A method, system, device and medium for differentiated multi-stream video transmission

Abstract

The application relates to the technical field of video transmission, and discloses a differentiated multi-stream video transmission method, system, device and medium. The method is applied to an edge computing node and comprises the following steps: receiving video content requests of a plurality of mobile terminal users carrying video content individualization demands; obtaining a standard video stream of the requested video content from a cloud server, wherein if the same video content is requested by different mobile terminal users, only one standard video stream is obtained; instantiating logical processing containers in real time locally according to all the individualization demands; processing the standard video stream through a service function chain composed of one or more logical processing containers to generate differentiated video streams satisfying the corresponding individualization demands of each mobile terminal user, wherein each logical processing container corresponds to an individualization operation of a standard video stream; and sending each differentiated video stream to the corresponding mobile terminal user, thereby solving the problem of downlink transmission redundancy when multiple users concurrently request the same video content.

Description

Technical Field

[0001] This invention relates to the field of video transmission technology, and in particular to a differentiated multi-stream video transmission method, system, device, and medium. Background Technology

[0002] Edge computing, as a novel computing paradigm, has become a key technology supporting high-bandwidth, low-latency applications such as high-definition video surveillance, real-time streaming media, and VR / AR by moving computing, storage, and network resources from the cloud center to the network edge. The core of this technology lies in providing localized services using edge nodes close to the data source, thereby alleviating transmission pressure on the core network and improving user experience. However, in real-world edge computing environments, computing nodes are typically heterogeneous and resource-constrained, and the network topology is highly dynamic and uncertain. Therefore, achieving a dynamic balance across layers and domains between energy consumption, latency, computing costs, and user experience while meeting the demands of high-definition video processing is a critical issue that urgently needs to be addressed in this field.

[0003] To improve video processing efficiency in edge environments, current mainstream technologies primarily focus on computation offloading strategies and resource scheduling algorithms. These technologies typically dynamically migrate video processing tasks (such as transcoding, analysis, and rendering) from terminal devices to edge servers or cloud servers. To address environmental uncertainties, existing research often employs optimization-based or online learning-based methods to intelligently decide the execution location and resource allocation of tasks, aiming to achieve a balance between high video quality and low service migration costs. Furthermore, some technologies introduce cloud-edge or edge-edge collaborative architectures, using multi-node collaboration to compensate for the insufficient computing power of a single edge node, thereby supporting complex video analytics applications.

[0004] For example, a proposed online algorithm for scheduling cloud-edge collaborative real-time video stream analysis tasks constructs a cloud-edge collaborative framework, in which the edge server acts as an intermediary between the video source and the cloud center, prioritizing the offloading of computing tasks to nodes closer to the video source. However, this method mainly optimizes the vertical offloading decision from the source to the cloud for a single video stream, neglecting the downlink transmission redundancy problem when multiple users concurrently request the same video content. In scenarios such as live streaming or monitoring, the core network still needs to bear the transmission pressure of a large amount of duplicate data, resulting in a waste of bandwidth resources. Summary of the Invention

[0005] The purpose of this invention is to provide a differentiated multi-stream video transmission method, system, device, and medium that can solve the video stream transmission problem in the prior art when multiple users concurrently request the same video content.

[0006] To address the aforementioned technical problems, embodiments of the present invention provide a differentiated multi-stream video transmission method applied to edge computing nodes, comprising the following steps: Receive video content requests sent by multiple mobile terminal users, wherein the video content requests carry the personalized needs of the mobile terminal users for video content; Obtain the standard video stream corresponding to the video content requested by each mobile terminal user from the cloud server; if different mobile terminal users request the same video content, only one standard video stream of the corresponding video content is obtained. Based on service function chain technology, logical processing containers are instantiated locally in real time according to the personalized needs of each mobile terminal user. The standard video stream is processed through a service function chain consisting of one or more logical processing containers to generate a differentiated video stream that meets the corresponding personalized needs of each mobile terminal user. Each logical processing container corresponds to a personalized operation of a standard video stream. Each differentiated video stream is sent to the corresponding mobile terminal user.

[0007] Furthermore, the edge computing node stores a service deployment mapping table and a user reference count table; Before instantiating the logic processing container locally in real time according to the personalized needs of each mobile terminal user based on the service function chain technology, the following is also included: For each mobile terminal user's personalized needs, the service deployment mapping table is queried to check whether the service function chain that meets the corresponding personalized needs is already running on the current edge computing node; the service deployment mapping table stores the running location of each logical processing container and the service function chain to which it belongs; If so, the corresponding personalized service function chain is reused to process the standard video stream, and the reference count is added to the user reference count table.

[0008] Furthermore, before instantiating the logic processing container locally in real time according to the personalized needs of each mobile terminal user based on the service function chain technology, the following is also included: With the objective of minimizing the total resource consumption when processing standard video streams according to the personalized needs of all mobile terminal users, a scheduling optimization problem is constructed as follows: ; In the formula, The resource overhead incurred for all service function chains deployed on edge computing nodes. For a set of service function chains, Let F be the chain index, and F be the set of logical processing containers in the service function chain. Container index; The resource overhead of offloading tasks to cloud servers or edge computing nodes has been quantified. This indicates that the logic processing container must be deployed on a cloud server or edge computing node; This means that the same logical processing container in the same service function chain can be deployed on at most one edge computing node; This indicates that the resource overhead incurred by the task on the cloud server does not exceed the resource threshold of the cloud server. , It refers to the resource overhead generated by the logical processing container f in the service function chain d during runtime; This indicates that the resource overhead generated by all service function chains deployed on edge computing nodes does not exceed the resource threshold of the edge computing nodes. ; For decision variables, the value is 1 if the logical processing container f of the service function chain d is deployed on the edge computing node s, and 0 otherwise. Solve the scheduling optimization problem to determine whether to process the standard video stream via edge computing nodes or offload it to a cloud server.

[0009] Furthermore, solving the scheduling optimization problem includes: By using non-negative Lagrange multipliers, the global coupling constraints in the scheduling optimization problem are relaxed and incorporated into the objective function as penalty terms, thereby decoupling the scheduling optimization problem and obtaining a resource allocation subproblem for cloud servers or each edge computing node, as follows: ; In the formula, For Lagrange multipliers, The resource overhead incurred by unloading the logic processing container f in the service function chain d to the cloud server. This is a constant term in the objective function of the scheduling optimization problem; Solve each resource allocation subproblem.

[0010] Furthermore, solving each resource allocation subproblem includes: An iterative solution method based on projected subgradients is used to iteratively update the Lagrange multipliers. The solution to each resource allocation subproblem is obtained by continuously updating the upper and lower bounds.

[0011] Furthermore, the cloud server stores video streams corresponding to various video content obtained from different data acquisition devices. By converting all video streams into a unified video stream with the same bit rate, resolution, and encoding format, a standard video stream corresponding to various video content is obtained.

[0012] Furthermore, after acquiring video streams corresponding to various video contents, the data acquisition device adjusts the bitrate of each video stream using a standardized compression algorithm, processes the video content using a visual quality preservation algorithm, and then sends the video streams to the cloud server.

[0013] Embodiments of the present invention also provide a differentiated multi-stream video transmission system, comprising: The request receiving module is used to receive video content requests sent by multiple mobile terminal users, wherein the video content requests carry the personalized requirements of the mobile terminal users for video content. The video acquisition module is used to obtain the standard video stream corresponding to the video content requested by each mobile terminal user from the cloud server; if different mobile terminal users request the same video content, only one standard video stream of the corresponding video content is obtained. The video processing module is used to instantiate logical processing containers locally in real time according to the personalized needs of each mobile terminal user based on service function chain technology, and process the standard video stream through the service function chain consisting of one or more logical processing containers to generate a differentiated video stream that meets the corresponding personalized needs of each mobile terminal user; wherein, each logical processing container corresponds to a personalized operation of a standard video stream. The video sending module is used to send each differentiated video stream to the corresponding mobile terminal user.

[0014] Embodiments of the present invention also provide a computer device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the above-described differentiated multi-stream video transmission method.

[0015] Embodiments of the present invention also provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described differentiated multi-stream video transmission method.

[0016] The differentiated multi-stream video transmission method provided by this invention has at least the following beneficial effects: This invention constructs a novel video transmission mechanism. When faced with video content requests from multiple mobile terminal users, edge computing nodes only obtain the standard video stream corresponding to the requested video content from the cloud server. Then, using service function chain technology, corresponding logical processing containers are instantiated locally in real time according to the personalized needs of each mobile terminal user. The standard video stream is processed by the service function chain consisting of one or more logical processing containers to generate a differentiated video stream that meets the corresponding personalized needs of each mobile terminal user. Finally, each differentiated video stream is sent to the corresponding mobile terminal user.

[0017] If different mobile terminal users request the same video content (which may have different personalized needs), only one standard video stream of the corresponding video content needs to be obtained. This process only occupies one bandwidth in the core network, and personalized needs are performed locally (such as super-resolution, transcoding, filtering, etc.), which solves the data transmission redundancy problem caused by traditional methods. Attached Figure Description

[0018] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this invention, illustrate exemplary embodiments of the invention and are used to explain the invention, but do not constitute an undue limitation of the invention. In the drawings:

[0019] Figure 1 A flowchart illustrating a differentiated multi-stream video transmission method provided by the present invention; Figure 2 A schematic diagram of the system architecture for a differentiated multi-stream video transmission method provided by the present invention; Figure 3 This invention provides an experimental comparison of the proposed method with three other algorithms. Figure 1 ; Figure 4 This invention provides an experimental comparison of the proposed method with three other algorithms. Figure 2 . Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0021] The technical solutions provided by the various embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0022] One embodiment of the present invention relates to a differentiated multi-stream video transmission method. The specific process of the differentiated multi-stream video transmission method in this embodiment can be as follows: Figure 1 As shown, it includes: Step 101: Receive video content requests sent by multiple mobile terminal users, each video content request carrying the personalized needs of the mobile terminal users for video content.

[0023] Step 102: Obtain the standard video stream corresponding to the video content requested by each mobile terminal user from the cloud server; if different mobile terminal users request the same video content, only one standard video stream of the corresponding video content is obtained.

[0024] Step 103: Based on the service function chain technology, according to the personalized needs of each mobile terminal user, a logic processing container is instantiated locally in real time, and the standard video stream is processed through the service function chain consisting of one or more logic processing containers to generate a differentiated video stream that meets the corresponding personalized needs requested by each mobile terminal user; wherein, each logic processing container corresponds to a personalized operation of a standard video stream.

[0025] Step 104: Send each differentiated video stream to the corresponding mobile terminal user.

[0026] The following is a detailed description of the implementation details of the differentiated multi-stream video transmission method in this embodiment. The following content is only for the convenience of understanding the implementation details and is not necessary for implementing this solution.

[0027] Figure 2This is the overall system architecture diagram of the technical solution in this embodiment. The system constructs a server-insensitive edge computing environment based on cloud-edge-device collaboration. The main participating entities include cloud servers, edge computing nodes, and mobile terminal users. The cloud server, as the system's global data center and control hub, possesses massive computing and storage resources and is responsible for processing large-scale, non-real-time video data. After receiving raw video streams from various acquisition sources, the cloud server uses a built-in unified transcoding module to convert video sources with different formats and resolutions into a standard basic stream with a unified bitrate, resolution, and encoding format, and performs persistent storage and backup. Edge computing nodes, deployed at the network edge, are the core processing units of this system, responsible for undertaking the computing tasks delegated by the cloud server and directly responding to user requests. Each edge computing node deploys a service function chain execution environment based on containers or microservices, capable of dynamically starting or releasing video processing services according to instructions. Mobile terminal users are the consumers of the video service, responsible for sending service requests with personalized needs to the edge nodes, such as specific viewing angles, resolutions, or image quality enhancement requirements, and receiving the processed video streams for playback. In the overall architecture, cloud servers are responsible for global scheduling and content distribution, edge computing nodes are responsible for real-time processing and the final step of differentiated delivery, and mobile terminals are responsible for interaction and presentation. The three work together to solve the problem of multi-stream transmission under the limitation of core network bandwidth.

[0028] The transmission process of the entire system is described in further detail below: (1) Video stream data input: The video end of this system consists of multiple video acquisition units (such as surveillance cameras and mobile recording devices) deployed in different geographical locations. These video acquisition units (data acquisition devices) are responsible for capturing visual data from diverse scenes in real time to obtain raw video streams corresponding to various video content. During the data acquisition phase, the video streams are acquired by multiple devices in different formats. The system then performs preprocessing operations on the raw video streams, covering format standardization, temporal filtering, and preliminary bitrate control based on standardized compression algorithms. For example, the bitrate of each video stream is adjusted using standardized compression algorithms, and video content is processed using visual quality preservation algorithms. This preprocessing process aims to achieve two goals: on the one hand, reducing transmission bandwidth requirements through intelligent bitrate adjustment, and on the other hand, ensuring the integrity of effective information in the video content using visual quality preservation algorithms. After preprocessing, the system standardizes and encapsulates the video data and transmits it to the cloud server for the next stage of the system. This layered processing mechanism ensures bandwidth efficiency during video data transmission and lays a quality foundation for subsequent processing.

[0029] (2) Video processing and forwarding on cloud servers: After receiving video streams from data acquisition devices, corresponding to various video content, the cloud server converts them into a unified video stream (a standard video stream corresponding to various video content) with the same bitrate, resolution (usually 720p or 1080p), and encoding format (H.264, HEVC). The cloud server then backs up and stores the processed video stream data to prevent data corruption. Next, the cloud server distributes the video data to the corresponding edge systems according to user requests. This mechanism not only ensures efficient distribution of video data but also effectively coordinates the needs of each node in the distributed network.

[0030] (3) Edge computing node forwarding and customized video processing: When an edge computing node receives a user request, it first retrieves the raw video stream from the cloud server. Subsequently, the edge computing node processes the video in real time according to the user's specific needs (such as filter enhancement, super-resolution, or video compression) to improve content quality and user experience. In a more macro-level differentiated multi-stream transmission architecture, multiple such edge computing nodes can be deployed. Each edge computing node consists of a network device and several processing units, dedicated to meeting the diverse needs of nearby users. The core advantage of this mechanism lies in resource reuse: if the requested video stream and its processing resources are already occupied by other users, the edge computing node can directly distribute the processed stream without re-requesting from the cloud server or performing duplicate processing. For example, an edge computing node only needs to retrieve three raw videos from the cloud server to generate dozens of video streams of different specifications (such as resolution and enhancement effects) locally. When multiple users simultaneously request the same video specification, the edge computing node only needs to process it once to respond to all requests. This design enhances the capabilities of edge computing nodes in a distributed computing architecture and effectively improves the efficiency and response speed of video services through localized processing.

[0031] To achieve efficient differentiated video services with limited bandwidth resources, this embodiment designs a unique multi-stream transmission and processing flow. When multiple mobile terminal users simultaneously request the same video content but have different image quality or special effects requirements (i.e., personalized needs), the edge computing node no longer simply transmits data but acts as an intelligent processing hub. The edge computing node first requests and pulls a unique standard basic video stream from the cloud server. This process only occupies one bandwidth channel in the core network, greatly reducing data redundancy. Subsequently, the edge computing node uses service function chain technology to instantiate corresponding logical processing containers locally in real time according to the personalized needs of different users. The standard video stream is processed through a service function chain consisting of one or more logical processing containers. For example, for users requesting high-definition image quality, the edge computing node starts a super-resolution enhancement container to process the standard basic video stream; for users requesting low-bandwidth mode, it starts a compression transcoding container to process the standard basic video stream; for users requesting high-definition and low-bandwidth, it starts a service function chain containing both a super-resolution enhancement container and a compression transcoding container to process the standard basic video stream. After processing, the multiple differentiated video streams are directly distributed by the edge node to the corresponding mobile terminal users through the access network. Through this mechanism, the system ensures that users receive customized service experiences while significantly reducing the data transmission pressure from cloud servers to edge computing nodes, effectively alleviating network congestion.

[0032] In some embodiments, to efficiently manage the computing resources of edge computing nodes and support the aforementioned differentiated transmission, this embodiment designs a refined resource reuse and scheduling mechanism within the edge computing nodes. The edge computing nodes maintain three core state lists: a user access record table, a service deployment mapping table, and a user reference count table. When a new user personalization request is received, the system first queries the service deployment mapping table to check if a service function chain instance matching the user's personalization needs is already running on the current edge computing node. The service deployment mapping table not only records the running location of the function instance but also stores the complete chain structure to which the function instance belongs. The reuse mechanism is only triggered when the requested chain structure completely matches an already deployed instance. If a matching instance is found, the system directly reuses the computing resources (i.e., the service function chain corresponding to the personalization needs), distributes the processed video stream to the new user, and increments the count in the user reference count table without restarting a new container or requesting data from the cloud. If no matching instance is found, the system determines it as a miss and then decides whether to start a new service locally or offload the task to the cloud server based on the current resource status. This process is completely transparent to end users; users only need to send requests and do not need to be aware of the complex container scheduling and reuse logic in the background.

[0033] This embodiment addresses the issue of severe data transmission redundancy in high-concurrency scenarios such as live sports broadcasts and multi-camera surveillance. It utilizes service function chaining technology to offload video processing and distribution tasks from cloud servers to network edge devices. This method uses edge computing nodes to pull raw video streams in a uniform format from the cloud server and, based on the personalized needs of different users (such as image quality, resolution, specific viewing angles), dynamically performs transcoding, super-resolution, or filtering tasks in real time at the edge using containerization technology. The transmission strategy employed in this embodiment significantly reduces redundant data transmission in the core network while satisfying users' customized viewing experience, achieving highly efficient resource utilization.

[0034] In some embodiments, in existing online algorithms for scheduling cloud-edge collaborative real-time video stream analysis tasks, this scheme establishes a measurement-based analysis model to balance network latency, inference latency, and analysis accuracy in edge-based real-time video analysis systems, and utilizes Lyapunov optimization techniques to decouple long-term constraints. Based on this, the scheme designs a resolution selection and bandwidth allocation algorithm based on Simulated Annealing (SA) heuristic search, aiming to dynamically adjust video resolution and allocate bandwidth between the edge and cloud servers without needing to know future information, thereby maximizing overall analysis accuracy while ensuring stable latency. While this method can balance latency and accuracy by adjusting resolution, it employs the SA heuristic search algorithm to solve the Mixed-Integer Nonlinear Programming (MINLP) problem of task allocation (resolution configuration). Although the SA algorithm can find approximate solutions, in high-concurrency and large-state-space edge computing environments, its convergence speed and computational efficiency often fail to meet millisecond-level real-time scheduling requirements, easily introducing additional computational time. Finally, this method lacks a fine-grained resource reuse mechanism based on Service Function Chain (SFC), and cannot accurately identify and reuse existing results like this invention, resulting in less than ideal utilization of edge computing resources in multi-stream concurrent scenarios.

[0035] In video streaming processing scenarios where servers are unaware of the edge environment, the efficient deployment of service function chains involves multi-dimensional resource scheduling and node selection issues. This embodiment models the function placement and cloud offloading decisions under limited computing resources. Given the scarcity and heterogeneity of computing capacity on edge computing nodes, excessively piling tasks on the edge can lead to node load imbalance and service congestion; while frequently offloading tasks to the cloud can alleviate edge pressure, it will bring significant transmission overhead and processing costs. Therefore, under the premise of strictly adhering to the resource capacity constraints of edge devices and cloud data centers, rationally planning the execution position of each functional module in the service chain and balancing edge computing overhead and cloud service offloading costs is crucial to improving the overall service efficiency of the system.

[0036] Based on the above analysis, this embodiment constructs an optimization model (i.e., a scheduling optimization problem) with the objective of minimizing the total resource consumption when processing standard video streams according to the personalized needs of all mobile terminal users, and minimizing the total system resource consumption when executing corresponding video stream processing tasks according to all personalized needs, as follows: ; In the formula, The resource overhead incurred for all service function chains deployed on edge computing nodes. For a set of service function chains, Let F be the chain index, and F be the set of logical processing containers in the service function chain. Container index; The resource overhead of offloading tasks to cloud servers or edge computing nodes has been quantified. This indicates that the logic processing container must be deployed on a cloud server or edge computing node; This means that the same logical processing container in the same service function chain can be deployed on at most one edge computing node; This indicates that the resource overhead incurred by the task on the cloud server does not exceed the resource threshold of the cloud server. , It refers to the resource overhead generated by the logical processing container f in the service function chain d during runtime; This indicates that the resource overhead generated by all service function chains deployed on edge computing nodes does not exceed the resource threshold of the edge computing nodes. ; The value is 1 if the logical processing container f of the service function chain d is deployed on the edge computing node s, and 0 otherwise.

[0037] This objective function-driven algorithm seeks the optimal balance between the low latency advantage of edge computing nodes and the abundant resources of cloud computing. By solving this scheduling optimization problem, it can be determined whether to offload video streaming processing tasks to cloud servers.

[0038] In one example, the aforementioned optimization problem involves complex integer constraints and resource coupling constraints, making direct solutions typically computationally intensive (NP-hard). To achieve efficient edge-cloud collaborative resource allocation, this embodiment proposes a primordial-dual decomposition algorithm based on Lagrange relaxation. This method introduces Lagrange multipliers to relax the global coupling constraints in the original problem into the objective function, thereby decomposing the complex global optimization problem into a series of subproblems that can be solved in parallel.

[0039] To reduce the difficulty of solving the problem, the decision variables of the original problem are first simplified. Regarding the global coupling constraints in the equation, the global coupling constraints in the scheduling optimization problem are relaxed and incorporated into the objective function as a penalty term using a non-negative Lagrange multiplier λ (used to penalize behaviors that violate resource constraints). This decouples the scheduling optimization problem, resulting in a resource allocation subproblem for cloud servers or each edge computing node. Each resource allocation subproblem is represented by the Lagrange function and the objective function of its dual problem as shown below:

[0040] ; In the formula, For Lagrange multipliers, The resource overhead incurred by unloading the logic processing container f in the service function chain d to the cloud server. The constant term in the objective function of the scheduling optimization problem For a given multiplier The coupling between different tasks is decoupled. At this point, the problem is transformed into an independent resource allocation subproblem for each edge computing node, which can be solved efficiently in parallel using existing optimization solvers (such as PuLP). Since the dual function is usually piecewise linear and non-differentiable, the traditional gradient descent method is no longer applicable. Therefore, this embodiment uses an iterative solution method based on projected subgradients, employing iteration to update the Lagrange multipliers. By continuously updating the upper and lower bounds, we can approximate the optimal solution to the dual problem, thereby obtaining the solution to each resource allocation subproblem.

[0041] This embodiment can dynamically adjust the transmission path and resource allocation according to the real-time network status, ensuring that low-latency and highly reliable video services can still be provided when the network load fluctuates, thereby significantly improving the user experience quality of the system with limited resources.

[0042] In summary, this invention proposes a differentiated multi-stream transmission strategy and a corresponding resource scheduling algorithm based on a server-insensitive edge computing environment to overcome the shortcomings of existing technologies, such as ignoring the core network bandwidth constraints, network congestion caused by concurrent transmission of multiple video streams, and high latency in personalized service response.

[0043] Addressing the pain point of resource constraints in high-concurrency video scenarios, this invention constructs a novel transmission mechanism that changes the data transmission redundancy problem caused by centralized processing in traditional cloud services. This invention utilizes service function chain technology, enabling edge nodes to obtain only one standard basic video stream from the cloud server and then dynamically generate various differentiated video streams in real time based on the personalized needs of different users (such as super-resolution, transcoding, and filter enhancements).

[0044] To further improve resource utilization, this invention designs a resource reuse mechanism based on a chain structure. By maintaining a global service deployment mapping table, the system can accurately compare the complete topology of the service function chain when receiving a request. Once a matching active instance is found, the calculation results are directly reused for distribution, avoiding redundant calculations and redundant startup of service containers.

[0045] Furthermore, considering the heterogeneous and constrained resource characteristics of edge environments, this invention proposes a primordial-dual decomposition algorithm based on Lagrange relaxation. This algorithm decomposes complex global resource coupling constraints into parallelizable subproblems, enabling real-time and rapid decisions on whether tasks should be executed locally at the edge or offloaded to cloud servers. This minimizes overall system resource consumption while ensuring a low-latency user experience.

[0046] Compared with existing cloud computing centralized processing solutions or traditional single-task offloading technologies, the Diverse Multi-Stream Video Transmission Strategy (DMTS) proposed in this invention, based on a server-insensitive edge computing environment, can significantly resolve the contradiction between resource constraints and high-concurrency transmission of multi-stream video, and has significant advantages in resource utilization, service response speed, and system scalability.

[0047] In a specific embodiment, the present invention conducts experiments in a simulated scenario to verify its effectiveness. In the experiments, DMTS is compared with the theoretical optimal solution and three classic heuristic algorithms. The three classic heuristic algorithms are: 1. Direct Greedy Algorithm: This algorithm focuses on minimizing the response latency of a single request. When processing user requests, it traverses all edge nodes and cloud servers that meet the current resource constraints, directly selecting the node with the lowest total cost for deployment, without considering the load balancing status of the nodes. 2. Leisure Algorithm: This algorithm aims to maximize the network's load balancing capability. When making decisions, it prioritizes selecting the edge device with the largest remaining computing resources (i.e., the capacity limit minus the occupied resources) to carry the task. 3. Random Algorithm: From all available nodes (edge ​​devices or cloud servers) that meet the resource requirements of the current task, it randomly and evenly selects one for deployment. This algorithm serves as the most basic control group, used to measure the performance gain brought by the intelligent scheduling strategy.

[0048] Experimental results are as follows Figure 3 As shown, this invention's DMTS can effectively reduce service latency and improve user service quality. With increasing video processing task complexity (e.g., service function chain length increasing from 3 to 7), the latency of traditional algorithms rises sharply, while the present invention maintains a low latency. For example, when the service function chain length is 7, the average latency of the idle resource priority algorithm is as high as 993.87ms, the random selection algorithm is 840.86ms, and the greedy algorithm is 763.57ms; in contrast, the average latency of the present invention is only 596.73ms, perfectly consistent with the theoretical optimal solution. This means that when processing complex video tasks, the present invention can reduce latency by approximately 40% compared to traditional load balancing algorithms, significantly improving the user's smooth experience. Even in scenarios where the number of edge servers changes (30 to 70) or the number of concurrent service chains increases (20 to 60), the present invention still demonstrates strong stability and rationality. As the number of edge servers increases, the average latency of all methods decreases, but DMTS consistently maintains the optimal latency result and is consistent with the theory. As the number of service function chains increases, the latency of this invention and other methods increases, but it still maintains the optimal results consistent with theory.

[0049] Besides improved service quality, this invention offers advantages in the execution efficiency of the algorithm itself. In edge computing environments, the computational overhead of the resource scheduling algorithm is a key factor affecting real-time performance. See also Figure 4The experimental results show that while the theoretically optimal solution offers lower transmission latency, its solution process is extremely time-consuming. In a scenario with a service function chain length of 7, the average computation time for solving the theoretically optimal solution is approximately 7.21 ms, while the DMTS algorithm proposed in this invention requires only 0.03 ms, representing a speed improvement of approximately 240 times. In most test scenarios (such as different SFC lengths and numbers of servers), the average computation time of this invention remains stable between 0.02 ms and 0.03 ms. This indicates that this invention can complete complex resource allocation decisions in a very short time with almost no additional system time, making it highly suitable for online video streaming and monitoring / early warning scenarios with stringent real-time requirements. This demonstrates that this invention can improve user experience while optimizing execution efficiency in large-scale edge network environments.

[0050] The steps of the various methods described above are only for clarity. In practice, they can be combined into one step or some steps can be split into multiple steps. As long as they include the same logical relationship, they are all within the protection scope of this invention. Adding insignificant modifications or introducing insignificant designs to the algorithm or process, without changing the core design of the algorithm and process, are also within the protection scope of this invention.

[0051] Another embodiment of the present invention relates to a differentiated multi-stream video transmission system. The implementation details of this differentiated multi-stream video transmission system are described below. The following details are provided for ease of understanding and are not essential for implementing this solution. The differentiated multi-stream video transmission system of this embodiment includes: The request receiving module is used to receive video content requests sent by multiple mobile terminal users. The video content requests carry the personalized requirements of the mobile terminal users for the video content. The video acquisition module is used to obtain the standard video stream corresponding to the video content requested by each mobile terminal user from the cloud server; if different mobile terminal users request the same video content, only one standard video stream of the corresponding video content is obtained. The video processing module is used to instantiate logical processing containers locally in real time according to the personalized needs of each mobile terminal user based on service function chain technology, and process the standard video stream through the service function chain consisting of one or more logical processing containers to generate a differentiated video stream that meets the corresponding personalized needs of each mobile terminal user; wherein, each logical processing container corresponds to a personalized operation of a standard video stream. The video sending module is used to send each differentiated video stream to the corresponding mobile terminal user.

[0052] It is not difficult to see that this embodiment is a system embodiment corresponding to the above method embodiments, and this embodiment can be implemented in conjunction with the above method embodiments. The relevant technical details and technical effects mentioned in the above embodiments are still valid in this embodiment, and will not be repeated here to reduce repetition. Accordingly, the relevant technical details mentioned in this embodiment can also be applied to the above embodiments.

[0053] It is worth mentioning that all modules involved in this embodiment are logical modules. In practical applications, a logical unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. Furthermore, to highlight the innovative aspects of this invention, this embodiment does not introduce units that are not closely related to solving the technical problem proposed by this invention; however, this does not mean that other units are absent from this embodiment.

[0054] Another embodiment of the present invention relates to a computer device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the differentiated multi-stream video transmission method described in the above embodiments.

[0055] The memory and processor are connected via a bus, which can include any number of interconnecting buses and bridges, connecting various circuits of one or more processors and memories. The bus can also connect various other circuits, such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and will not be described further herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by the processor is transmitted over the wireless medium via an antenna, which further receives data and transmits it to the processor.

[0056] The processor manages the bus and general processing, and also provides various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory is used to store data used by the processor during operation.

[0057] Another embodiment of the present invention relates to a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the method embodiments described above.

[0058] That is, those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0059] Those skilled in the art will understand that the above embodiments are specific embodiments for implementing the present invention, and in practical applications, various changes can be made to them in form and detail without departing from the spirit and scope of the present invention.

Claims

1. A differentiated multi-stream video transmission method, characterized in that, Applied to edge computing nodes, the method includes: Receive video content requests sent by multiple mobile terminal users, wherein the video content requests carry the personalized needs of the mobile terminal users for video content; Obtain the standard video stream corresponding to the video content requested by each mobile terminal user from the cloud server; if different mobile terminal users request the same video content, only one standard video stream of the corresponding video content is obtained. Based on service function chain technology, logical processing containers are instantiated locally in real time according to the personalized needs of each mobile terminal user. The standard video stream is processed through a service function chain consisting of one or more logical processing containers to generate a differentiated video stream that meets the corresponding personalized needs of each mobile terminal user. Each logical processing container corresponds to a personalized operation of a standard video stream. Each differentiated video stream is sent to the corresponding mobile terminal user.

2. The differentiated multi-stream video transmission method according to claim 1, characterized in that, The edge computing node stores a service deployment mapping table and a user reference count table; Before instantiating the logic processing container locally in real time according to the personalized needs of each mobile terminal user based on the service function chain technology, the following is also included: For each mobile terminal user's personalized needs, the service deployment mapping table is queried to check whether the service function chain that meets the corresponding personalized needs is already running on the current edge computing node; the service deployment mapping table stores the running location of each logical processing container and the service function chain to which it belongs; If so, the corresponding personalized service function chain is reused to process the standard video stream, and the reference count is added to the user reference count table.

3. The differentiated multi-stream video transmission method according to claim 1, characterized in that, Before instantiating the logic processing container locally in real time according to the personalized needs of each mobile terminal user based on the service function chain technology, the following is also included: With the objective of minimizing the total resource consumption when processing standard video streams according to the personalized needs of all mobile terminal users, a scheduling optimization problem is constructed as follows: ； In the formula, The resource overhead incurred for all service function chains deployed on edge computing nodes. For a set of service function chains, Let F be the chain index, and F be the set of logical processing containers in the service function chain. Container index; The resource overhead of offloading tasks to cloud servers or edge computing nodes has been quantified. This indicates that the logic processing container must be deployed on a cloud server or edge computing node; This means that the same logical processing container in the same service function chain can be deployed on at most one edge computing node; This indicates that the resource overhead incurred by the task on the cloud server does not exceed the resource threshold of the cloud server. , It refers to the resource overhead generated by the logical processing container f in the service function chain d during runtime; This indicates that the resource overhead generated by all service function chains deployed on edge computing nodes does not exceed the resource threshold of the edge computing nodes. ; For decision variables, the value is 1 if the logical processing container f of the service function chain d is deployed on the edge computing node s, and 0 otherwise. Solve the scheduling optimization problem to determine whether to process the standard video stream via edge computing nodes or offload it to a cloud server.

4. The differentiated multi-stream video transmission method according to claim 3, characterized in that, The solution to the scheduling optimization problem includes: By using non-negative Lagrange multipliers, the global coupling constraints in the scheduling optimization problem are relaxed and incorporated into the objective function as penalty terms, thereby decoupling the scheduling optimization problem and obtaining a resource allocation subproblem for cloud servers or each edge computing node, as follows: ； In the formula, For Lagrange multipliers, The resource overhead incurred by unloading the logic processing container f in the service function chain d to the cloud server. This is a constant term in the objective function of the scheduling optimization problem; Solve each resource allocation subproblem.

5. The differentiated multi-stream video transmission method according to claim 4, characterized in that, Solving each resource allocation subproblem includes: An iterative solution method based on projected subgradients is used to iteratively update the Lagrange multipliers. The solution to each resource allocation subproblem is obtained by continuously updating the upper and lower bounds.

6. The differentiated multi-stream video transmission method according to claim 1, characterized in that, The cloud server stores video streams corresponding to various video content obtained from different data acquisition devices. By converting all video streams into a unified video stream with the same bit rate, resolution, and encoding format, a standard video stream corresponding to various video content is obtained.

7. The differentiated multi-stream video transmission method according to claim 6, characterized in that, After acquiring video streams corresponding to various video contents, the data acquisition device adjusts the bitrate of each video stream using a standardized compression algorithm, processes the video content using a visual quality preservation algorithm, and then sends the video streams to the cloud server.

8. A differentiated multi-stream video transmission system, characterized in that, The system includes: The request receiving module is used to receive video content requests sent by multiple mobile terminal users, wherein the video content requests carry the personalized requirements of the mobile terminal users for video content. The video acquisition module is used to obtain the standard video stream corresponding to the video content requested by each mobile terminal user from the cloud server; if different mobile terminal users request the same video content, only one standard video stream of the corresponding video content is obtained. The video processing module is used to instantiate logical processing containers locally in real time according to the personalized needs of each mobile terminal user based on service function chain technology, and process the standard video stream through the service function chain consisting of one or more logical processing containers to generate a differentiated video stream that meets the corresponding personalized needs of each mobile terminal user; wherein, each logical processing container corresponds to a personalized operation of a standard video stream. The video sending module is used to send each differentiated video stream to the corresponding mobile terminal user.

9. A computer device, characterized in that, include: At least one processor; And a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the differentiated multi-stream video transmission method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the differentiated multi-stream video transmission method as described in any one of claims 1 to 7.

Images