An ultra-low latency live streaming method, system, device, and medium based on edge computing

Abstract

This application discloses an ultra-low latency live streaming method, system, device, and medium based on edge computing, relating to the technical field of real-time audio and video interaction. The method includes: a target edge computing node receiving the original media stream; real-time monitoring of the uplink and downlink network links between the broadcaster and the target edge computing node to determine network link quality parameters; obtaining historical network link quality parameters and setting link degradation and quality standards based on these parameters; determining a target compensation strategy based on the network link quality parameters and a first and second preset conditions; and the target edge computing node processing the original media stream using the target compensation strategy to generate a compensated media stream and forwarding it to the receiving end. Adopting this scheme can improve the overall performance of ultra-low latency live streaming.

Description

Technical Field

[0001] This application relates to the technical field of real-time audio and video interaction, specifically to an ultra-low latency live streaming method, system, device, and medium based on edge computing. Background Technology

[0002] With the rapid development of internet technology and the increasing demand from users for real-time interactive experiences, ultra-low latency live streaming technology has become a core technological requirement in fields such as game streaming, sports event streaming, and online education. In these application scenarios, reducing latency directly impacts user engagement and experience quality, especially in scenarios requiring real-time interaction, where even millisecond-level latency differences can lead to a significant decline in user experience.

[0003] Currently, the industry primarily employs two error control technologies to ensure the reliability of live stream transmission: Forward Error Correction (FEC) and Automatic Repeat Request (ARQ). Forward Error Correction adds redundant information at the sending end, enabling the receiving end to detect and correct errors during transmission. Its advantage lies in not requiring feedback information, effectively reducing transmission latency. Automatic Repeat Request, on the other hand, ensures transmission reliability by having the receiving end request the sending end to retransmit lost or erroneous data packets. While it provides higher data integrity, the retransmission mechanism increases transmission latency.

[0004] However, existing error control technologies often employ a static strategy selection approach in practical applications. This means that a specific error control method is predetermined during system deployment and consistently used throughout the transmission process. This static approach has significant limitations: under favorable network conditions, excessive error control can introduce unnecessary redundancy and latency; conversely, under deteriorating network conditions, a single error control strategy may fail to provide sufficient transmission guarantees. Especially in edge computing environments with complex and variable network conditions, this lack of adaptability makes it difficult to achieve the optimal balance between latency and reliability, thus impacting the overall performance of ultra-low latency live streaming. Summary of the Invention

[0005] This application provides an ultra-low latency live streaming method, system, device, and medium based on edge computing, which improves the overall performance of live streaming.

[0006] The first aspect of this application provides an ultra-low latency live streaming method based on edge computing, specifically including:

[0007] The target edge computing node receives the original media stream sent by the broadcaster. The edge computing network includes multiple edge computing nodes, and the target edge computing node is the edge computing node in the edge computing network that is closest to the broadcaster.

[0008] The target edge computing node monitors the uplink network link between the broadcaster and the target edge computing node, and the downlink network link between the target edge computing node and the receiver in real time, and determines the network link quality parameters based on the uplink network link and the downlink network link.

[0009] The target edge computing node obtains the historical network link quality parameters of the uplink network link and the downlink network link, and sets the link degradation standard and the link excellence standard based on the historical network link quality parameters;

[0010] The target edge computing node determines the target compensation strategy based on the network link quality parameters and the first and second preset conditions. The first preset condition is that the uplink network link or the downlink network link meets the link degradation standard, and the second preset condition is that both the uplink network link and the downlink network link meet the link excellence standard.

[0011] The target edge computing node processes the original media stream using the target compensation strategy to generate a compensated media stream, and then forwards it to the receiving end.

[0012] By adopting the above technical solution, the target edge computing node can accurately grasp the dynamic changes in network conditions by monitoring the network link quality parameters of uplink and downlink network links in real time, and combining these with link degradation and quality standards set based on historical network link quality parameters. Based on this, the target edge computing node dynamically determines the target compensation strategy according to the network link quality parameters and preset conditions. This allows for timely activation of the corresponding compensation mechanism when network conditions deteriorate, and appropriate reduction of the compensation intensity when network conditions are good, thus avoiding the resource waste caused by static compensation strategies. Simultaneously, by deploying the compensation strategy selection decision on edge computing nodes closer to the broadcaster, rapid response to network fluctuations can be achieved, and the transmission overhead of compensation decisions can be reduced. This solution, by introducing an adaptive compensation mechanism based on historical data, achieves a dynamic balance between network transmission reliability and transmission latency, effectively improving the overall performance of the live broadcast.

[0013] Optionally, setting link degradation criteria and link excellence criteria based on the historical network link quality parameters includes:

[0014] The target edge computing node extracts the first historical quality data of the uplink network link and the second historical quality data of the downlink network link from the historical network link quality parameters, respectively.

[0015] First degraded data and second degraded data are selected from the first historical quality data and the second historical quality data. Based on the first degraded data and the second degraded data, a first standard score of the uplink network link and a second standard score of the downlink network link are calculated.

[0016] The first standard score of the uplink network link and the second standard score of the downlink network link are used as the link degradation criteria;

[0017] The historical time period in which both the first standard score and the second standard score exceed the preset optimal score threshold is determined as the optimal candidate time period. The first excellent threshold of the uplink network link and the second excellent threshold of the downlink network link are determined based on the first historical quality data and the second historical quality data within the optimal candidate time period.

[0018] The first good threshold of the uplink network link and the second good threshold of the downlink network link are determined as the link goodness standard.

[0019] By employing the aforementioned technical solution, and extracting first historical quality data for uplink network links and second historical quality data for downlink network links from historical network link quality parameters, targeted analysis of the historical performance of both uplink and downlink links can be performed. Based on this, by filtering the first and second degraded data and calculating the first and second standard scores, a quantitative index reflecting the degree of network link degradation is established, providing a reliable decision-making basis for subsequent compensation strategy selection. Simultaneously, this solution identifies historical time periods where both the first and second standard scores exceed a preset optimal score threshold as optimal candidate time periods, and determines the first and second excellent thresholds based on the historical quality data within these time periods, constructing a dynamically adaptive network quality evaluation standard. This two-way evaluation mechanism based on historical data can not only accurately identify network quality degradation trends but also capture network state optimization opportunities, thereby supporting the system to make more precise adjustments to compensation strategies under different network conditions, effectively improving the service stability and transmission efficiency of the ultra-low latency live streaming system.

[0020] Optionally, determining the target compensation strategy based on the network link quality parameters, the first preset condition, and the second preset condition includes:

[0021] The target edge computing node determines whether the network link quality parameters meet the first preset condition;

[0022] If the first preset condition is met, then the forward error correction strategy is determined to be the target compensation strategy;

[0023] If the first preset condition is not met, then determine whether the network link quality parameters meet the second preset condition;

[0024] If the second preset condition is met, the automatic retransmission request strategy is determined to be the target compensation strategy.

[0025] If the second preset condition is not met, then determine whether the network link quality parameters show a deteriorating trend within a preset time window;

[0026] If a deterioration trend is observed, a target compensation strategy is determined based on the network link quality parameters of other edge computing nodes in the edge computing network.

[0027] If no deterioration trend is observed, the automatic retransmission request strategy is determined as the target compensation strategy.

[0028] By adopting the above technical solution, when the target edge computing node detects that the network link quality parameters meet the first preset condition, i.e., the network has significantly deteriorated, the system will prioritize the forward error correction strategy as the target compensation strategy, thereby ensuring transmission reliability through redundant data even in poor network conditions. When the network condition is good and meets the second preset condition, the system adopts an automatic retransmission request strategy, which ensures the integrity of data transmission and avoids unnecessary bandwidth consumption. In particular, this solution also introduces a trend analysis mechanism based on a preset time window. When the network quality parameters show a deteriorating trend, the compensation strategy is determined by referring to the network link quality parameters of other edge computing nodes in the edge computing network, effectively improving the system's ability to predict network fluctuations and the accuracy of its response measures. This compensation strategy selection mechanism based on multi-dimensional judgment can not only respond promptly to sudden changes in network status, but also predictively adjust the compensation strategy, maximizing the balance between transmission reliability and system overhead, and providing a more stable and efficient transmission guarantee for ultra-low latency live streaming applications.

[0029] Optionally, determining whether the network link quality parameters show a deteriorating trend within a preset time window includes:

[0030] The target edge computing node divides the preset time window into multiple continuous observation segments according to the sampling period;

[0031] Feature extraction is performed on the network link quality parameters within each observation segment to obtain the rising range of each parameter in the network link quality parameters, which include packet loss rate, network latency, and network jitter;

[0032] Calculate the first time intersection of the rising interval of the packet loss rate and the rising interval of the network latency, and the second time intersection of the rising interval of the packet loss rate and the rising interval of the network jitter.

[0033] In each of the first time intersections, the first time intersections in which the duration of the first time intersection exceeds a preset duration threshold, the increase in the packet loss rate exceeds a preset packet loss rate fluctuation threshold, and the increase in the network latency exceeds a preset latency fluctuation threshold are determined as the first degradation interval.

[0034] In each of the second time intersections, the target second time intersections where the duration of the second time intersection exceeds a preset duration threshold, the increment of the packet loss rate exceeds a preset packet loss rate fluctuation threshold, and the increment of the network jitter exceeds a preset jitter threshold are determined as the second degradation interval.

[0035] All time-continuous or overlapping first and second degradation intervals are merged to obtain one or more quality degradation intervals.

[0036] Each of the aforementioned quality degradation intervals is synthesized into an event cluster, and the network link is determined to be in a deteriorating trend state based on the event cluster.

[0037] By employing the aforementioned technical solution, and dividing a preset time window into multiple continuous observation segments, and extracting features from the network link quality parameters within each segment, fine-grained changes in key indicators such as packet loss rate, network latency, and network jitter can be captured. By calculating the temporal intersection of packet loss rate and network latency, and packet loss rate and network jitter, and combining this with preset multi-dimensional evaluation criteria such as duration thresholds, packet loss rate fluctuation thresholds, latency fluctuation thresholds, and jitter thresholds, the system can accurately identify time intervals with significant degradation characteristics. In particular, this solution intelligently merges temporally continuous or overlapping degradation intervals and organizes them into event clusters, establishing an analytical model from micro-level changes to macro-level trends. This trend judgment mechanism based on multi-parameter collaborative analysis not only effectively filters noise interference in network fluctuations but also promptly identifies inflection points in network quality degradation, providing a reliable decision-making basis for timely adjustments to compensation strategies. This ensures that the ultra-low latency live streaming system can respond quickly and accurately when network conditions change significantly, maximizing the continuity and stability of the live streaming service.

[0038] Optionally, the step of synthesizing each of the quality degradation intervals into an event cluster, and determining whether the network link has entered a deterioration trend state based on the event cluster, includes:

[0039] The target edge computing node calculates the correlation degree of each of the quality degradation intervals, and merges the quality degradation intervals with a correlation degree greater than a preset correlation degree threshold into an event cluster, thereby obtaining one or more event clusters;

[0040] When the number of quality degradation intervals contained in each event cluster exceeds a preset recurrence threshold, the network link is determined to have entered a deterioration trend state.

[0041] Alternatively, sequentially extract temporally adjacent quality degradation interval pairs from each event cluster, and calculate the degradation increase of each interval pair. If the degradation increase of any interval pair exceeds a preset intra-cluster aggravation threshold, then the network link is determined to have entered a degradation trend state.

[0042] Alternatively, the total duration of any event cluster can be calculated. The total duration is the difference between the start time of the first quality degradation interval and the end time of the last quality degradation interval within the event cluster. If the total duration exceeds a preset single-time tolerance threshold, the network link is determined to have entered a deterioration trend state.

[0043] By employing the aforementioned technical solution, and by calculating the correlation between quality degradation intervals and merging them based on a preset correlation threshold, the system can organize intrinsically related network quality fluctuations into meaningful event clusters, thereby grasping the evolutionary patterns of network status at a higher level. The solution designs three complementary trend judgment criteria: First, by monitoring the recurrence frequency of quality degradation intervals within an event cluster, repeated deterioration of network quality can be identified; second, by analyzing the deterioration rate of adjacent quality degradation intervals, the trend of rapid network quality deterioration can be captured; finally, by evaluating the total duration of the event cluster, a persistent decline in network quality can be detected. This multi-faceted trend judgment mechanism not only improves the system's accuracy in identifying network quality degradation but also distinguishes different types of network quality degradation patterns, providing sufficient decision-making basis for the precise adjustment of subsequent compensation strategies. By associating discrete quality degradation events into ordered event clusters and combining them with multi-dimensional judgment criteria, this solution greatly enhances the system's ability to perceive and predict changes in network quality, providing a more reliable guarantee for the stable operation of ultra-low latency live streaming systems.

[0044] Optionally, determining the target compensation strategy based on the network link quality parameters of other edge computing nodes in the edge computing network includes:

[0045] The target edge computing node obtains the target network link quality parameters between other edge computing nodes in the edge computing network and the receiving end;

[0046] Candidate edge computing nodes are determined based on the target network link quality parameters and the link excellence criteria;

[0047] The target edge computing node sends a transmission bearer query request to each of the candidate edge computing nodes to obtain the transmission bearer capability of each candidate edge computing node;

[0048] Generate a target compensation strategy that includes each candidate edge computing node coordinating the transmission of the original media stream according to its corresponding transmission carrying capacity.

[0049] By adopting the above technical solution, when the target edge computing node detects a deteriorating trend in network quality, it obtains the target network link quality parameters between other edge computing nodes and the receiving end in the edge computing network, and selects candidate edge computing nodes based on link quality criteria, establishing a dynamic transmission node candidate pool. By sending transmission bearer query requests to candidate edge computing nodes and obtaining their transmission bearer capabilities, the system can accurately assess the actual transmission potential of each candidate node. Based on this information, the solution generates a target compensation strategy for multi-node collaborative transmission, which fully utilizes the distributed characteristics of the edge computing network by rationally allocating transmission tasks. This intelligent scheduling mechanism based on network quality and transmission capacity can not only effectively avoid the impact of single transmission link quality deterioration, but also optimize resource allocation according to the actual bearer capacity of each node, thereby achieving optimal utilization of system resources while ensuring transmission reliability. By introducing a multi-node collaborative transmission mechanism, this solution significantly improves the adaptability and service quality of the ultra-low latency live streaming system in complex network environments.

[0050] Optionally, determining candidate edge computing nodes based on the target network link quality parameters and the link excellence criteria includes:

[0051] The target edge computing node selects edge computing nodes that meet the link quality criteria as initial candidate nodes based on the quality parameters of each target network link.

[0052] The quality difference vector is obtained by performing a quality difference analysis between the target network link quality parameters of each initial candidate node and the network link quality parameters of the target edge computing node.

[0053] Calculate the quality contribution of each initial candidate node relative to the target edge computing node based on the quality difference vector;

[0054] The target network link quality parameters of each initial candidate node are cross-compared with the target network link quality parameters of other initial candidate nodes to calculate the cooperative potential value between each initial candidate node and other initial candidate nodes.

[0055] The comprehensive score of each initial candidate node is calculated based on the quality contribution and the collaborative potential value. Initial candidate nodes with a comprehensive score greater than the preset score are selected as candidate edge computing nodes.

[0056] By adopting the above technical solution, firstly, by comparing the target network link quality parameters with the link excellence standard, initial candidate nodes with basic transmission capabilities are preliminarily screened, establishing a quality benchmark for the candidate pool. Further, by performing quality difference analysis between the initial candidate nodes and the target edge computing nodes and calculating the quality difference vector, the system can quantitatively evaluate the advantage of each candidate node relative to the current transmission node, thus obtaining the quality contribution of each node. In particular, this scheme calculates the collaborative potential value by conducting cross-comparison analysis among the initial candidate nodes, considering not only the transmission capability of individual nodes but also evaluating the collaborative effect between nodes. Finally, by comprehensively calculating the quality contribution and collaborative potential value, a comprehensive score is obtained, selecting the candidate edge computing node with the highest transmission value. This multi-dimensional node evaluation mechanism not only ensures that each selected collaborative transmission node has sufficient transmission quality assurance but also optimizes the collaborative effect between nodes, constructing an efficient and reliable multi-node transmission network for the ultra-low latency live streaming system, significantly improving the system's adaptability and service stability when network quality dynamically changes.

[0057] A second aspect of this application provides an ultra-low latency live streaming system based on edge computing, applied to a target edge computing node, specifically including:

[0058] The node determination module is used to receive the original media stream sent by the broadcaster terminal. The edge computing network includes multiple edge computing nodes, and the target edge computing node is the edge computing node in the edge computing network that is closest to the broadcaster terminal.

[0059] The parameter monitoring module is used to monitor the uplink network link between the broadcaster and the target edge computing node and the downlink network link between the target edge computing node and the receiver in real time, and to determine the network link quality parameters based on the uplink network link and the downlink network link.

[0060] The standard setting module is used to obtain the historical network link quality parameters of the uplink network link and the downlink network link, and set the link degradation standard and the link excellence standard according to the historical network link quality parameters.

[0061] The strategy determination module is used to determine a target compensation strategy based on the network link quality parameters and a first preset condition and a second preset condition. The first preset condition is that the uplink network link or the downlink network link meets the link degradation standard, and the second preset condition is that both the uplink network link and the downlink network link meet the link excellence standard.

[0062] The media compensation module is used to process the original media stream using the target compensation strategy, generate a compensated media stream, and forward it to the receiving end.

[0063] By adopting the above technical solution, the target edge computing node can accurately grasp the dynamic changes in network conditions by monitoring the network link quality parameters of uplink and downlink network links in real time, and combining these with link degradation and quality standards set based on historical network link quality parameters. Based on this, the target edge computing node dynamically determines the target compensation strategy according to the network link quality parameters and preset conditions. This allows for timely activation of the corresponding compensation mechanism when network conditions deteriorate, and appropriate reduction of the compensation intensity when network conditions are good, thus avoiding the resource waste caused by static compensation strategies. Simultaneously, by deploying the compensation strategy selection decision on edge computing nodes closer to the broadcaster, rapid response to network fluctuations can be achieved, and the transmission overhead of compensation decisions can be reduced. This solution, by introducing an adaptive compensation mechanism based on historical data, achieves a dynamic balance between network transmission reliability and transmission latency, effectively improving the overall performance of the live broadcast.

[0064] A third aspect of this application provides an electronic device including a processor, a memory, a user interface, and a network interface, wherein the memory is used to store instructions, the user interface and the network interface are both used to communicate with other devices, and the processor is used to execute the instructions stored in the memory to cause the electronic device to perform the method as described in any of the foregoing.

[0065] A fourth aspect of this application provides a computer-readable storage medium storing instructions that, when executed, perform the method described in any of the preceding descriptions. Attached Figure Description

[0066] Figure 1 This is an exemplary system architecture diagram of an ultra-low latency live streaming system provided in an embodiment of this application;

[0067] Figure 2 This is a flowchart illustrating an ultra-low latency live streaming method based on edge computing provided in an embodiment of this application.

[0068] Figure 3This is a flowchart illustrating a method for determining a target compensation strategy, as provided in an embodiment of this application.

[0069] Figure 4 This is a schematic diagram of the structure of an ultra-low latency live streaming system based on edge computing provided in an embodiment of this application;

[0070] Figure 5 This is a schematic diagram of the structure of an electronic device disclosed in an embodiment of this application.

[0071] Explanation of reference numerals in the attached figures: 901, processor; 902, communication bus; 903, user interface; 904, network interface; 905, memory. Detailed Implementation

[0072] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0073] In the description of the embodiments of this application, the words "for example" or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design that is described as "for example" or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design options. Rather, the use of the words "for example" or "for instance" is intended to present the relevant concepts in a specific manner.

[0074] In the description of the embodiments of this application, the term "multiple" means two or more. For example, multiple systems means two or more systems, and multiple screen terminals means two or more screen terminals. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. The terms "comprising," "including," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0075] Figure 1 An exemplary system architecture 010 for an ultra-low latency live streaming system is shown.

[0076] like Figure 1As shown, system architecture 010 may include a broadcast terminal 011, an edge computing network 012, and a receiver 013. The components in system architecture 010 are connected via a network, which may include various connection types, such as wired networks, wireless networks, or fiber optic networks. Specifically, the broadcast terminal 011 is connected to the edge computing network 012 via an uplink network link, and the edge computing network 012 is connected to the receiver 013 via a downlink network link.

[0077] The broadcast client 011 is used to capture and transmit raw media streams. The broadcast client 011 can have various media processing applications installed, such as audio and video capture applications, encoding applications, and data transmission applications. The broadcast client 011 is a hardware device and can be any type of device with media capture and processing capabilities, including but not limited to smartphones, professional live streaming equipment, and computers.

[0078] The edge computing network 012 includes multiple edge computing nodes 014, which are distributed at the network edge to provide media stream processing and forwarding services. Some of these edge computing nodes are selected as target edge computing nodes 015, responsible for receiving the raw media stream from the broadcaster 011, performing real-time processing and quality compensation on the media stream, and forwarding the processed media stream to the receiving end 013. Edge computing nodes 014 can be servers providing various media processing services, such as servers monitoring network quality or executing compensation strategies.

[0079] It should be noted that the edge computing node 014 in the edge computing network 012 can be either hardware or software. When the edge computing node 014 is hardware, it can be implemented as a distributed server cluster consisting of multiple servers, or as a single server. When the edge computing node 014 is software, it can be implemented as multiple software programs or software modules (e.g., multiple software programs or software modules used to provide distributed services), or as a single software program or software module. No specific limitations are imposed here.

[0080] Receiver 013 is used to receive and play media streams processed by target edge computing node 015. Receiver 013 can be various terminal devices with media playback capabilities, including but not limited to smartphones, tablets, and personal computers.

[0081] It should be understood that Figure 1The number of broadcast terminals 011, edge computing nodes 014 in the edge computing network 012, and receivers 013 shown are merely illustrative. Depending on implementation needs, there can be any number of broadcast terminals 011, edge computing nodes 014, and receivers 013. Specifically, the edge computing nodes 014 in the edge computing network 012 can be dynamically selected as target edge computing nodes 015 or candidate edge computing nodes 016 to achieve flexible media streaming processing and transmission.

[0082] This application provides an ultra-low latency live streaming method based on edge computing, referencing... Figure 2 , Figure 2 This is a flowchart illustrating an ultra-low latency live streaming method based on edge computing provided in an embodiment of this application, including steps S101 to S105, as follows:

[0083] S101: The target edge computing node receives the original media stream sent by the broadcaster. The edge computing network includes multiple edge computing nodes, and the target edge computing node is the edge computing node in the edge computing network that is closest to the broadcaster.

[0084] In this embodiment, the edge computing network is a distributed network system composed of multiple edge computing nodes with data processing and forwarding capabilities. These nodes are distributed at the network edge and can provide users with nearby computing and storage services. Among them, the target edge computing node, as the node closest to the broadcaster's network in the edge computing network, is responsible for receiving and processing the raw media stream sent by the broadcaster.

[0085] Specifically, the target edge computing node first actively acquires network distance information between all edge computing nodes in the edge computing network and the broadcast terminal. During network distance assessment, the target edge computing node uses network latency and hop count as evaluation metrics: on one hand, the target edge computing node coordinates with each edge computing node to send probe packets to the broadcast terminal and receive responses, obtaining network latency data by recording the round-trip time of the probe packets; on the other hand, the target edge computing node analyzes the transmission path of the probe packets, counting the number of routers the packets pass through, thereby determining the network hop count. The target edge computing node then performs a weighted calculation of the collected network latency and hop count information to obtain a comprehensive network distance metric reflecting actual transmission performance.

[0086] Next, the target edge computing node uses the calculated comprehensive network distance metric to determine whether it is the edge computing node with the closest network distance to the broadcaster. Once confirmed, the target edge computing node begins receiving the raw media stream sent by the broadcaster.

[0087] S102: The target edge computing node monitors the uplink network link between the broadcaster and the target edge computing node, as well as the downlink network link between the target edge computing node and the receiver in real time, and determines the network link quality parameters based on the uplink and downlink network links.

[0088] In this embodiment, to promptly detect fluctuations in network transmission quality and take corresponding compensation measures, the target edge computing node needs to monitor the network quality of the entire transmission link in real time. Specifically, the transmission link is divided into two key parts: an uplink network link connecting the broadcaster and the target edge computing node, and a downlink network link connecting the target edge computing node and the receiver.

[0089] Specifically, during real-time monitoring, the target edge computing node collects network quality data using a combination of active probing and passive measurement. For the uplink network link, the target edge computing node measures the link status by periodically sending probe packets to the broadcaster and receiving responses. Simultaneously, the target edge computing node records and analyzes the received raw media stream data packets in real time, including the timestamp, sequence number, and payload size of each packet. By comparing the timestamps of adjacent packets, the actual arrival time interval of the packets is calculated; by checking the continuity of packet sequence numbers, packet loss or out-of-order delivery is identified; and by analyzing changes in packet payload size, network bandwidth utilization is assessed. This information collectively constitutes the real-time network performance indicators of the uplink. For the downlink network link, the target edge computing node monitors the downlink transmission status in real time by sending probe packets to the receiver and collecting feedback information from the receiver.

[0090] The target edge computing node processes and analyzes the collected raw measurement data to calculate network link quality parameters, including packet loss rate, network latency, and network jitter. Specifically, packet loss rate is calculated by comparing the sequence numbers of sent and received data packets; network latency is obtained by measuring the round-trip time of data packets; and network jitter is calculated by acquiring a sequence of actual arrival time intervals of consecutive data packets and then calculating the deviation of these time intervals from the theoretical arrival time intervals. By statistically analyzing the standard deviation of these deviations, a jitter value reflecting the stability of network transmission is obtained.

[0091] S103: The target edge computing node obtains the historical network link quality parameters of the uplink and downlink network links, and sets the link degradation standard and the link excellence standard based on the historical network link quality parameters.

[0092] In this embodiment, the link degradation standard refers to an evaluation benchmark used to determine whether the network link quality is in a degraded state. When the network link quality parameters are worse than this standard, it indicates that the network transmission quality has significantly declined. The link excellence standard refers to an evaluation benchmark used to determine whether the network link quality is in a good state. When the network link quality parameters are better than this standard, it indicates that the network transmission quality is in an ideal state. In order to accurately determine changes in network quality and formulate reasonable compensation strategies, it is necessary to scientifically establish these two standards based on historical data to provide a reliable reference benchmark for network quality evaluation.

[0093] Specifically, the target edge computing node first retrieves historical network link quality parameters from the system database. These parameters are divided into two categories: first historical quality data reflecting uplink network link transmission quality, and second historical quality data reflecting downlink network link transmission quality. The system performs time-series analysis on this historical data to identify network quality degradation events that occurred during historical transmissions.

[0094] When determining the link degradation criteria, the target edge computing node first filters out the first and second historical quality data, selecting the first and second degradation data that exhibit significant quality deterioration characteristics. Statistical analysis is performed on these degradation data, and a weighted average of parameters such as packet loss rate, network latency, and network jitter is calculated to obtain the first standard score for the uplink network link and the second standard score for the downlink network link. The first and second standard scores together constitute the link degradation criteria, used to determine whether the current network quality is in a degraded state.

[0095] When determining the link quality standard, the target edge computing node first searches for the time period with the best network quality performance in historical data. Specifically, the target edge computing node identifies the time period in which both the first standard score and the second standard score exceed a preset optimal score threshold, and marks them as optimal candidate time periods. Then, the target edge computing node analyzes the first and second historical quality data within these optimal candidate time periods, and obtains the first excellent threshold for the uplink network link and the second excellent threshold for the downlink network link through statistical analysis. The first excellent threshold and the second excellent threshold together constitute the link quality standard.

[0096] Based on the above embodiments, as an optional embodiment, S103: the step of setting link degradation standards and link excellence standards according to historical network link quality parameters may specifically include the following steps:

[0097] S1031: The target edge computing node extracts the first historical quality data of the uplink network link and the second historical quality data of the downlink network link from the historical network link quality parameters.

[0098] Specifically, the target edge computing node first categorizes and organizes historical network link quality parameters based on the data collection timestamp and network link type. For uplink network links, the target edge computing node extracts the transmission data from the broadcaster to itself, including records of packet loss rate changes, network latency fluctuations, and network jitter statistics, integrating them to form the first historical quality data. Similarly, for downlink network links, the target edge computing node extracts the transmission data from itself to the receiver, including records of packet loss rate changes, network latency fluctuations, and historical data of network jitter statistics, integrating them to form the second historical quality data.

[0099] S1032: Select the first degraded data and the second degraded data from the first historical quality data and the second historical quality data, and calculate the first standard score of the uplink network link and the second standard score of the downlink network link based on the first degraded data and the second degraded data.

[0100] Specifically, the target edge computing node first analyzes the first historical quality data by comparing packet loss rate, network latency, and network jitter with their corresponding historical averages. When these three indicators exceed a preset multiple of their respective historical averages, they are marked as quality anomalies. The target edge computing node uses a sliding window approach; when the number of quality anomalies within any time window exceeds a preset proportion of the total number of sampling points in the window, the first historical quality data within that window is filtered out as the first degraded data. The same method is applied to the analysis of the second historical quality data, from which second degraded data representing downlink network link quality anomalies is filtered out.

[0101] After obtaining the first and second degraded data, the target edge computing node performs quantitative analysis on both. For the uplink network link, the target edge computing node calculates the statistical characteristics of various network quality indicators in the first degraded data, including mean, variance, and trend, and assigns preset weights to these statistical characteristics before performing a weighted calculation to obtain the first standard score for the uplink network link. Similarly, the target edge computing node performs similar analysis and calculation on the second degraded data to obtain the second standard score for the downlink network link.

[0102] S1033: Use the first standard score of the uplink network link and the second standard score of the downlink network link as the link degradation standard.

[0103] Specifically, the target edge computing node uses the first standard score of the uplink network link as the benchmark for judging uplink degradation. When the real-time monitored uplink network link quality is lower than this first standard score, it indicates that the uplink network link has degraded. Similarly, the target edge computing node also uses the second standard score of the downlink network link as the benchmark for judging downlink degradation. When the real-time monitored downlink network link quality is lower than this second standard score, it indicates that the downlink network link has degraded.

[0104] S1034: The historical time period in which both the first standard score and the second standard score exceed the preset optimal score threshold is determined as the optimal candidate time period, and the first excellent threshold and the second excellent threshold of the uplink network link and the downlink network link are determined based on the first historical quality data and the second historical quality data within the optimal candidate time period.

[0105] Specifically, the target edge computing node first sets a preset optimal scoring threshold. The target edge computing node compares the first standard score and the second standard score to find the time period in which the first standard score and the second standard score both exceed the preset optimal scoring threshold, and marks these time periods as the optimal candidate time periods.

[0106] After obtaining the optimal candidate time periods, the target edge computing nodes analyze the first and second historical quality data within these time periods. For uplink network links, the target edge computing nodes statistically analyze the packet loss rate, network latency, and network jitter of the first historical quality data within the optimal candidate time periods. By calculating the weighted average of these indicators, the first good / good threshold for the uplink network link is obtained. Similarly, the target edge computing nodes also analyze the second historical quality data within the optimal candidate time periods to calculate the second good / good threshold for the downlink network link.

[0107] S1035: The first good threshold for uplink network links and the second good threshold for downlink network links are determined as the link goodness standard.

[0108] Specifically, the target edge computing node uses the first good / good threshold of the uplink network link as the benchmark value for judging the good / good status of the uplink. When the real-time monitored uplink network link quality reaches or exceeds this first good / good threshold, it indicates that the uplink network link is in a good / good state. Similarly, the target edge computing node also uses the second good / good threshold of the downlink network link as the benchmark value for judging the good / good status of the downlink. When the real-time monitored downlink network link quality reaches or exceeds this second good / good threshold, it indicates that the downlink network link is in a good / good state.

[0109] S104: The target edge computing node determines the target compensation strategy based on the network link quality parameters and the first and second preset conditions. The first preset condition is that the uplink or downlink network link meets the link degradation standard, and the second preset condition is that both the uplink and downlink network links meet the link excellence standard.

[0110] Specifically, the target edge computing node first determines whether the network link quality parameters meet the first preset condition, i.e., whether the quality of the uplink or downlink network link reaches the link degradation standard. When the network link quality is lower than the first standard score or the second standard score, it indicates that the network transmission quality has significantly deteriorated. At this time, the target edge computing node determines to adopt a forward error correction strategy as the target compensation strategy. The forward error correction strategy adds redundant information to the data packets, enabling the receiving end to correct transmission errors on its own, and is suitable for situations with poor network quality.

[0111] If the network link quality parameters do not meet the first preset condition, the target edge computing node further determines whether the second preset condition is met, i.e., whether both the uplink and downlink network links simultaneously reach the excellent link standard. When the quality of both the uplink and downlink network links is higher than both the first and second excellent thresholds, it indicates that the network transmission quality is in an ideal state. At this time, the target edge computing node determines to adopt the automatic retransmission request strategy as the target compensation strategy. The automatic retransmission request strategy ensures transmission reliability by retransmitting lost data packets. This strategy can effectively reduce system overhead when the network quality is good.

[0112] If the network link quality parameters do not meet either the first or the second preset condition, the target edge computing node will further analyze the deterioration trend of the network link quality parameters within a preset time window. When a deterioration trend exists, the target compensation strategy will be determined based on the network link quality parameters of other edge computing nodes in the edge computing network.

[0113] refer to Figure 3 , Figure 3 This is a flowchart illustrating a method for determining a target compensation strategy according to an embodiment of this application, including steps S1041 to S1047, as follows:

[0114] S1041: The target edge computing node determines whether the network link quality parameters meet the first preset condition.

[0115] Specifically, the target edge computing node first collects network link quality parameters at the current moment. These parameters include indicators such as packet loss rate, network latency, and network jitter for both uplink and downlink network links. For the uplink network link, the target edge computing node uses the same scoring method as for calculating the first standard score to weight and calculate the currently collected quality parameters to obtain the real-time score for the current uplink. For the downlink network link, the same scoring method as for calculating the second standard score is used to weight and calculate the currently collected quality parameters to obtain the real-time score for the current downlink.

[0116] When the real-time score of the uplink network link is lower than the first standard score, or the real-time score of the downlink network link is lower than the second standard score, the target edge computing node determines that the network link quality parameters meet the first preset condition.

[0117] S1042: If the first preset condition is met, then the forward error correction strategy is determined to be the target compensation strategy.

[0118] Specifically, when the target edge computing node determines that the real-time score of the uplink network link is lower than the first standard score, or the real-time score of the downlink network link is lower than the second standard score, it confirms that the network link quality parameters meet the first preset condition. At this time, the target edge computing node determines the forward error correction strategy as the target compensation strategy. The core of the forward error correction strategy is to add redundant check information to the original data packet when sending data. This redundant check information enables the receiving end to recover or reconstruct the original data by parsing this check information when receiving damaged data packets, without requesting retransmission from the sending end.

[0119] S1043: If the first preset condition is not met, then determine whether the network link quality parameters meet the second preset condition.

[0120] Specifically, when the target edge computing node confirms that the real-time score of the uplink network link is not lower than the first standard score and the real-time score of the downlink network link is not lower than the second standard score, it confirms that the network link quality parameters do not meet the first preset condition. At this time, the target edge computing node will continue to determine whether the network link quality parameters meet the second preset condition. The specific determination process is that the target edge computing node compares the real-time score of the uplink network link with the first excellent threshold and the real-time score of the downlink network link with the second excellent threshold. Only when the real-time scores of both the uplink and downlink network links exceed their respective excellent thresholds is the network link quality parameter determined to meet the second preset condition.

[0121] S1044: If the second preset condition is met, the automatic retransmission request strategy is determined to be the target compensation strategy.

[0122] Specifically, when the network link quality parameters meet the second preset condition, it indicates that the network transmission quality is at its best. At this time, the target edge computing node needs to select a compensation strategy that can both ensure transmission reliability and reduce system overhead. Therefore, when the target edge computing node confirms that the real-time score of the uplink network link reaches or exceeds the first excellent threshold and the real-time score of the downlink network link reaches or exceeds the second excellent threshold, it determines the automatic retransmission request strategy as the target compensation strategy.

[0123] The automatic retransmission request (ATR) strategy works by sending a retransmission request to the sender when the receiver detects a lost or corrupted data packet. Upon receiving the request, the sender retransmits the corresponding data packet. This strategy is particularly effective when network quality is good, as a good network environment ensures fast and reliable transmission of both the retransmission request and the retransmitted data. When both uplink and downlink network links are in good condition, the additional latency caused by the retransmission mechanism can be kept to a low level. Furthermore, because the network packet loss rate is low at this time, the probability of triggering a retransmission is also low, thus effectively reducing system redundancy overhead.

[0124] S1045: If the second preset condition is not met, determine whether the network link quality parameters show a deteriorating trend within the preset time window.

[0125] Specifically, when the network link quality parameters do not meet either the first or the second preset condition, the target edge computing node needs to predict potential quality problems by analyzing the changing trend of network quality.

[0126] The target edge computing node first divides the preset time window into multiple consecutive observation segments according to the system's sampling period to ensure sufficient temporal granularity for network quality observation. Within each observation segment, the target edge computing node extracts features from network link quality parameters such as packet loss rate, network latency, and network jitter, identifying the time intervals in which these network link quality parameters continuously increase, i.e., the rising intervals.

[0127] Subsequently, the target edge computing node calculates the temporal overlap between the rising interval of packet loss rate and the rising interval of network latency, obtaining the first temporal intersection, and the temporal overlap between the rising interval of packet loss rate and the rising interval of network jitter, obtaining the second temporal intersection. In all first temporal intersections, the target edge computing node selects those intersections that simultaneously satisfy three conditions as the first degradation interval: the duration of the intersection exceeds a preset duration threshold, the increment of packet loss rate within the intersection exceeds a preset packet loss rate fluctuation threshold, and the increment of network latency within the intersection exceeds a preset latency fluctuation threshold. Similarly, in all second temporal intersections, the target edge computing node also selects intersections that simultaneously satisfy similar three conditions as the second degradation interval.

[0128] Next, the target edge computing node merges the first and second degradation intervals that are continuous or overlapping in time to form one or more quality degradation intervals. These quality degradation intervals represent periods in which network quality significantly declines. Finally, the target edge computing node merges these quality degradation intervals into event clusters and analyzes the characteristics of the event clusters (such as frequency of occurrence and duration) to determine whether the network link has entered a deterioration trend.

[0129] Based on the above embodiments, as an optional embodiment, S1045: the step of determining whether the network link quality parameters show a deteriorating trend within a preset time window may specifically include the following steps:

[0130] S10451: The target edge computing node divides the preset time window into multiple continuous observation segments according to the sampling period.

[0131] Specifically, the target edge computing node first defines a preset time window. The length of this time window is set according to the actual operational requirements of the system, typically ranging from several minutes to tens of minutes. Simultaneously, the target edge computing node divides this preset time window into multiple consecutive observation segments at equal intervals, based on the system's sampling period. For example, if the preset time window is 10 minutes and the sampling period is 1 second, 600 consecutive observation segments can be obtained. There are no time gaps between these observation segments, ensuring continuous monitoring of network quality.

[0132] S10452: Perform feature extraction on the network link quality parameters within each observation segment to obtain the rising range of each parameter in the network link quality parameters, including packet loss rate, network latency, and network jitter.

[0133] Specifically, the target edge computing node analyzes network link quality parameters within each observation segment, including packet loss rate, network latency, and network jitter. For the packet loss rate parameter, the target edge computing node compares the packet loss rate values ​​of adjacent sampling points. When a sustained increase in the packet loss rate is detected, this time point is marked as the start of the rising interval; when the packet loss rate stops increasing, this time point is marked as the end of the rising interval. The same method is applied to the analysis of the network latency parameter. The target edge computing node identifies the rising interval of network latency by recognizing the time period in which the latency value continuously increases. For the network jitter parameter, the target edge computing node also determines its rising interval by detecting a continuous increase in the jitter value.

[0134] S10453: Calculate the first time intersection of the rising interval of packet loss rate and the rising interval of network latency, and the second time intersection of the rising interval of packet loss rate and the rising interval of network jitter.

[0135] Specifically, the target edge computing node compares the obtained packet loss rate increase interval with the network latency increase interval along the time dimension, identifies the overlapping portion of these two packet loss rate and network latency increase intervals on the time axis, and marks these overlapping time periods as the first time intersection. For example, if the packet loss rate increases in the time period [t1, t3] and the network latency increases in the time period [t2, t4], then the time period [t2, t3] constitutes a first time intersection.

[0136] Similarly, the target edge computing node compares the packet loss rate increase interval with the network jitter increase interval over time, identifies the overlapping portion of these two parameter increase intervals on the time axis, and marks these overlapping time periods as the second temporal intersection. This calculation method ensures that it can capture the time periods when both packet loss rate and network jitter deteriorate simultaneously.

[0137] S10454: In each first time intersection, the first time intersection where the duration of the first time intersection exceeds a preset duration threshold, the increase in packet loss rate exceeds a preset packet loss rate fluctuation threshold, and the increase in network latency exceeds a preset latency fluctuation threshold is determined as the first degradation interval.

[0138] Specifically, the target edge computing node evaluates each first-time intersection in three dimensions. First, the target edge computing node calculates the duration of the first-time intersection and compares it with a preset duration threshold. This preset duration threshold is used to filter out network fluctuations with short durations. Second, the target edge computing node calculates the overall increment of packet loss rate within the first-time intersection, i.e., the difference between the maximum and minimum packet loss rate within that time period. This increment is compared with a preset packet loss rate fluctuation threshold to ensure that the change in packet loss rate reaches a significant level. Simultaneously, the target edge computing node also calculates the overall increment of network latency within that time period and compares the overall increment of network latency with a preset latency fluctuation threshold to ensure that the increase in network latency reaches a level requiring attention.

[0139] When the duration of a certain first-time intersection exceeds a preset duration threshold, the increase in packet loss rate exceeds a preset packet loss rate fluctuation threshold, and the increase in network latency exceeds a preset latency fluctuation threshold, the target edge computing node identifies it as the first degradation interval. This multi-dimensional judgment criterion ensures that the identified first degradation interval has significant network quality degradation characteristics: the duration is long enough, and both packet loss rate and network latency show a significant increase.

[0140] S10455: In each second time intersection, the target second time intersection where the duration of the second time intersection exceeds a preset duration threshold, the increase in packet loss rate exceeds a preset packet loss rate fluctuation threshold, and the increase in network jitter exceeds a preset jitter threshold is determined as the second degradation interval.

[0141] Specifically, the target edge computing node evaluates each second temporal intersection in three dimensions. First, the target edge computing node calculates the duration of the second temporal intersection and compares it with a preset duration threshold, which is used to filter out network fluctuations with short durations. Second, the target edge computing node calculates the overall increment of packet loss rate within the second temporal intersection, i.e., the difference between the maximum and minimum packet loss rate within that time period, and compares the overall increment with a preset packet loss rate fluctuation threshold to ensure that the change in packet loss rate reaches a significant level. Simultaneously, the target edge computing node also calculates the overall increment of network jitter within that time period and compares it with a preset jitter threshold to ensure that the increase in network jitter reaches a level requiring attention.

[0142] When a second temporal intersection simultaneously meets all three conditions, the target edge computing node identifies it as the second degradation interval. This multi-dimensional judgment criterion ensures that the identified second degradation interval has significant network quality degradation characteristics: the duration is long enough, and both packet loss rate and network jitter show a significant increase.

[0143] S10456: Merge all time-continuous or overlapping first and second degradation intervals to obtain one or more quality degradation intervals.

[0144] Specifically, the target edge computing node first sorts all the first and second degradation intervals in chronological order. Then, the target edge computing node traverses these degradation intervals, checking the temporal relationships between adjacent intervals. When it finds that the first and second degradation intervals are temporally continuous (i.e., the end time of one interval is adjacent to the start time of the other) or overlap (i.e., the two intervals share a common time period on the timeline), the target edge computing node merges these intervals into a larger quality degradation interval. The start time of the merged quality degradation interval is taken from the start time of the earliest first or second degradation interval, and the end time is taken from the end time of the latest degradation interval.

[0145] S10457: Combine each quality degradation interval into an event cluster, and determine whether the network link has entered a deterioration trend state based on the event cluster.

[0146] Specifically, the target edge computing node first calculates the correlation between each quality degradation interval. The correlation can be calculated based on multiple features such as the time interval between degradation intervals and the similarity of degradation levels. When the correlation between two quality degradation intervals exceeds a preset correlation threshold, the target edge computing node merges these intervals into an event cluster.

[0147] After obtaining the event clusters, the target edge computing node uses three methods to determine whether the network link has entered a deterioration trend state. The first method is recurrence frequency judgment. The target edge computing node counts the number of quality degradation intervals within each event cluster. When the number exceeds a preset recurrence frequency threshold, it indicates that network quality degradation is occurring frequently, and the network link is determined to be in a deterioration trend state. The second method is degradation aggravation judgment. The target edge computing node sequentially extracts temporally adjacent quality degradation intervals from each event cluster to form interval pairs, and calculates the degradation increase of each interval pair. When the degradation increase of any interval pair exceeds a preset in-cluster aggravation threshold, it indicates that network quality has significantly deteriorated in a short period of time, and the network link is determined to be in a deterioration trend state. The third method is duration judgment. The target edge computing node calculates the total duration of the event cluster, which is the time difference from the start time of the first quality degradation interval to the end time of the last quality degradation interval within the event cluster. When the total duration exceeds a preset single-incident tolerance duration threshold, it indicates that the network quality degradation has lasted too long, and the network link is determined to be in a deterioration trend state.

[0148] Based on the above embodiments, as an optional embodiment, S10457: the step of synthesizing event clusters from each quality degradation interval and determining whether the network link has entered a deterioration trend state based on the event clusters may specifically include the following steps:

[0149] S104571: The target edge computing node calculates the correlation degree of each quality degradation interval, and merges the quality degradation intervals with a correlation degree greater than the preset correlation degree threshold into an event cluster, thereby obtaining one or more event clusters.

[0150] The target edge computing node first calculates the time distance coefficient between the two quality degradation intervals. Assuming the end and start times of the two quality degradation intervals are t1_end and t2_start, respectively, the time distance coefficient can be expressed as: time_correlation = 1 - (t2_start - t1_end) / T, where T is the preset time window size. The closer the time distance between the two intervals, the closer their time distance coefficient is to 1; when the time distance exceeds the preset time window T, the time distance coefficient is 0.

[0151] Next, the target edge computing node calculates the degradation parameter similarity coefficient between the two quality degradation intervals. For each quality degradation interval, a feature vector can be extracted, including parameters such as average packet loss rate, average network latency, and average network jitter. The degradation parameter similarity coefficient is obtained by calculating the cosine similarity of the feature vectors of the two quality degradation intervals.

[0152] Finally, the target edge computing node performs a weighted sum of the temporal distance coefficient and the degradation parameter similarity coefficient to obtain the final correlation score. When the final correlation score exceeds a preset correlation score threshold, the target edge computing node merges the two quality degradation intervals into the same event cluster. By iteratively processing all quality degradation intervals, one or more event clusters are ultimately formed.

[0153] S104572: When the number of quality degradation intervals contained in each event cluster exceeds the preset recurrence threshold, the network link is determined to be in a deterioration trend state.

[0154] Specifically, the target edge computing node first traverses each event cluster and counts the number of quality degradation intervals contained within it. When the number of quality degradation intervals within an event cluster exceeds a preset recurrence threshold, the target edge computing node determines that the network link has entered a deterioration trend.

[0155] S104573: Alternatively, extract temporally adjacent quality degradation interval pairs in each event cluster, and calculate the degradation increase of each interval pair. If the degradation increase of any interval pair exceeds the preset intra-cluster aggravation threshold, then the network link is determined to enter a deterioration trend state.

[0156] The rate of deterioration reflects the degree of deterioration in network quality parameters in the later quality degradation interval relative to the previous quality degradation interval.

[0157] Specifically, the target edge computing node first extracts adjacent quality degradation intervals in each event cluster in chronological order, forming multiple interval pairs. For each interval pair, the target edge computing node calculates the degradation increase of the interval pair. First, it calculates the average packet loss rate, average network latency, and average network jitter for each quality degradation interval within the interval pair; then, it calculates the relative rates of change of these parameters, i.e.: Packet loss rate change rate = (average packet loss rate of the later interval - average packet loss rate of the earlier interval) / average packet loss rate of the earlier interval; Latency change rate = (average latency of the later interval - average latency of the earlier interval) / average latency of the earlier interval; Jitter change rate = (average jitter of the later interval - average jitter of the earlier interval) / average jitter of the earlier interval.

[0158] Preset weights are assigned to the packet loss rate change rate, latency change rate, and jitter change rate, respectively. The weighted sum is used to obtain the deterioration increase. When the calculated deterioration increase exceeds the preset intra-cluster aggravation threshold, the target edge computing node determines that the network link has entered a deterioration trend state.

[0159] S104574: Alternatively, calculate the total duration of any event cluster. The total duration is the difference between the start time of the first quality degradation interval and the end time of the last quality degradation interval within the event cluster. If the total duration exceeds the preset single tolerance duration threshold, the network link is determined to be in a deterioration trend state.

[0160] Specifically, the target edge computing node first obtains the start time of the earliest quality degradation interval and the end time of the last quality degradation interval within each event cluster. By calculating the difference between these two time points, the total duration of the event cluster is obtained. When the total duration of an event cluster exceeds a preset single-instance tolerance threshold, the target edge computing node determines that the network link has entered a deterioration trend state.

[0161] S1046: If a deterioration trend is observed, the target compensation strategy shall be determined based on the network link quality parameters of other edge computing nodes in the edge computing network.

[0162] Specifically, once a network link is determined to be deteriorating, the target edge computing node first acquires the target network link quality parameters between other edge computing nodes in the edge computing network and the receiving end, including key indicators such as network latency, packet loss rate, and network jitter. These parameters reflect the network transmission quality status of other edge computing nodes.

[0163] Subsequently, the target edge computing node compares the acquired target network link quality parameters with preset link quality standards. These standards include threshold requirements such as upper limits for network latency, packet loss rate, and network jitter. When the network link quality parameters of an edge computing node meet these standards, the target edge computing node designates it as a candidate edge computing node. This selection method ensures that only nodes with good network conditions can participate in compensation transmission.

[0164] To further evaluate the actual transmission capabilities of candidate edge computing nodes, the target edge computing node sends a transmission bearer query request to each candidate edge computing node. Upon receiving the request, each candidate edge computing node returns its own transmission bearer capabilities, including available bandwidth and computing resources. This information reflects the actual transmission guarantee capabilities of the candidate edge computing nodes.

[0165] Finally, the target edge computing node generates a target compensation strategy based on the transmission capacity of each candidate edge computing node. This strategy specifies how each candidate edge computing node should collaboratively transmit the original media stream according to its respective transmission capacity. For example, candidate nodes with stronger transmission capacity can undertake a larger proportion of media stream transmission tasks, while candidate nodes with weaker transmission capacity can undertake a relatively smaller transmission task.

[0166] Based on the above embodiments, as an optional embodiment, S1046: the step of determining the target compensation strategy based on the network link quality parameters of other edge computing nodes in the edge computing network may specifically include the following steps:

[0167] S10461: The target edge computing node obtains the target network link quality parameters between other edge computing nodes in the edge computing network and the receiving end.

[0168] Specifically, the target edge computing node obtains the target network link quality parameters between other edge computing nodes and the receiving end through the edge computing network's network quality monitoring system. These target network link quality parameters mainly include key indicators such as network latency, packet loss rate, and network jitter.

[0169] S10462: Determine candidate edge computing nodes based on the target network link quality parameters and link excellence criteria.

[0170] Specifically, the target edge computing node first obtains the target network link quality parameters for each edge computing node. Based on these parameters, it calculates a first score reflecting uplink network link quality and a second score reflecting downlink network link quality. The first and second scores are then compared to the link quality standard. When an edge computing node simultaneously satisfies the condition that its first score is greater than a first good threshold and its second score is greater than a second good threshold, the target edge computing node designates this edge computing node as an initial candidate node.

[0171] Next, the target edge computing node performs a quality difference analysis on each initial candidate node. By comparing the target network link quality parameters of the initial candidate nodes with the current network link quality parameters of the target edge computing node, the differences in network latency, packet loss rate, and network jitter are calculated, forming a quality difference vector.

[0172] Based on the quality difference vector, the target edge computing nodes calculate the quality contribution of each initial candidate node. The quality contribution can be obtained by weighted summation of the components of the quality difference vector, with the weight coefficients reflecting the importance of different network quality parameters. A higher quality contribution indicates that the node can provide better network quality improvement in compensation transmission.

[0173] Subsequently, the target edge computing nodes perform pairwise cross-comparisons with the initial candidate nodes to calculate their collaborative potential value. The collaborative potential value reflects the performance of two nodes in terms of network quality complementarity, and can be calculated by analyzing the degree of complementarity of their network quality parameters.

[0174] Finally, the target edge computing node calculates a comprehensive score for each initial candidate node based on its quality contribution and collaborative potential. The comprehensive score is obtained by weighted combining the quality contribution and average collaborative potential, with the weighting coefficients adjustable according to actual needs. The target edge computing node identifies the initial candidate nodes whose comprehensive scores exceed a preset score as the final candidate edge computing nodes.

[0175] Based on the above embodiments, as an optional embodiment, S10462: the step of determining candidate edge computing nodes according to the target network link quality parameters and link excellence criteria may specifically include the following steps:

[0176] S104621: The target edge computing nodes select edge computing nodes that meet the link quality standards as initial candidate nodes based on the link quality parameters of each target network.

[0177] Specifically, the target edge computing node first analyzes the target network link quality parameters between each edge computing node and the receiver. Using the same weighted calculation method as the link quality standard, it calculates a first score reflecting uplink network link quality and a standard score reflecting downlink network link quality. Then, it compares the calculated first score with a first quality threshold determined beforehand using historical data, and compares the second score with a second quality threshold. When an edge computing node simultaneously satisfies the condition that its first score is greater than the first quality threshold and its second score is greater than the second quality threshold, the target edge computing node designates this edge computing node as an initial candidate node.

[0178] S104622: Perform quality difference analysis between the target network link quality parameters of each initial candidate node and the network link quality parameters of the target edge computing node to obtain a quality difference vector.

[0179] Specifically, the target edge computing nodes undergo quality difference analysis. For each initial candidate node, the target edge computing node calculates the difference between the network latency of the initial candidate node and its current network latency to obtain a latency difference component; the difference between the packet loss rate of the initial candidate node and its current packet loss rate to obtain a packet loss rate difference component; and the difference between the network jitter of the initial candidate node and its current network jitter to obtain a jitter difference component. These latency difference, packet loss rate difference, and jitter difference components together constitute a three-dimensional quality difference vector, where each component is "initial candidate node parameter value - target node parameter value". Using this calculation method, negative values ​​in any component of the quality difference vector indicate that the initial candidate node's network quality is superior to the target edge computing node in that dimension.

[0180] S104623: Calculate the quality contribution of each initial candidate node relative to the target edge node based on the quality difference vector.

[0181] Specifically, the target edge computing node first sets the weight coefficients for the network quality parameters. For the three components in the quality difference vector, namely network latency difference, packet loss rate difference, and network jitter difference, preset weight coefficients are set respectively.

[0182] Then, the target edge computing node performs a weighted summation of the quality difference vector of each initial candidate node to obtain the quality contribution of that initial candidate node.

[0183] S104624: Cross-compare the target network link quality parameters of each initial candidate node with those of other initial candidate nodes to calculate the collaborative potential value between each initial candidate node and other initial candidate nodes.

[0184] Specifically, the target edge computing node first pairs and compares the target network link quality parameters of each initial candidate node with those of all other initial candidate nodes. For each pair of initial candidate nodes, the target edge computing node compares their differences in parameters such as network latency, packet loss rate, and network jitter. When one node performs better on some network quality parameters, while another node performs better on others, this complementarity can improve overall transmission performance.

[0185] The target edge computing node calculates the collaboration potential value between two nodes based on parameter differences. Specifically, when two nodes exhibit complementary advantages in different network quality parameters, their collaboration potential value is increased; when the advantages of two nodes overlap or their performance is similar, their collaboration potential value is decreased. For example, one node has low network latency but a high packet loss rate, while another node has high network latency but a low packet loss rate; this complementarity leads to a higher collaboration potential value.

[0186] After calculating the collaborative potential values ​​of all node pairs, the target edge computing node can obtain a complete collaborative potential matrix, where each element represents the collaborative potential value between the corresponding two nodes.

[0187] S104625: Calculate the comprehensive score of each initial candidate node based on the quality contribution and collaborative potential value, and select the initial candidate node with the comprehensive score greater than the preset score as the candidate edge computing node.

[0188] Specifically, for each initial candidate node, the target edge computing node multiplies its quality contribution by the corresponding preset weight coefficient, and then averages the collaborative potential value of the node with all other initial candidate nodes, multiplies it by the preset weight coefficient of the collaborative potential value, and adds them together to obtain the comprehensive score of the initial candidate node.

[0189] The target edge computing node compares the overall score of each initial candidate node with a preset score. The preset score is a threshold pre-set by the system based on actual performance requirements, used to filter nodes with sufficient performance guarantees. When the overall score of an initial candidate node is greater than the preset score, the target edge computing node confirms it as a candidate edge computing node.

[0190] S10463: The target edge computing node sends a transmission bearer query request to each candidate edge computing node to obtain the transmission bearer capability of each candidate edge computing node.

[0191] Specifically, the target edge computing node first constructs a transport bearer query request. This request contains key information such as the target edge computing node's identification information and the data transmission requirements of the current service flow. The target edge computing node concurrently sends the constructed transport bearer query request to all candidate edge computing nodes to improve query efficiency.

[0192] Upon receiving a transmission capacity inquiry request, each candidate edge computing node conducts a comprehensive assessment of its own resource status. When assessing available bandwidth, candidate edge computing nodes calculate the currently available bandwidth resources by monitoring real-time network interface traffic statistics, including used bandwidth and total bandwidth capacity. For processing capacity assessment, candidate edge computing nodes calculate the available computing resource capacity by detecting CPU utilization, the number of currently running tasks, and task queue length, combined with historical processing performance data. Regarding storage space assessment, candidate edge computing nodes statistically analyze current disk space usage, including system occupancy, allocated business data space, and available space, while also considering performance indicators such as data read / write speed. The bandwidth resources, computing resource capacity, and disk space usage are integrated to form a transmission capacity response. The candidate edge computing node returns the assessed transmission capacity information to the target edge computing node.

[0193] S10464: Generate a target compensation strategy that includes each candidate edge computing node coordinating the transmission of the original media stream according to its corresponding transmission carrying capacity.

[0194] Specifically, the target edge computing node first analyzes the original media stream to determine the amount of data requiring compensation for transmission and the required transmission quality. Then, based on the transmission capacity of each candidate edge computing node, the target edge computing node calculates the proportion of transmission tasks each node can handle. Based on this proportion, transmission task shares are allocated.

[0195] When generating the target compensation strategy, the target edge computing node also needs to consider the coordination mechanism among candidate edge computing nodes. This includes determining the data distribution method, planning the transmission sequence of each node, and designing a data redundancy mechanism. When a candidate edge computing node experiences a transmission anomaly, the redundancy mechanism in the target compensation strategy ensures that other nodes can promptly take over the transmission task, maintaining transmission continuity.

[0196] S1047: If no deterioration trend is observed, the automatic retransmission request strategy is determined as the target compensation strategy.

[0197] Specifically, the target edge computing node determines whether the second preset condition is met by detecting whether the real-time score of the uplink network link reaches or exceeds the first excellent threshold, and whether the real-time score of the downlink network link reaches or exceeds the second excellent threshold. When the above conditions are met simultaneously, the target edge computing node determines the automatic retransmission request strategy as the target compensation strategy.

[0198] S105: The target edge computing node uses a target compensation strategy to process the original media stream, generate a compensated media stream, and forward it to the receiving end.

[0199] Specifically, the target edge computing node first receives the original media stream from the sender. When the target compensation strategy is an automatic repeat request strategy, the target edge computing node identifies lost data packets during transmission by detecting the continuity of data packet sequence numbers; when the target compensation strategy is a multipath transmission strategy, the target edge computing node distributes the original media stream to multiple candidate edge computing nodes for parallel transmission.

[0200] The target edge computing node performs corresponding compensation processing on the original media stream according to the specific requirements of the target compensation strategy. During the processing, the target edge computing node may need to perform operations such as packet reassembly, timing adjustment, and redundant data removal. The purpose of these processing operations is to restore data integrity, ensure the continuity of the media stream, and optimize transmission efficiency. After processing is complete, the target edge computing node sends the generated compensated media stream to the receiving end according to the predetermined transmission path.

[0201] refer to Figure 4 This application also provides a schematic diagram of an ultra-low latency live streaming system based on edge computing. This system is applied to a target edge computing node and specifically includes:

[0202] The node determination module is used to receive the original media stream sent by the broadcaster terminal. The edge computing network includes multiple edge computing nodes, and the target edge computing node is the edge computing node in the edge computing network that is closest to the broadcaster terminal.

[0203] The parameter monitoring module is used to monitor the uplink network link between the broadcaster and the target edge computing node and the downlink network link between the target edge computing node and the receiver in real time, and to determine the network link quality parameters based on the uplink network link and the downlink network link.

[0204] The standard setting module is used to obtain the historical network link quality parameters of the uplink network link and the downlink network link, and set the link degradation standard and the link excellence standard according to the historical network link quality parameters.

[0205] The strategy determination module is used to determine a target compensation strategy based on the network link quality parameters and a first preset condition and a second preset condition. The first preset condition is that the uplink network link or the downlink network link meets the link degradation standard, and the second preset condition is that both the uplink network link and the downlink network link meet the link excellence standard.

[0206] The media compensation module is used to process the original media stream using the target compensation strategy, generate a compensated media stream, and forward it to the receiving end.

[0207] Optionally, the standard setting module is specifically used for:

[0208] Extract the first historical quality data of the uplink network link and the second historical quality data of the downlink network link from the historical network link quality parameters, respectively.

[0209] First degraded data and second degraded data are selected from the first historical quality data and the second historical quality data. Based on the first degraded data and the second degraded data, a first standard score of the uplink network link and a second standard score of the downlink network link are calculated.

[0210] The first standard score of the uplink network link and the second standard score of the downlink network link are used as the link degradation criteria;

[0211] The historical time period in which both the first standard score and the second standard score exceed the preset optimal score threshold is determined as the optimal candidate time period. The first excellent threshold of the uplink network link and the second excellent threshold of the downlink network link are determined based on the first historical quality data and the second historical quality data within the optimal candidate time period.

[0212] The first good threshold of the uplink network link and the second good threshold of the downlink network link are determined as the link goodness standard.

[0213] Optionally, the strategy determination module is specifically used for:

[0214] Determine whether the network link quality parameters meet the first preset condition;

[0215] If the first preset condition is met, then the forward error correction strategy is determined to be the target compensation strategy;

[0216] If the first preset condition is not met, then determine whether the network link quality parameters meet the second preset condition;

[0217] If the second preset condition is met, the automatic retransmission request strategy is determined to be the target compensation strategy.

[0218] If the second preset condition is not met, then determine whether the network link quality parameters show a deteriorating trend within a preset time window;

[0219] If a deterioration trend is observed, a target compensation strategy is determined based on the network link quality parameters of other edge computing nodes in the edge computing network.

[0220] If no deterioration trend is observed, the automatic retransmission request strategy is determined as the target compensation strategy.

[0221] Optionally, the strategy determination module is further specifically used for:

[0222] The preset time window is divided into multiple consecutive observation segments according to the sampling period;

[0223] Feature extraction is performed on the network link quality parameters within each observation segment to obtain the rising range of each parameter in the network link quality parameters, which include packet loss rate, network latency, and network jitter;

[0224] Calculate the first time intersection of the rising interval of the packet loss rate and the rising interval of the network latency, and the second time intersection of the rising interval of the packet loss rate and the rising interval of the network jitter.

[0225] In each of the first time intersections, the first time intersections in which the duration of the first time intersection exceeds a preset duration threshold, the increase in the packet loss rate exceeds a preset packet loss rate fluctuation threshold, and the increase in the network latency exceeds a preset latency fluctuation threshold are determined as the first degradation interval.

[0226] In each of the second time intersections, the target second time intersections where the duration of the second time intersection exceeds a preset duration threshold, the increment of the packet loss rate exceeds a preset packet loss rate fluctuation threshold, and the increment of the network jitter exceeds a preset jitter threshold are determined as the second degradation interval.

[0227] All time-continuous or overlapping first and second degradation intervals are merged to obtain one or more quality degradation intervals.

[0228] Each of the aforementioned quality degradation intervals is synthesized into an event cluster, and the network link is determined to be in a deteriorating trend state based on the event cluster.

[0229] Optionally, the strategy determination module is further specifically used for:

[0230] Calculate the correlation degree of each of the quality degradation intervals, and merge the quality degradation intervals with a correlation degree greater than a preset correlation degree threshold into an event cluster to obtain one or more event clusters;

[0231] When the number of quality degradation intervals contained in each event cluster exceeds a preset recurrence threshold, the network link is determined to have entered a deterioration trend state.

[0232] Alternatively, sequentially extract temporally adjacent quality degradation interval pairs from each event cluster, and calculate the degradation increase of each interval pair. If the degradation increase of any interval pair exceeds a preset intra-cluster aggravation threshold, then the network link is determined to have entered a degradation trend state.

[0233] Alternatively, the total duration of any event cluster can be calculated. The total duration is the difference between the start time of the first quality degradation interval and the end time of the last quality degradation interval within the event cluster. If the total duration exceeds a preset single-time tolerance threshold, the network link is determined to have entered a deterioration trend state.

[0234] Optionally, the strategy determination module is further specifically used for:

[0235] Obtain the target network link quality parameters between other edge computing nodes in the edge computing network and the receiving end;

[0236] Candidate edge computing nodes are determined based on the target network link quality parameters and the link excellence criteria;

[0237] The target edge computing node sends a transmission bearer query request to each of the candidate edge computing nodes to obtain the transmission bearer capability of each candidate edge computing node;

[0238] Generate a target compensation strategy that includes each candidate edge computing node coordinating the transmission of the original media stream according to its corresponding transmission carrying capacity.

[0239] Optionally, the strategy determination module is further specifically used for:

[0240] Based on the quality parameters of each target network link, edge computing nodes that meet the excellent link criteria are selected as initial candidate nodes.

[0241] The quality difference vector is obtained by performing a quality difference analysis between the target network link quality parameters of each initial candidate node and the network link quality parameters of the target edge computing node.

[0242] Calculate the quality contribution of each initial candidate node relative to the target edge computing node based on the quality difference vector;

[0243] The target network link quality parameters of each initial candidate node are cross-compared with the target network link quality parameters of other initial candidate nodes to calculate the cooperative potential value between each initial candidate node and other initial candidate nodes.

[0244] The comprehensive score of each initial candidate node is calculated based on the quality contribution and the collaborative potential value. Initial candidate nodes with a comprehensive score greater than the preset score are selected as candidate edge computing nodes.

[0245] It should be noted that the above embodiments of the apparatus are only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the apparatus and method embodiments provided above belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.

[0246] This embodiment also discloses an electronic device, as shown in the reference. Figure 5 , Figure 5 This is a schematic diagram of the structure of an electronic device disclosed in an embodiment of this application. The electronic device may include: at least one processor 901, at least one communication bus 902, a user interface 903, a network interface 904, and at least one memory 905.

[0247] The communication bus 902 is used to enable communication between these components.

[0248] The user interface 903 may include a display screen and a camera. Optionally, the user interface 903 may also include a standard wired interface and a wireless interface.

[0249] The network interface 904 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0250] The processor 901 may include one or more processing cores. The processor 901 connects to various parts of the server using various interfaces and lines, and performs various server functions and processes data by running or executing instructions, programs, code sets, or instruction sets stored in the memory 905, and by calling data stored in the memory 905. Optionally, the processor 901 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array. The processor 901 may integrate one or a combination of several of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content to be displayed on the screen; and the modem handles wireless communication. It is understood that the modem may also not be integrated into the processor 901 and may be implemented as a separate chip.

[0251] The memory 905 may include random access memory (RAM) or read-only memory. Optionally, the memory 905 may include a non-transitory computer-readable storage medium. The memory 905 may be used to store instructions, programs, code, code sets, or instruction sets. The memory 905 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-described method embodiments, etc.; the data storage area may store data involved in the above-described method embodiments, etc. Optionally, the memory 905 may also be at least one storage device located remotely from the aforementioned processor 901. (See reference...) Figure 5 The memory 905, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and an application for ultra-low latency live streaming based on edge computing.

[0252] exist Figure 5In the electronic device shown, the user interface 903 is mainly used to provide an input interface for the user and to obtain the user input data; while the processor 901 can be used to call an application for ultra-low latency live streaming based on edge computing stored in the memory 905. When executed by one or more processors 901, the electronic device performs one or more methods as described in the above embodiments.

[0253] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0254] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0255] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the shown or discussed mutual couplings or direct couplings or communication connections may be through some service interfaces; indirect couplings or communication connections between apparatuses or units may be electrical or other forms.

[0256] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0257] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0258] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, portable hard drives, magnetic disks, or optical disks.

[0259] The foregoing description is merely an exemplary embodiment of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure. Other embodiments of this disclosure will be readily apparent to those skilled in the art upon consideration of the disclosure in this specification. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described in this disclosure. The specification and embodiments are to be considered exemplary only, and the scope and spirit of this disclosure are defined by the claims.

Claims

1. A method for ultra-low latency live streaming based on edge computing, characterized in that, The method includes: The target edge computing node receives the original media stream sent by the broadcaster. The edge computing network includes multiple edge computing nodes, and the target edge computing node is the edge computing node in the edge computing network that is closest to the broadcaster. The target edge computing node monitors the uplink network link between the broadcaster and the target edge computing node, and the downlink network link between the target edge computing node and the receiver in real time, and determines the network link quality parameters based on the uplink network link and the downlink network link. The target edge computing node obtains the historical network link quality parameters of the uplink network link and the downlink network link, and sets the link degradation standard and the link excellence standard based on the historical network link quality parameters; The target edge computing node determines the target compensation strategy based on the network link quality parameters and the first and second preset conditions. The first preset condition is that the uplink network link or the downlink network link meets the link degradation standard, and the second preset condition is that both the uplink network link and the downlink network link meet the link excellence standard. The target edge computing node processes the original media stream using the target compensation strategy to generate a compensated media stream, and then forwards it to the receiving end. The target edge computing node determines a target compensation strategy based on the network link quality parameters and the first and second preset conditions, including: The target edge computing node determines whether the network link quality parameters meet the first preset condition; If the first preset condition is met, then the forward error correction strategy is determined to be the target compensation strategy; If the first preset condition is not met, then determine whether the network link quality parameters meet the second preset condition; If the second preset condition is met, the automatic retransmission request strategy is determined to be the target compensation strategy. If the second preset condition is not met, then determine whether the network link quality parameters show a deteriorating trend within a preset time window; If a deterioration trend is observed, a target compensation strategy is determined based on the network link quality parameters of other edge computing nodes in the edge computing network. If no deterioration trend is observed, the automatic retransmission request strategy is determined as the target compensation strategy.

2. The ultra-low latency live streaming method based on edge computing according to claim 1, characterized in that, The step of setting link degradation standards and link excellence standards based on the historical network link quality parameters includes: The target edge computing node extracts the first historical quality data of the uplink network link and the second historical quality data of the downlink network link from the historical network link quality parameters, respectively. First degraded data and second degraded data are selected from the first historical quality data and the second historical quality data. Based on the first degraded data and the second degraded data, a first standard score of the uplink network link and a second standard score of the downlink network link are calculated. The first standard score of the uplink network link and the second standard score of the downlink network link are used as the link degradation criteria; The historical time period in which both the first standard score and the second standard score exceed the preset optimal score threshold is determined as the optimal candidate time period. The first excellent threshold of the uplink network link and the second excellent threshold of the downlink network link are determined based on the first historical quality data and the second historical quality data within the optimal candidate time period. The first good threshold of the uplink network link and the second good threshold of the downlink network link are determined as the link goodness standard.

3. The ultra-low latency live streaming method based on edge computing according to claim 1, characterized in that, The step of determining whether the network link quality parameters show a deteriorating trend within a preset time window includes: The target edge computing node divides the preset time window into multiple continuous observation segments according to the sampling period; Feature extraction is performed on the network link quality parameters within each observation segment to obtain the rising range of each parameter in the network link quality parameters, which include packet loss rate, network latency, and network jitter; Calculate the first time intersection of the rising interval of the packet loss rate and the rising interval of the network latency, and the second time intersection of the rising interval of the packet loss rate and the rising interval of the network jitter. In each of the first time intersections, the first time intersections in which the duration of the first time intersection exceeds a preset duration threshold, the increase in the packet loss rate exceeds a preset packet loss rate fluctuation threshold, and the increase in the network latency exceeds a preset latency fluctuation threshold are determined as the first degradation interval. In each of the second time intersections, the target second time intersections where the duration of the second time intersection exceeds a preset duration threshold, the increment of the packet loss rate exceeds a preset packet loss rate fluctuation threshold, and the increment of the network jitter exceeds a preset jitter threshold are determined as the second degradation interval. All time-continuous or overlapping first and second degradation intervals are merged to obtain one or more quality degradation intervals. Each of the aforementioned quality degradation intervals is synthesized into an event cluster, and the network link is determined to be in a deteriorating trend state based on the event cluster.

4. The ultra-low latency live streaming method based on edge computing according to claim 3, characterized in that, The step of synthesizing each of the quality degradation intervals into an event cluster, and determining whether the network link has entered a deterioration trend state based on the event cluster, includes: The target edge computing node calculates the correlation degree of each of the quality degradation intervals, and merges the quality degradation intervals with a correlation degree greater than a preset correlation degree threshold into an event cluster, thereby obtaining one or more event clusters; When the number of quality degradation intervals contained in each event cluster exceeds a preset recurrence threshold, the network link is determined to have entered a deterioration trend state. Alternatively, sequentially extract temporally adjacent quality degradation interval pairs from each event cluster, and calculate the degradation increase of each interval pair. If the degradation increase of any interval pair exceeds a preset intra-cluster aggravation threshold, then the network link is determined to have entered a degradation trend state. Alternatively, the total duration of any event cluster can be calculated. The total duration is the difference between the start time of the first quality degradation interval and the end time of the last quality degradation interval within the event cluster. If the total duration exceeds a preset single-time tolerance threshold, the network link is determined to have entered a deterioration trend state.

5. The ultra-low latency live streaming method based on edge computing according to claim 1, characterized in that, The step of determining the target compensation strategy based on the network link quality parameters of other edge computing nodes in the edge computing network includes: The target edge computing node obtains the target network link quality parameters between other edge computing nodes in the edge computing network and the receiving end; Candidate edge computing nodes are determined based on the target network link quality parameters and the link excellence criteria; The target edge computing node sends a transmission bearer query request to each of the candidate edge computing nodes to obtain the transmission bearer capability of each candidate edge computing node; Generate a target compensation strategy that includes each candidate edge computing node coordinating the transmission of the original media stream according to its corresponding transmission carrying capacity.

6. The ultra-low latency live streaming method based on edge computing according to claim 5, characterized in that, The step of determining candidate edge computing nodes based on the target network link quality parameters and the link excellence criteria includes: The target edge computing node selects edge computing nodes that meet the link quality criteria as initial candidate nodes based on the quality parameters of each target network link. The quality difference vector is obtained by performing a quality difference analysis between the target network link quality parameters of each initial candidate node and the network link quality parameters of the target edge computing node. Calculate the quality contribution of each initial candidate node relative to the target edge computing node based on the quality difference vector; The target network link quality parameters of each initial candidate node are cross-compared with the target network link quality parameters of other initial candidate nodes to calculate the cooperative potential value between each initial candidate node and other initial candidate nodes. The comprehensive score of each initial candidate node is calculated based on the quality contribution and the collaborative potential value. Initial candidate nodes with a comprehensive score greater than the preset score are selected as candidate edge computing nodes.

7. An ultra-low latency live streaming system based on edge computing, characterized in that, For executing the ultra-low latency live streaming method based on edge computing as described in claim 1, and applied to a target edge computing node, the system comprises: The node determination module is used to receive the original media stream sent by the broadcaster terminal. The edge computing network includes multiple edge computing nodes, and the target edge computing node is the edge computing node in the edge computing network that is closest to the broadcaster terminal. The parameter monitoring module is used to monitor the uplink network link between the broadcaster and the target edge computing node and the downlink network link between the target edge computing node and the receiver in real time, and to determine the network link quality parameters based on the uplink network link and the downlink network link. The standard setting module is used to obtain the historical network link quality parameters of the uplink network link and the downlink network link, and set the link degradation standard and the link excellence standard according to the historical network link quality parameters. The strategy determination module is used to determine a target compensation strategy based on the network link quality parameters and a first preset condition and a second preset condition. The first preset condition is that the uplink network link or the downlink network link meets the link degradation standard, and the second preset condition is that both the uplink network link and the downlink network link meet the link excellence standard. The media compensation module is used to process the original media stream using the target compensation strategy, generate a compensated media stream, and forward it to the receiving end.

8. An electronic device, characterized in that, The device includes a processor, a memory, a user interface, and a network interface. The memory is used to store instructions. The user interface and the network interface are both used to communicate with other devices. The processor is used to execute the instructions stored in the memory to cause the electronic device to perform the method as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed by a processor, perform the method as described in any one of claims 1-6.

Images