicast streams in the two areas can be transmitted consistently. According to the multicast synchronization mechanism, the multicast convergence point of the first area is configured to receive the initial multicast stream. At the same time, the multicast convergence point of the second area is set as a synchronous receiving end to ensure that the multicast streams of the two areas are synchronized, and the transmission path configuration of the multicast stream is completed, so that the multicast stream can be stably transmitted between the first area and the second area. In order to ensure that the multicast transcoding servers of the two areas can share the data stream, the IP addresses of the multicast transcoding servers of the two areas are uniformly assigned. Through this configuration, the multicast transcoding servers in the two regions use the same IP address to form an initial dual-node multicast network. Under this network structure, the transcoding server can realize automatic traffic switching and load balancing, so that even when a node fails, the other node can seamlessly take over the traffic to ensure the continuity of the multicast service. Based on the transcoding server configuration, the multicast transcoding servers in the first and second regions are bound to the CDN server respectively, and a real-time recording link is established to obtain a storage mechanism for the program. By binding the CDN server with the transcoding server, it is ensured that the multicast stream can be efficiently distributed between multiple regions, and the establishment of the real-time recording link provides technical support for the instant storage of program content. The recorded program data is deployed in the CDN servers in the two regions to ensure high availability and redundant backup of the program. The playback and time-shift functions are configured for the program, so that users can replay and pause according to their needs, providing users with a more flexible viewing experience. The recording service configuration is divided into service zones. According to the actual needs of the business, the multicast convergence points in the first and second regions are set as the live service access nodes of users in their respective regions to obtain the service partition strategy. According to the service partitioning strategy, the same external service address is configured for the multicast convergence points in the first area and the second area to generate an initial two-node multicast network.

[0031] Step S2: According to the initial two-node multicast network, the multicast transcoding server is dynamically detected by ICMP, and a routing target strategy is established through IGP and BGP protocols to obtain a multicast service routing mechanism;

[0032] Specifically, NQA is configured on the egress switch in the initial two-node multicast network to analyze the network quality in real time. Through NQA, the quality of each link in the network is evaluated, including bandwidth, delay, and packet loss rate, to help the system determine whether the network meets the transmission requirements of the multicast service. Based on the network quality analysis mechanism, the IP addresses of the multicast transcoding servers in the first area and the second area are detected for survival. The ICMP protocol ping operation is used to monitor the heartbeat of each transcoding server to ensure that the server is in normal operation and to provide real-time feedback on the survival status of the server. If a transcoding server fails or is unreachable, the system will automatically adjust the routing policy and transfer it to other healthy servers to ensure that the multicast service is not interrupted. According to the monitoring results of the multicast server, the OSPF (Open Short Path First) neighbor relationship is established between the multicast aggregation points in the first area and the second area and the corresponding egress switches, and an IGP routing connection is generated. OSPF is an internal gateway protocol that calculates the shortest path based on the link status information of each switch to ensure that data packets can be efficiently and reliably transmitted within the area. After establishing neighbor relationships through OSPF, the egress switch configures the detailed routes of the multicast transcoding server according to the route calculation results, and publishes the route information to the multicast aggregation point through the IGP protocol to complete the routing policy configuration within the area. The BGP (Border Gateway Protocol) neighbor relationship is configured for the multicast aggregation points in the first area and the second area, and a BGP connection channel is established. The BGP protocol is an external gateway protocol used for route selection across autonomous systems. The multicast aggregation points publish the detailed routes of the multicast transcoding server to each other through the BGP protocol to implement inter-area routing policy formulation. The introduction of BGP ensures the correct transmission of routing information between different areas and improves the reliability of cross-area transmission. Based on the intra-area and inter-area routing policies, the priority of the intra-area routing policy is set higher than the priority of the inter-area routing policy. When the multicast aggregation point faces multiple routing paths, the local IGP route will be selected first to reduce the overhead and delay of cross-area transmission. In this way, the traffic within the area is transmitted through the shortest path, thereby improving the efficiency and stability of the network. At the same time, the routing table of the multicast aggregation point is configured with routing targets to ensure that each aggregation point can select the best path according to the current network conditions. When the network changes, the route selection in the routing table is dynamically adjusted according to the route priority mechanism. For example, under normal circumstances, the multicast aggregation point selects the local IGP route as the preferred path. Once the path fails, the system will automatically switch to the BGP route to maintain business continuity.Through the dynamic update and route switching mechanism, the routes of IPTV multicast services can be seamlessly switched between different areas, thus ensuring high availability and disaster recovery capabilities of the service.

[0033] Step S3: Create a VPN instance for the multicast service routing mechanism, and bind the CDN server and the multicast transcoding server to the corresponding VPN instance to obtain a service isolation environment;

[0034] Specifically, a vlanif interface is created for the physical ports of the egress switches of the first area and the second area and the multicast convergence point. The VLAN interface (vlanif) isolates each service flow in the physical network and provides a clear routing channel by dividing the physical ports, thereby ensuring the effective transmission of different service flows and avoiding mutual interference of network flows. Based on the created service bearer channel, the ports of the CDN server egress switches in the first area and the second area are marked and divided to form a CDN access link, thereby ensuring that each access port of the CDN server can be clearly associated with the corresponding network channel and that the data flow can be transmitted through a suitable path, thereby avoiding service flow interruption or performance degradation due to port configuration errors. Through the marking and division of the CDN access link, efficient management of different service flows is achieved and the traffic transmission path of the CDN is optimized. The vlanif interface and the CDN server are associated and bound to the VPN instance, and the multicast protocol is enabled in the VPN. The entire service flow is divided into the multicast domain of the VPN, thereby ensuring that all multicast flows can be transmitted through a dedicated path in the service isolation environment and will not be interfered with by other service flows. By creating VPN instances and enabling multicast protocols, the service traffic in multiple regions is managed uniformly and the efficient transmission of multicast streams is ensured. Routing configuration is performed on VPN multicast domains, and the default route is directed to the multicast convergence point of each region. This configuration ensures that the data stream is always transmitted along the preset path before reaching the target multicast convergence point. At the same time, the original tree of the CDN server is forced to bypass the multicast convergence point to control the specific traffic path and avoid unnecessary traffic from going to the multicast convergence point. This process helps to enhance the controllability of the network and optimize the network traffic distribution, making the transmission of multicast traffic more efficient and stable. After completing the routing configuration, based on the VPN routing policy, the CDN servers and multicast transcoding servers in the first and second regions are isolated and configured to establish a logical isolation structure. The logical isolation structure separates the network environment of different regions and servers to ensure that the traffic of each region and server does not affect each other, thereby realizing the service isolation function. Each region and its related servers run in an isolated environment to avoid network performance problems caused by traffic conflicts or resource contention between different regions. The NQA detection results are associated with the VPN instances in the logical isolation structure. The detailed routing status of the multicast transcoding server is synchronized to the VPN routing table through the NQA detection results. NQA detection (Network Quality Analysis) monitors the network quality in real time to ensure the operation status of the multicast transcoding server in the network. By synchronizing the detailed routing status of the multicast transcoding server, the VPN routing table can dynamically reflect the availability of the server, ensuring that the routing path can be quickly adjusted in the event of a network failure to avoid data flow loss or transmission delay.Based on the routing synchronization mechanism, the VPN routing table will be dynamically updated regularly to ensure that the paths in the routing table are always up to date, and when the local detailed routing fails, it will automatically switch to the BGP routing. Get the VPN switching strategy. Through the VPN switching strategy, when the original multicast path fails, the system can automatically switch the traffic to the backup path to avoid service interruption, thereby protecting the continuity of the CDN server's recording business and obtaining a business isolation environment.

[0035] Step S4: Based on the service isolation environment, the multicast stream received in the first area is transcoded and synchronized to the second area, and the multicast stream is transmitted between the two areas through the MSDP peer mechanism to obtain multicast service data;

[0036] Specifically, the initial multicast stream received in the first area is parsed and transcoding parameters are configured to ensure that the transcoding task can be performed in accordance with the predetermined format and requirements. By parsing the multicast stream, the format and parameters of the original multicast stream are analyzed, and then the corresponding transcoding configuration is performed to ensure that it can meet the playback requirements and standards of different areas. After the transcoding task definition is completed, the initial multicast stream is transcoded to generate a transcoded multicast stream that meets the playback standards of the target area. After the transcoding is completed, the IP address of the transcoded multicast stream is registered through the multicast transcoding server in the first area, and registered with the multicast convergence point in the area. Through the registration mechanism, the source information of the multicast stream and its transmission path are recorded to generate local registration data. Through the local registration data, the multicast stream is correctly identified and transmitted in the first area. In order to realize cross-regional multicast stream transmission, the MSDP (Multicast Source Discovery Protocol) Peer mechanism is used to synchronize the multicast stream to the multicast convergence point in the second area. MSDP is a protocol for sharing multicast source information across regions. It can ensure that source address information is correctly propagated between different regions, realize cross-regional synchronous transmission of multicast streams, and obtain synchronized multicast streams. After the synchronized multicast stream reaches the second region, the IP address is registered with the multicast transcoding server in the second region according to the data of the synchronized stream, and the multicast convergence point in the region is registered to generate remote registration data. Through this process, the network equipment in the second region can identify and process the multicast stream from the first region, ensuring the smooth transmission and processing of the multicast stream in the second region. For the synchronized multicast stream, secondary transcoding is performed to ensure that the multicast stream meets the format standard of the first region and generates a remote multicast stream with the same format as the first region. The remote multicast stream is input to the CDN server in the second region for recording and storage of multicast services. Through the recording and storage functions of the CDN server, the remote multicast stream is backed up and becomes the key data for disaster recovery. The uniqueness of the (S, G) table entry of the multicast transcoding server is verified and confirmed based on the local registration data and the remote registration data. S stands for source address, and G stands for multicast group address. Through verification and confirmation, the source and destination addresses of each multicast stream are ensured to be unique, avoiding data conflicts and duplications. After verification, the multicast service data generated finally contains all necessary information, ensuring the stable transmission of multicast streams from the source to the destination between the two areas, as well as the consistency and integrity of data during disaster recovery.

[0037] Step S5: input the multicast service data into the double-layer proxy model, perform deep reinforcement learning on the service scheduling through the upper-layer policy network and the lower-layer execution network in the double-layer proxy model, and obtain an intelligent scheduling strategy;

[0038] Specifically, feature extraction is performed on multicast service data to separate the load status of each regional multicast convergence point (RP), the CPU occupancy rate, memory usage rate, network bandwidth utilization rate, and user access volume distribution data of the multicast transcoding server. These data represent the health status of the network and server, resource usage, and service load. These feature data are constructed into a multidimensional service feature vector to describe the current state of the entire system. The multidimensional service feature vector is input into the state encoding layer of the upper policy network in the two-layer proxy model. The state encoding layer extracts time series features through a three-layer long short-term memory network (LSTM), and each LSTM layer contains 128 hidden units. LSTM is a neural network structure that processes time series data and can capture long-term and short-term dependencies in the data, thereby effectively modeling the dynamic changes of the network state. Through the extraction process of the three-layer LSTM, the current time series state representation of the system is obtained, which reflects the change trend of various indicators in the system and can help the upper network predict and make decisions on the future state. The time series state representation is processed by the decision layer of the upper policy network. The decision layer contains a bidirectional attention mechanism and a three-layer fully connected network. The bidirectional attention mechanism is used to dynamically allocate weights for different state representations, thereby focusing on the key features that have the greatest impact on the decision. After being processed by the attention mechanism and the fully connected network, the decision layer can output global decision results such as service switching strategy, load balancing strategy, and resource allocation strategy. These strategies indicate how to allocate network resources between different regions, when to switch services, and how to balance server loads to achieve efficient service scheduling. The global decision results are input into the action generation layer of the upper policy network. The action generation layer adopts the Actor-Critic architecture to generate specific decision action sequences through interactive optimization of the policy network and the value network. The Actor part is responsible for generating actions, while the Critic part evaluates the generated actions, thereby continuously optimizing the action selection strategy. In this way, the system can gradually learn the optimal scheduling strategy so that each decision can maximize the overall utility of the system. The decision action sequence is task-decomposed. It is processed by the task planning layer of the lower execution network in the two-layer proxy model. The task planning layer contains a graph convolutional network (GCN) for dependency analysis of the generated routing switching instructions, VPN configuration instructions, and multicast stream switching instructions. In this process, the interdependencies between tasks are identified and an execution dependency graph is constructed. Through the processing of the graph convolutional network, the dependencies between these tasks are clarified to ensure that there will be no conflicts or order errors during execution. The execution dependency graph is input into the parameter generation layer of the lower execution network. The parameter generation layer contains five layers of residual blocks, each of which consists of two convolutional layers and a short-circuit connection. The introduction of residual blocks effectively solves the problem of gradient vanishing in deep neural networks, allowing the network to learn deeper features.Through these residual blocks, the system can generate specific parameter configuration schemes, including execution parameters for each task, such as the time of route switching, VPN configuration details, and switching conditions for multicast streams. Based on the generated parameter configuration scheme, the execution control layer of the lower-level execution network is responsible for specific task execution. The execution control layer uses a reinforcement learning algorithm and uses the network's NQA detection data, IGP routing status, BGP routing status, and multicast stream status as environmental feedback. The core of reinforcement learning is to continuously adjust the execution strategy through interaction with the environment to maximize the accumulated rewards. During the execution process, the execution strategy is continuously adjusted according to the feedback from the environment. For example, when the NQA detection finds that the network quality of a certain path has deteriorated, the system will automatically adjust the route to select a better path. When the BGP or IGP status changes, the system will also re-evaluate the execution plan based on the new status information, thereby achieving dynamic network optimization. The resulting intelligent scheduling strategy efficiently manages network resources and ensures load balancing between regions and seamless switching of services.

[0039] Step S6: Perform load sharing calculation and establish a dual-active backup mechanism for the first area and the second area according to the intelligent scheduling strategy, and generate a remote disaster recovery transmission solution.

[0040] Specifically, key scheduling control parameters are extracted from the intelligent scheduling strategy, including service switching rules, load threshold parameters, and resource allocation coefficients. Service switching rules refer to how to switch service flows between different areas under specific conditions; load threshold parameters refer to the need for the system to perform load balancing when the service flow reaches a certain threshold; and resource allocation coefficients refer to how to perform optimal resource allocation according to the resource conditions of different areas. Based on the scheduling control parameter set, the geographical partitioning calculation of the service coverage of the first area and the second area is performed. The user distribution and load conditions in the two areas are analyzed to formulate a suitable user service strategy. Under this strategy, the system can reasonably divide the user's access area and ensure that the load of each area does not exceed its carrying capacity through different access strategies, while maintaining the efficient operation of the entire system. According to the user service strategy, the multicast aggregation points in the first area and the second area are respectively configured for access to user live broadcast services. As an access point, the multicast aggregation point is responsible for distributing the multicast stream to the terminal users. By configuring the multicast aggregation points in the two areas for load sharing, it is ensured that when the traffic in one area is too high, part of the traffic load is transferred to another area, so as to achieve more balanced resource use. The load sharing solution obtained in this process ensures that resources between the two areas are reasonably allocated and avoids service interruption caused by overload in a single area. After completing the load sharing configuration, a backup path is configured for the BGP route between the first area and the second area. As an important routing protocol, BGP can provide path selection services between different areas and configure a backup path for BGP routes. Through intelligent scheduling strategies, the weight of BGP routes is dynamically adjusted to ensure that when the main path fails, it can quickly switch to the backup path to ensure the stability and continuity of data transmission. The implementation of the routing disaster recovery mechanism can improve the reliability of the system and reduce the impact on the service when a network failure occurs. Cross-regional backup configuration is performed for the multicast transcoding servers in the first area and the second area. The formulation of the backup switching strategy includes the backup of the transcoding server, as well as the dynamic adjustment of traffic and the redistribution of loads to ensure that the system can automatically adjust and continue to run when a failure occurs. When the backup switching strategy is determined, the strategy is deployed to the CDN servers in the first area and the second area to configure active-active disaster recovery for the recording service. According to the storage backup mechanism, an independent multicast stream source is introduced to the second area. The second area no longer relies solely on the stream source of the first area, but can independently generate and synchronize multicast streams, thereby improving the disaster recovery capability of the system. This is accomplished through the recording synchronization function of the CDN server, ensuring that the multicast stream data of the two areas are synchronized and consistent, and ensuring that the backup data of the two areas is equal. Based on the peer-to-peer backup structure, the load sharing solution, routing disaster recovery mechanism, and storage backup mechanism are integrated and associated. By organically combining these mechanisms, efficient management and disaster recovery capabilities of IPTV multicast services are achieved.Under the integrated off-site disaster recovery transmission solution, whether in terms of business switching, load balancing or data backup, the system can be dynamically adjusted according to the real-time status to ensure that it can quickly recover and continue to provide services when any failure occurs.

[0041] In the embodiment of the present invention, by deploying a multicast transcoding server and an RP node in a different location and adopting the same IP address configuration, the single point failure problem is solved, the system reliability is significantly improved, and the interruption of the whole network service caused by equipment failure is avoided; the NQA function and the IGP / BGP hybrid routing mechanism are introduced to realize accurate server status monitoring and intelligent routing switching, and the timeliness and accuracy of service switching are guaranteed; the VPN isolation technology is adopted to logically isolate the CDN server from the multicast transcoding server, effectively solve the continuity problem of the CDN recording service, and ensure the reliability of value-added services such as playback time shift; a two-layer proxy model is adopted for service scheduling, and intelligent load balancing and fault recovery are realized through deep reinforcement learning, which greatly improves the automation level of the system; a complete remote disaster recovery solution is constructed, dual-active deployment and dynamic load sharing of the service are realized, and the overall service quality of the IPTV live broadcast service is significantly improved; the present invention does not require additional dedicated switching equipment, and remote disaster recovery can be realized purely through network configuration, which reduces deployment costs and simplifies operation and maintenance difficulties.

[0042] In a specific embodiment, the process of executing step S1 may specifically include the following steps:

[0043] Physical port configuration is performed on the multicast convergence points of the first area and the second area, and transmission network connection channels are established respectively to obtain a regional interconnection structure;

[0044] Based on the regional interconnection structure, MSDP protocol configuration and peer connection relationship establishment are performed on the multicast convergence points of the first area and the second area to obtain a multicast synchronization mechanism;

[0045] Performing an initial multicast stream receiving configuration on the multicast convergence point in the first area according to the multicast synchronization mechanism, setting the multicast convergence point in the second area as a synchronization receiving end, and obtaining a multicast stream transmission path;

[0046] Uniformly assign IP addresses to multicast transcoding servers in the first area and the second area in the multicast stream transmission path to generate transcoding server configurations;

[0047] Based on the transcoding server configuration, CDN server binding and real-time recording link establishment are performed on the multicast transcoding servers in the first area and the second area respectively to obtain a program storage mechanism;

[0048] Deploy the program storage mechanism in the CDN servers in the first area and the second area, configure playback and time shift functions, and obtain recording service configuration;

[0049] Divide the recording service configuration into service zones, set the multicast convergence points in the first zone and the second zone as live service access nodes for users in the respective zones, and obtain a service zone strategy;

[0050] According to the service partitioning strategy, the same external service address is configured for the multicast convergence points in the first area and the second area to generate an initial two-node multicast network.

[0051] Specifically, physical ports are configured for the multicast convergence points in the first area and the second area. These ports are interfaces for connecting the multicast convergence point with other devices in the network and provide a transmission path for the multicast stream. Assume that the IP address of the multicast convergence point in the first area is , the IP address of the multicast aggregation point in the second area is Through the configuration of these physical ports, multicast streams are smoothly transmitted from one area to another. The multicast aggregation point in each area is connected to the switch or router in the network. The bandwidth of the connection between areas is set according to the needs. The bandwidth should take into account the actual data volume of the multicast stream. ,satisfy To ensure stable transmission of traffic. Through the regional interconnection structure, the multicast convergence points of the first and second regions are configured with the MSDP (Multicast Source Discovery Protocol) protocol based on the requirements of the network topology to achieve synchronization of multicast sources. The MSDP protocol is used to propagate multicast source information between different regions. MSDP ensures that multicast convergence points in different regions exchange source information by establishing peer connections. After setting up the MSDP connection, multicast sources are transmitted between regions through the following relationship: ;

[0052] in, Indicates the synchronization mechanism of the MSDP protocol. and are the IP addresses of the multicast convergence points in the first area and the second area respectively, and Source Data represents the multicast source data to be transmitted. After the MSDP protocol configuration is completed, the multicast convergence point in the first area is configured to receive the initial multicast stream according to the multicast synchronization mechanism. After the multicast convergence point in the first area receives the multicast stream, it synchronizes the stream to the multicast convergence point in the second area through the specified path according to the synchronization mechanism. Assume that the source IP of the initial multicast stream is , the synchronous IP of the receiving end is , the synchronous transmission path of the multicast stream is expressed as: ;

[0053] in, is the transmission path of the multicast stream, is the multicast aggregation point address of the source area, is the multicast aggregation point address of the target area, is the traffic data size. The IP addresses of the multicast transcoding servers in the first and second regions in the multicast stream transmission path are uniformly assigned to ensure that the multicast stream can be accurately routed to the transcoding servers in each region. Assume that the multicast transcoding servers in the first and second regions are and , the IP address of each transcoding server is configured as:

[0054] ;

[0055] ;

[0056] in, and Represent the multicast transcoding server configurations of the first and second regions respectively. The IP addresses are configured by the network administrator and ensure that the transcoding servers are uniquely identified in the network. Based on the transcoding server configurations, the multicast transcoding servers of the first and second regions are bound to the CDN (Content Delivery Network) servers respectively, and a real-time recording link is established. Assume that the CDN servers of the first and second regions are and , the recording link is established by the following formula: ;

[0057] in, Indicates the recording link. and It is a transcoding server. and It is the CDN server in each region. This link ensures that the CDN server in each region can receive and store the stream from the multicast transcoding server in real time. After completing the recording link, the program storage mechanism is deployed to the CDN servers in the first and second regions, and the playback and time-shift functions are configured. The playback and time-shift functions allow users to view recorded programs at any time, and the time-shift function allows users to pause the currently playing program and continue watching it later. The implementation of these functions depends on the storage mechanism, and the data is cached and distributed in the CDN server. The storage configuration for the playback and time-shift functions is set as follows: ,TimeShift,Playback);

[0058] Among them, TimeShift and Playback represent the configuration options of time shift and playback functions respectively. The recording service configuration is divided into service zones, and the multicast convergence points of the first zone and the second zone are set as the live service access nodes of the users in their respective zones. By dividing different service zones, the multicast convergence points in each zone only provide multicast services to users in the zone. Assume that the service zone strategy is and , corresponding to the service configuration of the first area and the second area. The business partition strategy is expressed by the following formula:

[0059] ;

[0060] ;

[0061] in, and The business zoning strategy for each area ensures that different user groups access different multicast content. According to the business zoning strategy, the same external service address is configured for the multicast convergence points in the first area and the second area to ensure that external users can access the content of the two areas through a unified address without being restricted by the area. The external service address configuration is expressed by the following formula: ;

[0062] Through the above steps, the configuration of the initial two-node multicast network is completed, and efficient, redundant and sustainable multicast transmission between different areas is achieved.

[0063] In a specific embodiment, the process of executing step S2 may specifically include the following steps:

[0064] Perform NQA function configuration on the egress switch in the initial two-node multicast network to obtain the network quality analysis mechanism;

[0065] Based on the network quality analysis mechanism, the survival status of the multicast transcoding server IP addresses in the first area and the second area is detected to generate a multicast server monitoring result;

[0066] According to the monitoring result of the multicast server, an OSPF neighbor relationship is established between the multicast convergence point in the first area and the second area and the corresponding egress switch to obtain an IGP routing connection;

[0067] Based on the IGP routing connection, the egress switch is configured with detailed routing for the multicast transcoding server, and the local routing is published to the multicast aggregation point through the IGP routing to obtain the intra-regional routing policy;

[0068] Configure the BGP neighbor relationship between the multicast convergence points in the first area and the second area, establish a BGP connection channel, and mutually publish detailed routes of the multicast transcoding servers to obtain an inter-area routing strategy;

[0069] The priority parameter of the intra-regional routing policy is set higher than the priority parameter of the inter-regional routing policy, and the routing target is configured for the multicast convergence point to obtain a routing priority mechanism;

[0070] According to the routing priority mechanism, the routing table of the multicast aggregation point is prioritized for routing selection, and the local IGP route is used as the preferred routing path to obtain the routing switching strategy;

[0071] Based on the routing switching strategy, the routing table of the multicast convergence point is dynamically updated. When the local route fails, it automatically switches to the BGP route to obtain a multicast service routing mechanism.

[0072] Specifically, enable the NQA function on the egress switch. The NQA function is a mechanism for real-time monitoring of network quality, evaluating key performance indicators such as latency, bandwidth, and packet loss rate between nodes in the network. Assume that the IP address of the egress switch in the first area of ​​the network is , the IP address of the egress switch in the second area is , through NQA configuration, the connection quality between these switches and other network devices is monitored in real time. NQA obtains network quality data between the switch and the target server by setting regular test tasks on each node. Assume that the monitoring result is and , respectively represent the NQA detection results of the first area and the second area. These monitoring results are expressed as:

[0073] ;

[0074] ;

[0075] Among them, Test Parameters includes network quality related indicators such as packet loss rate, delay, and bandwidth. After obtaining the network quality analysis mechanism based on the NQA function, the IP address survival status of the multicast transcoding server in the first area and the second area is detected to ensure that the IP address of the multicast transcoding server is reachable in the network to avoid the situation where the multicast stream cannot be transmitted. Assume that the IP addresses of the multicast transcoding servers in the first area and the second area are and , through the survival status detection mechanism, the IP addresses of the two servers are regularly detected to ensure their normal operation. The survival status detection results generate multicast server monitoring results. Assume that the result is and , indicating whether the multicast transcoding server is reachable. The detection formula of the monitoring result is expressed as:

[0076] ;

[0077] ;

[0078] Among them, Ping Test is a detection method used to detect whether the multicast transcoding server is in an active state. Based on the monitoring results of the multicast server, OSPF (Open Shortest Path First) neighbor relationships are established between the multicast aggregation points in the first area and the second area and the export switches according to the status of the multicast server, forming an IGP (Interior Gateway Protocol) routing connection. OSPF is an internal gateway protocol that is applicable to the exchange of routing information within the same autonomous system. Assume that the IP address of the multicast aggregation point in the first area is , the IP address of the multicast aggregation point in the second area is The export switches are and , in these two areas, the routing information exchange between the multicast aggregation point and the egress switch is realized by establishing OSPF neighbor relationships. The process of OSPF neighbor relationship configuration is expressed by the following formula:

[0079] ;

[0080] ,Neighbor Config);

[0081] After the OSPF neighbor relationship is established, the detailed routing configuration of the multicast transcoding server is performed on the egress switch based on the IGP routing connection. The detailed routing configuration of the multicast transcoding server ensures that the egress switch can publish local routing information to the multicast aggregation point according to the specific needs of the multicast stream. Assume that the routing configuration result is and , respectively representing the IGP routing information of the first area and the second area, and publishing the local routes to the multicast aggregation point through IGP to form the routing policy within the area. The routing policy formula within the area is:

[0082] ;

[0083] ;

[0084] In order to implement the inter-regional routing strategy, the BGP (Border Gateway Protocol) neighbor relationship is configured for the multicast convergence points in the first region and the second region. BGP is a routing protocol used between different autonomous systems and can implement inter-regional routing information exchange. Assume that the IP addresses of the multicast convergence points in the first region and the second region are and After the BGP neighbor relationship is configured, a routing connection channel between regions is established. Assume that the BGP connection is configured as and , indicating the BGP neighbor relationship established between the two areas, is expressed by the following formula:

[0085] ;

[0086] ;

[0087] Through the connection of BGP protocol, the detailed route of multicast transcoding server is published, and the inter-regional routing strategy is obtained to ensure that the multicast convergence points in each region can maintain the integrity and accuracy of data in cross-regional transmission. In order to optimize routing performance and ensure efficient traffic management, the priority parameters of intra-regional routing policies are set higher than the priority parameters of inter-regional routing policies to ensure that the routing priority within the local area is higher and the delay of cross-regional routing is reduced. Assume that the intra-regional routing priority parameter is , the inter-area routing priority is , the rule for setting the priority is expressed by the following formula: ;

[0088] According to this priority rule, the routing target of the multicast convergence point is configured to obtain the routing priority mechanism. The routing priority mechanism ensures that local routes will be given priority when selecting routes, reducing the dependence on cross-regional routes. By configuring the routing priority system, a more efficient routing switching strategy can be achieved. The formula is as follows:

[0089] ;

[0090] Based on the routing switching strategy, the routing table of the multicast aggregation point is dynamically updated according to the routing priority mechanism. When the local IGP route fails, the system automatically switches to the BGP route to ensure the continuity and stability of the multicast flow. The multicast service routing mechanism ensures that the multicast flow can be transmitted through the optimal path under any circumstances. The formula for the routing switching strategy is: ;

[0091] Through the dynamic update mechanism, the routing of multicast services can be quickly switched, thus ensuring high reliability and stability of the network.

[0092] In a specific embodiment, the process of executing step S3 may specifically include the following steps:

[0093] Create a vlanif interface for the physical ports of the egress switches and the multicast convergence point in the first area and the second area to obtain a service bearer channel;

[0094] Based on the service bearer channel, the ports of the CDN server uplink egress switches in the first area and the second area are marked and divided to obtain CDN access links;

[0095] According to the CDN access link, associate and bind the vlanif interface and the CDN server with the VPN instance, and enable the multicast protocol in the VPN to obtain the VPN multicast domain.

[0096] Configure the routing for the VPN multicast domain, point the default route to the multicast aggregation point in the corresponding area, force the original tree of the CDN server to bypass the multicast aggregation point, and obtain the VPN routing policy;

[0097] Based on the VPN routing policy, the CDN servers and the multicast transcoding servers in the first area and the second area are isolated and configured to obtain a logical isolation structure;

[0098] The NQA detection results are associated with the VPN instances in the logical isolation structure, and the detailed routing status of the multicast transcoding server is synchronized to the VPN routing table to obtain a routing synchronization mechanism;

[0099] Based on the routing synchronization mechanism, the VPN routing table is dynamically updated. When the local detailed routing fails, it automatically switches to the BGP routing to obtain the VPN switching strategy. According to the VPN switching strategy, the CDN server's recording service is continuously protected to obtain a service isolation environment.

[0100] Specifically, a virtual LAN interface (vlanif) is configured on the physical ports of the egress switch and the multicast aggregation point to separate and manage data flows, so that different network flows are isolated and carried in the same physical connection. Assume that the egress switch in the first area is , the egress switch in the second area is , the multicast aggregation point in the first area is , the multicast aggregation point in the second area is , configure virtual LAN interfaces for them respectively and , and create a service bearer channel. The creation of a service bearer channel requires the configuration of the relevant VLAN (virtual local area network) ID so that data can be diverted and forwarded according to the predefined VLAN. The formula for this step is:

[0101] ;

[0102] ;

[0103] Among them, VLAN ID represents the VLAN number assigned to each area. After the configuration of the vlanif interface is completed, the network traffic will be logically isolated in each area to form two independent service bearer channels. Based on the service bearer channel, the ports of the CDN (Content Distribution Network) server uplink export switches in the first area and the second area are identified and divided to obtain CDN access links. The uplink port of each CDN server needs to match the physical port of the export switch and assign it a unique identifier (for example, port number or VLAN ID). Assume that the uplink port of the CDN server in the first area is , the uplink port of the CDN server in the second area is , and its identification division is achieved by assigning specific VLANs or IP addresses to these ports. These identification divisions ensure that each CDN server can be correctly connected to the upstream egress switch and form a stable CDN access link. The formula is:

[0104] ;

[0105] ;

[0106] Among them, Port ID represents the identifier assigned to each CDN server uplink port. After completing the configuration of the access link, the CDN server exchanges data with other devices in the network (such as multicast transcoding servers) through these links. According to the CDN access link, the vlanif interface and the CDN server are associated with the VPN instance, and the multicast protocol is enabled in the VPN to build a VPN multicast domain. The purpose of VPN instance binding is to configure VPN in each area, interconnect different regional networks, enable the multicast protocol, and ensure the transmission of multicast streams. Assume that the name of the VPN instance is and , corresponding to the VPN configuration of the first area and the second area respectively, a VPN binding relationship is created between the vlanif interface of each area and the CDN server, and the multicast protocol is enabled in the VPN to support the transmission of multicast streams. The establishment of the VPN multicast domain is expressed by the following formula:

[0107] ;

[0108] ;

[0109] Multicast Protocol represents the multicast protocol (such as PIM-SM) enabled in VPN. By enabling the multicast protocol, the VPN multicast domain ensures the correct forwarding and transmission of multicast streams across regions. Based on the VPN multicast domain, the routing configuration is performed, and the default route is pointed to the multicast aggregation point of the corresponding region, forcing the original tree of the CDN server to bypass the multicast aggregation point. Through this configuration, the multicast streams of the CDN server are ensured to pass through the multicast aggregation point, thus avoiding the bottleneck and delay caused by the multicast aggregation point. Assume that the default route is and , pointing to the multicast convergence point in the first and second areas and , the routing configuration is expressed as:

[0110] ;

[0111] ;

[0112] Through this configuration, the traffic of the CDN server can bypass the multicast aggregation point, avoiding unnecessary traffic concentration at the aggregation point, thereby improving the efficiency and stability of the network. Based on the VPN routing strategy, the CDN servers and multicast transcoding servers in the first and second regions are isolated and configured to obtain a logical isolation structure, thereby achieving resource and traffic independence in the same VPN environment and avoiding resource conflicts in different regions. Assume that the IP address of the multicast transcoding server is and , configure independent subnets or VLANs for them in the VPN instance to ensure isolation between them. The configuration formula for the logical isolation structure is:

[0113] ;

[0114] ,Subnet Config);

[0115] With this configuration, the CDN servers and multicast transcoding servers in the first and second regions will be isolated and form independent network structures. By associating the NQA detection results, the detailed routing status of the multicast transcoding server is synchronized to the VPN routing table, and a routing synchronization mechanism is obtained. The NQA detection function is used to monitor the status of each multicast transcoding server in real time, including performance indicators such as network delay and packet loss rate. By associating the NQA detection results with the VPN instance, the multicast flow routing status in the VPN routing table is synchronized with the actual network environment. Assume that the NQA detection result is and , and associate it with the VPN routing table through the following formula:

[0116] ;

[0117] ;

[0118] The synchronization mechanism can dynamically adjust the information in the routing table according to the network conditions to ensure the stability and reliability of the network. Based on the routing synchronization mechanism, the VPN routing table will be dynamically updated. When the local detailed routing fails, the system will automatically switch to the BGP routing to ensure the continuity of network traffic. The use of BGP routing ensures that the network can quickly switch to the backup path when the local routing fails. The VPN switching strategy ensures the continuous transmission of multicast streams.

[0119] In a specific embodiment, the process of executing step S4 may specifically include the following steps:

[0120] Performing code stream parsing and transcoding parameter configuration on the initial multicast stream received in the first area to obtain a transcoding processing task, and performing transcoding processing on the initial multicast stream based on the transcoding processing task to generate a transcoded multicast stream;

[0121] Register the transcoded multicast stream with the multicast transcoding server in the first area through the IP address, register it with the multicast convergence point in the first area, and obtain local registration data;

[0122] According to the local registration data, the multicast stream is transmitted synchronously across regions through the MSDP peer mechanism, and the transcoded multicast stream is synchronized to the multicast convergence point in the second region to obtain a synchronized multicast stream;

[0123] Based on the synchronous multicast stream, register the IP address of the multicast transcoding server in the second area, register with the multicast convergence point in the second area, and obtain remote registration data;

[0124] Perform secondary transcoding on the synchronous multicast stream to generate a multicast stream in the same format as that in the first area, obtain a remote multicast stream, and input the remote multicast stream into a CDN server in the second area to record and store the multicast service to obtain backup multicast data;

[0125] According to the local registration data and the remote registration data, the (S, G) table entry of the multicast transcoding server is uniquely verified and confirmed to obtain the multicast service data.

[0126] Specifically, the initial multicast stream received in the first area is parsed and transcoding parameters are configured to generate a transcoding processing task. Stream parsing is to analyze the received original video stream data, extract the audio, video, subtitles and other information, and configure the transcoding parameters according to the preset transcoding requirements. Transcoding parameter configuration involves selecting key parameters such as video encoding format, resolution, frame rate, bit rate, etc. These parameters directly affect the quality and bandwidth occupancy of the final generated transcoded stream. After completing the transcoding parameter configuration, based on these parameters, the transcoding processing task is started to transcode the original multicast stream to generate a transcoded multicast stream that meets the target specifications. This process involves re-encoding and compressing the video to adapt to different network environments and user equipment requirements. For example, the original stream is in H.264 encoding format, and after transcoding, it is converted to H.265 or VP9 format to reduce bandwidth occupancy and improve video transmission efficiency. After the initial multicast stream is transcoded, the generated transcoded multicast stream is registered with the IP address through the multicast transcoding server in the first area, and is registered with the multicast convergence point in the first area to obtain local registration data. After receiving the transcoded multicast stream, the multicast transcoding server allocates a unique multicast address to the stream in the local multicast routing table, and publishes the registration information of the multicast stream to the multicast convergence point. This process ensures that other devices in the network can know the existence of the multicast stream and can correctly subscribe to and forward the stream. The multicast convergence point is responsible for receiving streams from different multicast sources and forwarding the streams to the required users and devices according to the registration data. According to the local registration data, the multicast stream is synchronously transmitted across regions through the MSDP (Multicast Source Discovery Protocol) peer mechanism. The MSDP mechanism allows multicast convergence points in different areas to communicate with each other and exchange information about multicast sources, thereby realizing synchronous transmission of multicast streams between regions. The multicast convergence point in the first area will pass the information of the transcoded multicast stream to the multicast convergence point in the second area. The MSDP mechanism establishes an MSDP peer connection to synchronously transmit the multicast stream in the first area to the multicast convergence point in the second area, thereby ensuring that the multicast streams in the two areas are consistent and realizing remote disaster recovery transmission. After the synchronized multicast stream arrives at the second area, the IP address is registered on the multicast transcoding server in the second area, and the multicast convergence point in the second area is registered to obtain remote registration data. The multicast transcoding server registers the synchronized multicast stream to the local routing table and publishes the registration information to the multicast convergence point so that devices in other areas can receive the synchronized stream. The multicast convergence point serves as the distribution center of the multicast stream and distributes the stream to all users and devices subscribed to the multicast address according to the registration information. The synchronized multicast stream is transcoded twice to generate a multicast stream in the same format as the first area to obtain a remote multicast stream.The purpose of the secondary transcoding process is to ensure that the final transmitted multicast stream is consistent with the multicast stream in the first area in terms of format and quality, even in different areas, to achieve seamless service connection. For example, assuming that the multicast stream in the first area uses H.264 encoding and 1080p resolution, after secondary transcoding, it is ensured that the multicast stream in the second area is also transcoded with H.264 encoding and 1080p resolution. The generated remote multicast stream is input to the CDN server in the second area for multicast service recording and storage. The CDN server not only needs to cache the real-time multicast stream, but also needs to record and store the stream for subsequent playback, on-demand and other services. In the backup process, by recording and storing the remote multicast stream in the CDN server, it is ensured that when the multicast stream in the first area fails, the backup remote multicast stream can be quickly replaced, thereby providing high-availability services. The (S, G) table entry of the multicast transcoding server is uniquely verified and confirmed according to the local registration data and the remote registration data to ensure the correct transmission of the multicast service data. The (S, G) entry in the multicast routing table is used to identify the mapping relationship between a specific source (S) and a multicast group address (G). By verifying the uniqueness of these entries, it is ensured that all multicast streams can be correctly routed and forwarded according to their source address and target group address, avoiding multicast stream conflicts or incorrect stream forwarding problems. During the transmission of multicast service data, S represents the source address of the multicast stream, and G represents the target group address for receiving the stream. Through this mechanism, the distribution path of the traffic is accurately controlled to ensure that users in each area can receive accurate multicast data.

[0127] In a specific embodiment, the process of executing step S5 may specifically include the following steps:

[0128] Extract features from multicast service data, separate the load status of RPs in each region, CPU occupancy rate, memory usage rate, network bandwidth utilization rate, and user access volume distribution data of multicast transcoding servers, and construct a multi-dimensional service feature vector.

[0129] The multi-dimensional business feature vector is input into the state encoding layer of the upper strategy network in the two-layer proxy model, and the time series feature is extracted through the three-layer long short-term memory network in the state encoding layer. Each layer of the long short-term memory network contains 128 hidden units to obtain the time series state representation;

[0130] The temporal state representation is processed through the decision layer of the upper-layer strategy network. The decision layer includes a bidirectional attention mechanism and a three-layer fully connected network, and outputs the global decision results of the service switching strategy, load balancing strategy, and resource allocation strategy.

[0131] The global decision result is input into the action generation layer of the upper policy network. The action generation layer adopts the Actor-Critic architecture and generates a decision action sequence through interactive optimization of the policy network and the value network.

[0132] The decision action sequence is decomposed into tasks and processed by the task planning layer of the lower execution network in the two-layer proxy model. The task planning layer contains a graph convolutional network, which analyzes the dependency relationship of routing switching instructions, VPN configuration instructions, and multicast stream switching instructions to obtain an execution dependency graph.

[0133] The execution dependency graph is input into the parameter generation layer of the lower execution network. The parameter generation layer contains five layers of residual blocks. Each residual block contains two convolutional layers and a short-circuit connection to generate a parameter configuration scheme.

[0134] Based on the parameter configuration scheme, it is processed by the execution control layer of the lower execution network. The execution control layer adopts a reinforcement learning algorithm, takes NQA detection data, IGP routing status, BGP routing status and multicast flow status as environmental feedback, dynamically adjusts the execution strategy, and obtains an intelligent scheduling strategy.

[0135] Specifically, feature extraction is performed on multicast service data to separate multidimensional data such as the load status of RPs in each region, CPU occupancy rate of multicast transcoding servers, memory usage rate, network bandwidth utilization rate, and user access volume distribution, and a multidimensional service feature vector is constructed. [Load status, CPU occupancy, memory occupancy, bandwidth utilization, user access distribution]. The feature vector is input into the state encoding layer of the upper policy network in the two-layer proxy model, and the time series feature is extracted through the three-layer long short-term memory network (LSTM) in this layer. Each LSTM layer contains 128 hidden units, and the update formula of LSTM is:

[0136] ;

[0137] ;

[0138] ;

[0139] ;

[0140] ;

[0141] in, They are input gate, forget gate and output gate. is the cell state, is the output state. The output of LSTM As a representation of the time series state, it is passed to the decision layer for processing. In the decision layer of the upper policy network, an intelligent scheduling strategy is generated through a structure including a bidirectional attention mechanism and a three-layer fully connected network, which includes service switching strategy, load balancing strategy and resource allocation strategy. Assume that the output of the decision layer is a vector [Business switching strategy, load balancing strategy, resource allocation strategy], these strategies are passed as decision results to the lower-level execution network for subsequent task planning. In the action generation layer, combined with the Actor-Critic architecture, the decision action sequence is generated by alternating optimization strategies (Actor) and value functions (Critic). The optimization goal of the Actor-Critic method is to update the policy parameters through the policy gradient method, and its formula is: ;

[0142] in, is the output of the policy network in state Next select action The probability of is the advantage function, which represents the choice of action Relative to the current strategy in state Advantage under The calculation formula is: ;

[0143] in, is the state-action value function, is the state value function. Through the interactive optimization of Actor-Critic, the strategy parameters generated by Actor The value function estimated by the Critic will be updated together to generate a sequence of decision actions In the lower execution network, the task planning layer analyzes the dependency relationship between routing switching instructions, VPN configuration instructions, and multicast stream switching instructions through a graph convolutional network (GCN) to obtain an execution dependency graph. The basic operation formula of the graph convolutional network is: ;

[0144] in, and The nodes are The input and output characteristics of Is a node The set of neighbor nodes of is the weight matrix of the graph convolution operation. The result of dependency analysis generates an execution dependency graph, which is used to guide the subsequent execution control layer. The execution control layer dynamically adjusts task execution through reinforcement learning algorithms and optimizes based on feedback from NQA detection data, IGP routing status, BGP routing status, and multicast flow status. The specific reinforcement learning update formula is: ;

[0145] in, Is the current state and actions The state action value of is the learning rate, is the discount factor. Based on the parameter configuration scheme and intelligent scheduling strategy, the execution control layer dynamically adjusts the execution strategy of the task to ensure that the multicast services in each area can be carried out efficiently and stably. The whole process forms an adaptive and optimized scheduling system based on deep reinforcement learning.

[0146] In a specific embodiment, the process of executing step S6 may specifically include the following steps:

[0147] Based on the intelligent scheduling strategy, the service switching rules, load threshold parameters and resource allocation coefficients are extracted to obtain the scheduling control parameter set;

[0148] Based on the scheduling control parameter set, geographical partition calculation is performed on the service coverage of the first area and the second area to generate a user service strategy;

[0149] According to the user service policy, user live broadcast service access configuration is performed on the multicast aggregation points in the first area and the second area respectively to obtain a load sharing solution;

[0150] A backup path is configured for the BGP route between the first area and the second area, and the route weight is dynamically adjusted through an intelligent scheduling strategy to obtain a route disaster recovery mechanism;

[0151] Based on the routing disaster recovery mechanism, a cross-region backup configuration is performed on the multicast transcoding servers in the first region and the second region to obtain a backup switching strategy;

[0152] Deploy the backup switching strategy to the CDN servers in the first and second regions, configure active-active disaster recovery for the recording service, and obtain a storage backup mechanism;

[0153] According to the storage backup mechanism, an independent multicast stream source is introduced to the second area, and recording synchronization is performed through the CDN server to obtain a peer-to-peer backup structure;

[0154] Based on the peer-to-peer backup structure, the load sharing solution, routing disaster recovery mechanism and storage backup mechanism are integrated and associated to generate a remote disaster recovery transmission solution.

[0155] Specifically, the scheduling control parameters are extracted through the intelligent scheduling strategy, including service switching rules, load threshold parameters and resource allocation coefficients. The load threshold parameters and resource allocation coefficients are set to measure the resource allocation efficiency and availability of each region and each node under different workloads. These parameters constitute the scheduling control parameter set. The scheduling control parameter set is expressed as ,in is the load threshold, is the resource allocation coefficient, is the service switching rule. Based on the scheduling control parameter set, the service coverage of the first area and the second area is geographically partitioned and a user service policy is generated. The goal of geographical partitioning is to ensure that the service coverage can be efficiently allocated to each area so that users in each area can access the service through the optimal access path. The partitioning process is completed by comprehensively analyzing the regional load and resource availability, and the calculation result is expressed as a user service policy vector To indicate that and Represent the user service coverage strategy of the first area and the second area respectively. According to the service strategy, the multicast convergence points of the first area and the second area are configured to provide access points for the live broadcast services of users in each area, so as to realize the generation of load sharing solutions. The backup path is configured for the BGP route between the first area and the second area, and the route weight is dynamically adjusted through the intelligent scheduling strategy to build a route disaster recovery mechanism. Assume that the route weight is ,This weight is closely related to the routing reliability and bandwidth of the network. The adjustment formula of BGP routing weight is: reliability bandwidth;

[0156] in, and It is used to adjust the importance coefficient of route reliability and bandwidth. Reliability represents the reliability score of the route, and bandwidth represents the bandwidth index of the path. Parameters, automatically select the optimal routing path under different network conditions to ensure high availability of services. Configure the multicast transcoding servers in the first and second regions for cross-regional backup, form a backup switching strategy, and ensure that when a path fails, it can switch to the backup path in time to continue business processing. In order to ensure the efficiency of backup switching, the configuration of the multicast transcoding server needs to dynamically adjust its IP address and routing configuration, and optimize it by continuously monitoring the network status and calculating the load and performance of the routing path. Set the target path in the backup switching strategy to , which is a redundant path generated based on the current network topology and load status. On this basis, in order to ensure the high availability of the recording service, the backup switching strategy is deployed to the CDN servers in the first and second regions, and the recording service is configured with active-active disaster recovery to generate a storage backup mechanism. The core of this mechanism is to synchronize data to the CDN servers in the two regions in real time to ensure that the data can be seamlessly restored even if one of the servers fails. The storage backup mechanism is expressed by the following formula: ;

[0157] in, Indicates backup data. and Represents the data on the CDN servers in the first area and the second area respectively. Through real-time synchronization, redundant storage of data is ensured and the disaster recovery capability of the system is improved. After the storage backup mechanism is deployed, an independent multicast stream source is introduced to the second area according to the mechanism, and recording synchronization is performed through the CDN server to obtain a peer-to-peer backup structure. The peer-to-peer backup structure is designed to ensure that the data source and backup source in each area can be independent of each other and back up each other, thereby avoiding the impact of single point failures on the system. On this basis, the synchronization mechanism between CDN servers will be dynamically adjusted according to the real-time network conditions to provide highly reliable recording synchronization. Through the peer-to-peer backup structure, the load sharing solution, routing disaster recovery mechanism and storage backup mechanism are integrated and associated to form a complete off-site disaster recovery transmission solution.

[0158] The above describes the IPTV multicast service remote disaster recovery transmission method in the embodiment of the present invention. The following describes the IPTV multicast service remote disaster recovery transmission device in the embodiment of the present invention. Figure 2 In one embodiment of the present invention, an IPTV multicast service remote disaster recovery transmission device includes:

[0159] A building module, used to respectively use a multicast convergence point and a multicast transcoding server to build a networking architecture for the first area and the second area, configure the multicast transcoding servers in the first area and the second area to have the same IP address, and obtain an initial two-node multicast network;

[0160] The detection module is used to dynamically detect the multicast transcoding server in ICMP mode according to the initial two-node multicast network, establish a routing target strategy through IGP and BGP protocols, and obtain a multicast service routing mechanism;

[0161] A binding module is used to create a VPN instance for the multicast service routing mechanism and bind the CDN server and the multicast transcoding server to the corresponding VPN instance to obtain a service isolation environment;

[0162] The transmission module is used to transcode the multicast stream received in the first area and synchronize it to the second area based on the service isolation environment, realize the transmission of the multicast stream between the two areas through the MSDP peer mechanism, and obtain the multicast service data;

[0163] The scheduling module is used to input the multicast service data into the two-layer proxy model, and perform deep reinforcement learning on the service scheduling through the upper-layer policy network and the lower-layer execution network in the two-layer proxy model to obtain an intelligent scheduling strategy;

[0164] The generation module is used to perform load sharing calculation and establish a dual-active backup mechanism for the first area and the second area according to the intelligent scheduling strategy, and generate a remote disaster recovery transmission plan.

[0165] Through the collaborative cooperation of the above-mentioned components, by deploying multicast transcoding servers and RP nodes in different locations and using the same IP address configuration, the single point failure problem is solved, the system reliability is significantly improved, and the interruption of the entire network service caused by equipment failure is avoided; the NQA function and the IGP / BGP hybrid routing mechanism are introduced to realize accurate server status monitoring and intelligent routing switching, and ensure the timeliness and accuracy of service switching; the VPN isolation technology is used to logically isolate the CDN server from the multicast transcoding server, which effectively solves the continuity problem of the CDN recording service and ensures the reliability of value-added services such as playback time shift; a two-layer proxy model is used for service scheduling, and intelligent load balancing and fault recovery are realized through deep reinforcement learning, which greatly improves the automation level of the system; a complete remote disaster recovery solution is constructed, which realizes dual-active deployment and dynamic load sharing of the service, and significantly improves the overall service quality of the IPTV live broadcast service; the present invention does not require additional dedicated switching equipment, and remote disaster recovery can be achieved purely through network configuration, which reduces deployment costs and simplifies operation and maintenance difficulties.

[0166] Reference Figure 3 In an embodiment of the present invention, a computer device is also provided. The computer device may be a server, and its internal structure may be as follows: Figure 3 As shown. The computer device includes a processor, a memory, a display screen, an input device, a network interface and a database connected through a system bus. Among them, the processor designed by the computer is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the corresponding data in this embodiment. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the above method is implemented.

[0167] Those skilled in the art will understand that Figure 3 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied.

[0168] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the above method is implemented. It can be understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0169] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided by the present invention and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double-speed data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM.

[0170] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0171] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the whole or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program code.

[0172] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for remote disaster recovery transmission of IPTV multicast services, characterized in that: The method comprises: A multicast convergence point and a multicast transcoding server are respectively used to build a networking architecture for the first area and the second area, and the multicast transcoding servers in the first area and the second area are configured with the same IP address to obtain an initial two-node multicast network; According to the initial two-node multicast network, the multicast transcoding server is dynamically detected by ICMP, and a routing target strategy is established through IGP and BGP protocols to obtain a multicast service routing mechanism; Creating a VPN instance for the multicast service routing mechanism, and binding the CDN server and the multicast transcoding server to the corresponding VPN instance to obtain a service isolation environment; Based on the service isolation environment, transcode the multicast stream received in the first area and synchronize it to the second area, and realize the transmission of the multicast stream between the two areas through the MSDP peer mechanism to obtain multicast service data; The multicast service data is input into a two-layer proxy model, and deep reinforcement learning is performed on the service scheduling through the upper-layer strategy network and the lower-layer execution network in the two-layer proxy model to obtain an intelligent scheduling strategy; specifically, the following steps are performed: feature extraction is performed on the multicast service data to separate the load status of the RP in each area, the CPU occupancy rate, memory usage rate, network bandwidth utilization rate and user access volume distribution data of the multicast transcoding server, and a multidimensional service feature vector is constructed; the multidimensional service feature vector is input into the state encoding layer of the upper-layer strategy network in the two-layer proxy model, and time series feature extraction is performed through the three-layer long short-term memory network in the state encoding layer, and each layer of the long short-term memory network contains 128 hidden units to obtain a time series state representation; the time series state representation is processed through the decision layer of the upper-layer strategy network, and the decision layer contains a bidirectional attention mechanism and a three-layer fully connected network, and outputs the global decision results of the service switching strategy, load balancing strategy and resource allocation strategy; the global decision result is input into The action generation layer of the upper-layer policy network adopts the Actor-Critic architecture, and generates a decision action sequence through interactive optimization of the policy network and the value network; the decision action sequence is task-decomposed and processed by the task planning layer of the lower-layer execution network in the two-layer proxy model, the task planning layer includes a graph convolutional network, and the dependency analysis of the routing switching instructions, VPN configuration instructions and multicast stream switching instructions is performed to obtain an execution dependency graph; the execution dependency graph is input into the parameter generation layer of the lower-layer execution network, the parameter generation layer includes five layers of residual blocks, each residual block includes two convolutional layers and a short-circuit connection, and a parameter configuration scheme is generated; based on the parameter configuration scheme, it is processed by the execution control layer of the lower-layer execution network, the execution control layer adopts a reinforcement learning algorithm, and takes NQA detection data, IGP routing status, BGP routing status and multicast stream status as environmental feedback to dynamically adjust the execution strategy to obtain an intelligent scheduling strategy; According to the intelligent scheduling strategy, load sharing calculation and active-active backup mechanism are performed on the first area and the second area to generate a remote disaster recovery transmission solution.

2. The IPTV multicast service remote disaster recovery transmission method according to claim 1, characterized in that: The method of respectively using a multicast convergence point and a multicast transcoding server to construct a networking architecture for the first area and the second area, configuring the multicast transcoding servers in the first area and the second area to have the same IP address, and obtaining an initial two-node multicast network includes: Physical port configuration is performed on the multicast convergence points of the first area and the second area, and transmission network connection channels are established respectively to obtain a regional interconnection structure; Based on the regional interconnection structure, MSDP protocol configuration and peer connection relationship establishment are performed on the multicast convergence points of the first area and the second area to obtain a multicast synchronization mechanism; Performing an initial multicast stream receiving configuration on the multicast convergence point in the first area according to the multicast synchronization mechanism, setting the multicast convergence point in the second area as a synchronization receiving end, and obtaining a multicast stream transmission path; uniformly assigning IP addresses to multicast transcoding servers in the first area and the second area in the multicast stream transmission path to generate transcoding server configurations; Based on the transcoding server configuration, CDN server binding and real-time recording link establishment are performed on the multicast transcoding servers in the first area and the second area respectively to obtain a program storage mechanism; Deploy the program storage mechanism in the CDN servers in the first area and the second area, configure playback and time-shift functions, and obtain recording service configuration; Dividing the recording service configuration into service zones, setting the multicast convergence points of the first zone and the second zone as live broadcast service access nodes for users in the respective zones, and obtaining a service zone strategy; According to the service partitioning strategy, the same external service address is configured for the multicast convergence points in the first area and the second area to generate an initial two-node multicast network.

3. The IPTV multicast service remote disaster recovery transmission method according to claim 2, characterized in that: The method of dynamically detecting the multicast transcoding server in an ICMP manner according to the initial two-node multicast network, establishing a routing target strategy through IGP and BGP protocols, and obtaining a multicast service routing mechanism includes: Performing NQA function configuration on the egress switch in the initial dual-node multicast network to obtain a network quality analysis mechanism; Based on the network quality analysis mechanism, performing survival status detection on the IP addresses of the multicast transcoding servers in the first area and the second area to generate multicast server monitoring results; According to the monitoring result of the multicast server, an OSPF neighbor relationship is established between the multicast convergence point in the first area and the multicast convergence point in the second area and the corresponding egress switch to obtain an IGP routing connection; Based on the IGP routing connection, the egress switch is configured with detailed routing of the multicast transcoding server, and the local routing is published to the multicast convergence point through the IGP routing to obtain the intra-regional routing strategy; Configure BGP neighbor relationship between the multicast convergence points of the first area and the second area, establish a BGP connection channel, and mutually publish detailed routes of multicast transcoding servers to obtain an inter-area routing strategy; The priority parameter of the intra-area routing policy is set higher than the priority parameter of the inter-area routing policy, and the routing target is configured for the multicast convergence point to obtain a routing priority mechanism; According to the routing priority mechanism, the routing table of the multicast convergence point is prioritized for routing selection, the local IGP routing is used as the preferred routing path, and a routing switching strategy is obtained; Based on the routing switching strategy, the routing table of the multicast convergence point is dynamically updated, and when the local routing fails, it automatically switches to the BGP routing to obtain a multicast service routing mechanism.

4. The IPTV multicast service remote disaster recovery transmission method according to claim 3, characterized in that: The VPN instance is created for the multicast service routing mechanism, and the CDN server and the multicast transcoding server are bound to the corresponding VPN instance to obtain a service isolation environment, including: Create a vlanif interface for the physical ports of the egress switches and the multicast convergence point in the first area and the second area to obtain a service bearer channel; Based on the service bearer channel, marking and dividing ports of CDN server uplink egress switches in the first area and the second area to obtain CDN access links; According to the CDN access link, the vlanif interface and the CDN server are associated and bound with a VPN instance, and a multicast protocol is enabled in the VPN to obtain a VPN multicast domain; Perform routing configuration on the VPN multicast domain, point the default route to the multicast convergence point of the corresponding area, force the original tree of the CDN server to bypass the multicast convergence point, and obtain the VPN routing strategy; Based on the VPN routing policy, the CDN servers and the multicast transcoding servers in the first area and the second area are isolated and configured to obtain a logical isolation structure; Associating the VPN instance in the logical isolation structure with an NQA detection result, synchronizing the detailed routing state of the multicast transcoding server to the VPN routing table, and obtaining a routing synchronization mechanism; Based on the routing synchronization mechanism, the VPN routing table is dynamically updated, and when the local detailed routing fails, it automatically switches to the BGP routing to obtain the VPN switching strategy, and according to the VPN switching strategy, the recording service of the CDN server is continuously protected to obtain a service isolation environment.

5. The IPTV multicast service remote disaster recovery transmission method according to claim 4, characterized in that: Based on the service isolation environment, transcoding the multicast stream received in the first area and synchronizing it to the second area, and transmitting the multicast stream between the two areas through the MSDP peer mechanism to obtain multicast service data, including: Performing code stream parsing and transcoding parameter configuration on the initial multicast stream received in the first area to obtain a transcoding processing task, and performing transcoding processing on the initial multicast stream based on the transcoding processing task to generate a transcoded multicast stream; Registering the transcoded multicast stream with the multicast convergence point in the first area through the multicast transcoding server in the first area to obtain local registration data; According to the local registration data, the multicast stream is synchronously transmitted across regions through the MSDP peer mechanism, and the transcoded multicast stream is synchronized to the multicast convergence point of the second region to obtain a synchronous multicast stream; Based on the synchronous multicast stream, register the IP address of the multicast transcoding server in the second area, register with the multicast convergence point in the second area, and obtain remote registration data; Perform secondary transcoding processing on the synchronous multicast stream to generate a multicast stream in the same format as that in the first area to obtain a remote multicast stream, and input the remote multicast stream into a CDN server in the second area to record and store multicast services to obtain backup multicast data; According to the local registration data and the remote registration data, the (S, G) table entry of the multicast transcoding server is uniquely verified and confirmed to obtain multicast service data.

6. The IPTV multicast service remote disaster recovery transmission method according to claim 1, characterized in that: The performing load sharing calculation and establishing a dual-active backup mechanism for the first area and the second area according to the intelligent scheduling strategy to generate a remote disaster recovery transmission solution includes: Based on the intelligent scheduling strategy, extracting service switching rules, load threshold parameters and resource allocation coefficients to obtain a scheduling control parameter set; Based on the scheduling control parameter set, performing geographical partition calculation on the service coverage of the first area and the second area to generate a user service policy; According to the user service policy, user live broadcast service access configuration is performed on the multicast aggregation points in the first area and the second area respectively to obtain a load sharing solution; A backup path is configured for the BGP route between the first area and the second area, and the route weight is dynamically adjusted through the intelligent scheduling strategy to obtain a route disaster recovery mechanism; Based on the routing disaster recovery mechanism, cross-region backup configuration is performed on the multicast transcoding servers in the first region and the second region to obtain a backup switching strategy; Deploy the backup switching strategy to the CDN servers in the first area and the second area, perform active-active disaster recovery configuration on the recording service, and obtain a storage backup mechanism; According to the storage backup mechanism, an independent multicast stream source is introduced into the second area, and recording synchronization is performed through the CDN server to obtain a peer-to-peer backup structure; Based on the peer-to-peer backup structure, the load sharing solution, the routing disaster recovery mechanism and the storage backup mechanism are integrated and associated to generate a remote disaster recovery transmission solution.

7. A remote disaster recovery transmission device for IPTV multicast services, characterized in that: Used to execute the IPTV multicast service remote disaster recovery transmission method according to any one of claims 1 to 6, the IPTV multicast service remote disaster recovery transmission device comprises: A building module, used to respectively use a multicast convergence point and a multicast transcoding server to build a networking architecture for the first area and the second area, configure the multicast transcoding servers in the first area and the second area with the same IP address, and obtain an initial two-node multicast network; A detection module is used to dynamically detect the multicast transcoding server in an ICMP manner according to the initial two-node multicast network, establish a routing target strategy through IGP and BGP protocols, and obtain a multicast service routing mechanism; A binding module is used to create a VPN instance for the multicast service routing mechanism and bind the CDN server and the multicast transcoding server to the corresponding VPN instance to obtain a service isolation environment; A transmission module, configured to transcode the multicast stream received in the first area and synchronize it to the second area based on the service isolation environment, and realize the transmission of the multicast stream between the two areas through the MSDP peer mechanism to obtain multicast service data; A scheduling module, used for inputting the multicast service data into a double-layer proxy model, performing deep reinforcement learning on service scheduling through an upper-layer policy network and a lower-layer execution network in the double-layer proxy model, and obtaining an intelligent scheduling strategy; A generation module is used to perform load sharing calculation and establish a dual-active backup mechanism for the first area and the second area according to the intelligent scheduling strategy, and generate a remote disaster recovery transmission plan.

8. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and is characterized in that when the processor executes the computer program, the method for disaster recovery transmission of IPTV multicast services in different locations as described in any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the processor is enabled to execute the IPTV multicast service disaster recovery transmission method according to any one of claims 1 to 6.

Images