A game network acceleration method for multi-terminal cooperation

Abstract

This invention discloses a game network acceleration method for multi-terminal collaboration, belonging to the field of computer network and edge computing technology. The method includes the following steps: acquiring network status data of multiple game terminals and corresponding edge node operation data to generate a standard network status sequence; constructing terminal-side network quality descriptions and node-side network quality descriptions based on the standard network status sequence to generate a set of effective links; forming a candidate link subset based on the set of effective links; selecting a target transmission link from the candidate link subset and transmitting data; continuously updating the standard network status sequence during data transmission, and performing link replacement after the replaceable state persists to a preset stable stage, thus completing the game network acceleration for multi-terminal collaboration. This invention improves transmission stability in complex network environments by performing path reconstruction processing on links with degraded performance.

Description

Technical Field

[0001] This invention relates to the field of computer networks and edge computing technology, and in particular to a method for accelerating game networks for multi-terminal collaboration. Background Technology

[0002] As online games evolve towards multi-terminal collaboration and high real-time interaction, game networks typically refer to the communication link system used to carry game data transmission. This includes multiple transmission paths between game terminals and servers or edge nodes. Game network acceleration, on the other hand, involves selecting, scheduling, and optimizing these transmission paths to reduce latency, suppress jitter, and minimize packet loss, thereby improving the real-time performance and stability of game interactions.

[0003] In existing technologies, game network acceleration typically involves detecting the current network status of the terminal and ranking multiple candidate links based on metrics such as latency, jitter, and packet loss, thereby selecting one as the data transmission link. When the link performance degrades, the link is switched by re-detecting the network status and reselecting the link. Some solutions also combine the resource status of edge nodes to schedule the links in order to improve link utilization and transmission stability.

[0004] However, existing technologies typically rely solely on real-time detection results to directly select candidate links during link switching, lacking the ability to predict the transmission performance of candidate links under the target transmission state. Furthermore, they do not perform path reconstruction for links with degraded performance, making it difficult to obtain stable alternative links when all available links are fluctuating. In addition, existing technologies are mostly based on fixed topologies or static connection relationships during path construction, lacking the ability to dynamically generate intermediate node paths based on link states. As a result, they cannot achieve adaptive optimization of link structures in complex network environments, thus affecting the overall network acceleration effect.

[0005] Therefore, there is an urgent need to provide a game network acceleration method for multi-terminal collaboration to solve the above problems. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to overcome the shortcomings of the prior art, which usually relies on real-time detection results to directly select candidate links during link switching, lacks the ability to predict the transmission performance of candidate links under the target transmission state, and does not perform path reconstruction processing on links with degraded performance. This makes it difficult to obtain stable alternative links when all available links fluctuate. At the same time, the prior art is mostly based on fixed topology or static connection relationship during path construction, and lacks the ability to dynamically generate intermediate node paths based on link status. As a result, it cannot achieve adaptive optimization of link structure in complex network environments, thus affecting the overall network acceleration effect. The present invention provides a game network acceleration method for multi-terminal collaboration.

[0007] To solve the above-mentioned technical problems, one technical solution adopted by the present invention is to provide a game network acceleration method for multi-terminal collaboration, comprising the following steps: Acquire network status data from multiple game terminals and corresponding edge node operation data, preprocess the network status data and the edge node operation data, and generate a standard network status sequence. Based on the standard network state sequence, construct terminal-side network quality descriptions and node-side network quality descriptions respectively. Based on the terminal-side network quality description and the node-side network quality description, a set of links between each game terminal and each edge node is constructed, and the terminal-side network quality description and the node-side network quality description are respectively associated with the corresponding links to form a transmission performance sequence of each link. Sequence analysis is performed on the transmission performance sequence of each link based on a continuous preset time window to generate an effective link set. Based on the set of valid links, the current game data is divided into data types, and link matching is performed according to the data type division results to form a candidate link subset; From the candidate link subset, a target transmission link is selected, and data transmission is performed. During data transmission, the standard network state sequence is continuously updated, and the transmission performance sequence of the target transmission link is continuously monitored. The target transmission link is marked as replaceable according to preset rules, and the link replacement is performed after the replaceable state continues to a preset stable stage. Otherwise, the current target transmission link remains unchanged, thus completing the game network acceleration for multi-terminal collaboration.

[0008] The present invention is further configured such that: the method for obtaining the network status data and the edge node operation data is as follows: by deploying a network status acquisition module on each game terminal side to monitor the current communication link of the terminal in real time, and to obtain latency performance, jitter performance, packet loss performance and bandwidth usage performance; by deploying an operation status acquisition module on the edge node side to monitor the resource usage and data processing queue of the node to obtain the node load status and node queuing status; and by synchronously collecting and uploading the network status data and the edge node operation data according to a unified time base. The preprocessing steps include time alignment, source identification, and order reordering of the network status data and the edge node operation data.

[0009] The present invention is further configured such that the method for constructing the terminal-side network quality description includes: The latency, jitter, packet loss, and bandwidth usage of each game terminal within a continuous time window are updated using a preset sampling period to form a corresponding transmission performance subsequence. Outlier removal and continuity verification are performed on the delay performance, jitter performance, packet loss performance and bandwidth usage performance in the transmission performance subsequence, respectively, to obtain the processed delay performance sequence, jitter performance sequence, packet loss performance sequence and bandwidth usage performance sequence; Based on the latency performance sequence, jitter performance sequence, packet loss performance sequence, and bandwidth usage performance sequence, the magnitude, direction, and degree of change within the continuous time window are calculated respectively, and a terminal-side network quality description corresponding to each game terminal is generated in the order of latency performance first, jitter performance second best, packet loss performance third best, and bandwidth usage performance third best. The steps for constructing the node-side network quality description are as follows: The node load status and node queuing status of each edge node are time-aligned using a unified time base, and the node load status and node queuing status of each edge node are updated in a continuous time window according to the preset sampling period to form a corresponding node running subsequence. Outlier removal and continuity verification are performed on the node load status and node queuing status in the node running sub-sequence, respectively, to obtain the processed node load status sequence and node queuing status sequence. Based on the node load state sequence and the node queuing state sequence, the magnitude, direction, and degree of change within a continuous time window are calculated respectively. The node load state sequence and the node queuing state sequence are then filtered and integrated step by step in the order of priority for node queuing state and second priority for node load state to generate a node-side network quality description corresponding to each edge node.

[0010] The present invention is further configured such that the method for forming the link set is as follows: Based on the terminal-side network quality description and the node-side network quality description, initial link pairs are generated according to the correspondence between each game terminal and each edge node, and the terminal-side network quality description and the node-side network quality description are mapped to the corresponding initial link pairs to form an initial link set; Based on the initial link set, the synchronization data of each initial link pair within a continuous time window is extracted according to a unified time base, and the synchronization data is time-aligned and combined to generate the basic link description of each initial link pair. Based on the link basic description, the latency performance, jitter performance, packet loss performance, bandwidth utilization performance, node load status and node queuing status of each initial link pair are compared, and the links are filtered in the order of priority of latency performance and node queuing status, second best of jitter performance and node load status, and third best of packet loss performance and bandwidth utilization performance to obtain a candidate link set. Based on the candidate link set, the consistency of the transmission performance of each initial link pair within a continuous time window is determined. Specifically, the direction of change of the delay performance, jitter performance, packet loss performance and bandwidth usage performance at adjacent times is compared. When the direction of change remains consistent within a preset number of consecutive times and the change amplitude is less than a preset range, the corresponding initial link pair is retained. When the direction of change does not remain consistent within a preset number of consecutive times or the change amplitude is greater than or equal to a preset range, the corresponding initial link pair is removed, thus obtaining the link set.

[0011] The present invention is further configured such that the method for generating the effective link set is as follows: Based on the link set, the node load status and node queuing status of each initial link pair in the link set are extracted according to the link basic description. The node load status and node queuing status of each initial link pair are compared within the same time window. When multiple initial link pairs correspond to the same edge node, the multiple initial link pairs are sorted according to the value of the node load status. If the node load status values ​​are the same, they are sorted according to the value of the node queuing status. The initial link pairs whose sorting results are in a preset later position are marked as competing link pairs and removed. For the remaining initial link pairs, the initial link pairs whose sorting results are in a preset earlier position are retained as valid link pairs, thus generating the valid link set.

[0012] The present invention is further configured such that: the current game data is divided into data types, and the data type division results include control data, state synchronization data, and image rendering data; The link matching rules include: executing in the following order: control data first, state synchronization data second, and image rendering data last; The specific steps for generating the candidate link subset are as follows: Firstly, select the first link that satisfies the continuous interaction of the control data from the set of valid links. Then, further filter the links that do not match the control data to ensure the continuous transmission of state synchronization data. Finally, filter the remaining links to ensure that the third link can carry the transmission of image rendering data. Integrate the first link, the second link, and the third link to generate a candidate link subset.

[0013] The present invention is further configured such that the method for selecting the target transmission link includes: Based on the candidate link subset, the corresponding transmission performance sequences are extracted for the first link, the second link and the third link respectively, and the transmission performance sequences are aligned under a unified time reference to form a standard link sequence for comparison. Based on the standard link sequence, the latency, jitter, packet loss, and bandwidth usage of the first link, the second link, and the third link are compared within a continuous time window. Based on the matching results of the control data, the state synchronization data, and the image rendering data, the latency and packet loss of the first link are used as the first judgment indicator, the packet loss and bandwidth usage of the second link are used as the second judgment indicator, and the bandwidth usage of the third link is used as the third judgment indicator. Based on the first judgment index, the second judgment index, and the third judgment index, the comprehensive performance value of the first link, the second link, and the third link within a continuous time window is calculated, wherein the comprehensive performance value is determined by the change magnitude and change direction of the corresponding judgment index, and the first link, the second link, and the third link are sorted according to the priority order of the comprehensive performance value. The ranking results are corrected by combining the historical stability of the first link, the second link, and the third link, and the link in the first position in the corrected ranking results is selected as the target transmission link.

[0014] The present invention is further configured such that: after the target transmission link is generated, an alternative link needs to be selected synchronously, and the method for selecting the alternative link includes: Based on the corrected sorting results, after selecting the link in the first position as the target transmission link, the remaining links in the sorting results other than the target transmission link are taken as a set of candidate alternative links, and the corresponding transmission performance sequence is extracted for each link in the set of candidate alternative links to form a candidate alternative link sequence. Based on the candidate alternative link sequence, each candidate alternative link is input into a preset edge computing simulation node, and the transmission performance sequence of the target transmission link in the data transmission process is used as the simulation input to perform synchronous transmission process simulation on each candidate alternative link, thereby obtaining the simulation transmission performance sequence corresponding to each candidate alternative link. Based on the simulated transmission performance sequence, the delay performance, jitter performance, packet loss performance, and bandwidth usage performance of each candidate alternative link are compared with the corresponding transmission performance of the target transmission link. When the candidate alternative link has at least one of the delay performance, jitter performance, or packet loss performance that changes more significantly within a continuous time window than the corresponding performance of the target transmission link, the corresponding candidate alternative link is marked as a link to be reassembled. Based on the link to be reconstructed, a path reconstruction process is performed on the corresponding link basic description. Specifically, at least one intermediate edge node is inserted between the game terminal and the edge node corresponding to the link to be reconstructed, and the inserted intermediate edge nodes are combined according to the bandwidth usage and node load status in the link basic description to generate at least one reconstruction path, and the reconstruction path is used as the reconstruction link. Based on the reconstructed link, a transmission process simulation is performed on it in a preset edge computing simulation node under the same input conditions as the target transmission link to obtain the reconstructed transmission performance sequence corresponding to the reconstructed link. Based on the reconstructed transmission performance sequence, the delay performance, jitter performance, and packet loss performance of the reconstructed link are compared item by item with the corresponding transmission performance of the target transmission link. When the change amplitude of each of the delay performance, jitter performance, and packet loss performance of the reconstructed link within a continuous time window is not greater than the change amplitude of the corresponding performance of the target transmission link, the reconstructed link is marked as an alternative link. The alternative link is associated with and stored with the target transmission link.

[0015] The present invention is further configured such that the method for generating the reconstructed path is as follows: Based on the link to be reassembled, the original transmission path between the game terminal and the edge node corresponding to the link to be reassembled is split. According to the basic description of the link, the bandwidth usage and node load status of each transmission segment in the original transmission path are extracted. Based on the bandwidth usage and node load status, replaceable transmission segments are determined to form a set of path segments to be replaced. Based on the set of path segments to be replaced, an extended search is performed on the edge nodes corresponding to each path segment to be replaced. A set of candidate intermediate edge nodes is established among the edge nodes involved in the set of path segments to be replaced. The set of candidate intermediate edge nodes is then filtered according to the node load status and node queuing status corresponding to each candidate intermediate edge node. Candidate intermediate edge nodes whose node load status change direction is consistent with the node load status change direction of the corresponding node in the path segment to be replaced within a continuous time window and whose node queuing status has not shown continuous growth are retained, thus generating an intermediate node set. Based on the set of intermediate nodes, path combination processing is performed on each path segment to be replaced in the set of path segments to be replaced. According to the connection relationship between each intermediate edge node in the set of intermediate nodes and the game terminal and the edge node, at least one candidate path is generated from the game terminal to the edge node via one or more intermediate edge nodes. Each candidate path is represented as a path sequence composed of multiple consecutive transmission segments to form a candidate path set. Based on the candidate path set, the continuity of the path sequence in each candidate path is checked. Specifically, the bandwidth usage and node load status of adjacent transmission segments in the path sequence are compared segment by segment. When the bandwidth usage of adjacent transmission segments changes in the same direction and the node load status does not change in the opposite direction, the corresponding path sequence is marked as a continuous path sequence. Otherwise, the corresponding path sequence is removed to generate a continuous path set. Based on the set of continuous paths, the number of transmission segments in each continuous path sequence and the transmission performance sequence corresponding to each transmission segment are compared. When the number of transmission segments in the continuous path sequence is less than the number of transmission segments in the corresponding original transmission path and the change amplitude of the transmission performance sequence corresponding to each transmission segment does not increase synchronously within the continuous time window, the corresponding continuous path sequence is determined as the reconstructed path.

[0016] The present invention is further configured such that: the preset rule includes: based on the transmission performance sequence of the target transmission link, comparing the change direction of delay performance, jitter performance and packet loss performance at each time step within a continuous time window, and counting the number of consecutive occurrences of the same change direction and the cumulative change amplitude; when at least two of the delay performance, jitter performance and packet loss performance show a continuous consistent change direction within the continuous time window and the corresponding change amplitude shows a continuous increasing trend during the continuous comparison, the target transmission link is marked as a replaceable state, and after the target transmission link is marked as a replaceable state, continuous monitoring of the transmission performance sequence continues; when the number of consecutive occurrences of the same change direction does not stop in subsequent time windows and the corresponding change amplitude does not show a reverse change, the replaceable state is determined to enter the preset stable stage; After link replacement is performed, the transmission performance sequence of the replacement link is continuously monitored. If the direction of change of each of the delay performance, jitter performance and packet loss performance of the replacement link within a continuous time window is consistent with the direction of change of the target transmission link before replacement and the corresponding change amplitude does not increase continuously, the replacement link is marked as a stable link and the current transmission state is maintained. Otherwise, the replacement link is marked as an abnormal link and the target transmission link is reselected based on the candidate link subset.

[0017] The beneficial effects of this invention are as follows: 1. This invention improves the reliability of link switching by introducing a candidate link simulation mechanism based on the target transmission link state to predict the transmission performance of alternative links under actual transmission conditions. 2. This invention performs path reconstruction processing on performance-degraded links and introduces intermediate edge nodes to generate reconstruction paths, thereby achieving dynamic adjustment of the link structure and improving transmission stability in complex network environments; 3. This invention improves the overall network acceleration effect by performing continuity verification and multi-index comparison and screening on the reconstructed path to ensure that the alternative link maintains stable transmission characteristics within a continuous time window. Attached Figure Description

[0018] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0019] The preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, so that the advantages and features of the present invention can be more easily understood by those skilled in the art, thereby providing a clearer and more explicit definition of the scope of protection of the present invention.

[0020] Please see Figure 1 A method for accelerating game networks for multi-terminal collaboration includes the following steps: The network status data of multiple game terminals and the corresponding edge node operation data are acquired, and the network status data and edge node operation data are preprocessed to generate a standard network status sequence. Based on the standard network state sequence, construct network quality descriptions for the terminal side and the node side respectively; Based on the terminal-side network quality description and the node-side network quality description, a set of links between each game terminal and each edge node is constructed. The terminal-side network quality description and the node-side network quality description are associated with the corresponding links to form a transmission performance sequence of each link. The transmission performance sequence of each link is analyzed based on a continuous preset time window to generate an effective link set. Based on the set of valid links, the current game data is divided into data types, and link matching is performed according to the data type division results to form a candidate link subset; From the subset of candidate links, select the target transmission link and perform data transmission; During data transmission, the standard network status sequence is continuously updated, and the transmission performance sequence of the target transmission link is continuously monitored. The target transmission link is marked as replaceable according to preset rules, and the link replacement is performed after the replaceable state continues to a preset stable stage. Otherwise, the current target transmission link remains unchanged, thus completing the game network acceleration for multi-terminal collaboration.

[0021] This method constructs terminal-side network quality descriptions and node-side network quality descriptions based on standard network state sequences, and generates effective link sets and candidate link subsets by combining transmission performance sequences. This enables dynamic selection and replacement control of target transmission links, thereby improving the stability and continuity of game data transmission and reducing the impact of link fluctuations on the transmission process.

[0022] In one implementation, the method for obtaining network status data and edge node operation data is as follows: by deploying a network status acquisition module on each game terminal side to monitor the current communication link of the terminal in real time, and to obtain latency performance, jitter performance, packet loss performance and bandwidth usage performance; by deploying an operation status acquisition module on the edge node side to monitor the resource usage and data processing queue of the node to obtain the node load status and node queuing status; and by synchronously collecting and uploading network status data and edge node operation data according to a unified time base. Game terminals specifically include: personal computer terminals, smartphone terminals, tablet terminals, game console terminals, and cloud gaming access terminals. All game terminals have network communication interfaces and can execute game client programs to generate game data.

[0023] The network status acquisition module is specifically a network performance monitoring component deployed in the game terminal. It obtains round-trip latency measurements, packet arrival interval differences, packet sending and receiving statistics, and port bandwidth usage statistics through the operating system network interface to form latency performance, jitter performance, packet loss performance, and bandwidth usage performance.

[0024] The operational status acquisition module is specifically a system resource monitoring component deployed in edge nodes. It generates node load status and node queuing status by acquiring CPU utilization, memory usage, network interface queue length, and task processing queue length.

[0025] The unified time base is a unified timestamp sequence generated synchronously based on the network time protocol, and the network status data and edge node operation data are aligned using the unified timestamp sequence.

[0026] The preprocessing steps include time alignment, source identification, and order reordering of network status data and edge node operation data.

[0027] In one implementation, the method for constructing a terminal-side network quality description includes: The latency, jitter, packet loss, and bandwidth usage of each game terminal within a continuous time window are updated using a preset sampling period to form a corresponding transmission performance subsequence. Preset sampling period: The preferred value is 20 milliseconds to 100 milliseconds, with 40 milliseconds to 60 milliseconds being the preferred range. The reason is that the frequency at which the network status acquisition module acquires latency performance, jitter performance, packet loss performance, and bandwidth usage performance needs to match the data interaction frequency of the game terminal. When the preset sampling period is less than 20 milliseconds, the network status data update frequency is too high, which can easily introduce instantaneous fluctuation interference and increase the computational burden. When the preset sampling period is greater than 100 milliseconds, the network status data responds to link changes in a lag, which affects the accuracy of the construction of network quality descriptions on the terminal side and the node side. Therefore, the preset sampling period is limited to the above range to balance the change capture capability and computational stability.

[0028] Continuous time window: The preferred value is a time length of 5 to 15 times the preset sampling period, with 8 to 12 times the preset sampling period being the preferred range. The reason is that the continuous time window is used to perform sequence analysis on the transmission performance sequence and generate a set of effective links. If the continuous time window is too short, it is impossible to form a stable trend judgment and it is easily affected by single fluctuations. If the continuous time window is too long, it will weaken the response capability to link changes and affect the dynamic selection efficiency of the target transmission link. Therefore, by limiting the continuous time window to a multiple range of the preset sampling period, the calculation of the change amplitude, change direction and fluctuation degree can reflect the continuous trend and maintain a timely response to link changes.

[0029] Outlier removal and continuity verification are performed on the delay performance, jitter performance, packet loss performance and bandwidth usage performance in the transmission performance subsequence, respectively, to obtain the processed delay performance sequence, jitter performance sequence, packet loss performance sequence and bandwidth usage performance sequence; For the latency, jitter, packet loss, and bandwidth usage subsequences, outlier removal and continuity verification are performed respectively, as follows: For each performance value in the transmission performance subsequence, an original sequence is constructed according to a unified time base, and the difference sequence between adjacent time points is calculated. Performance values ​​whose absolute values ​​exceed twice the average value of the difference sequence are marked as abnormal performance values. The original sequence is replaced based on the abnormal performance values, and the abnormal performance values ​​are replaced with the average of the performance values ​​of the adjacent time points to form a corrected sequence. The time interval between adjacent moments is calculated based on the corrected sequence, and it is determined whether there is a time interval greater than twice the preset sampling period. If such a situation exists, padding values ​​are inserted at the corresponding positions to form a continuous sequence. Based on the consistency of the change direction of each adjacent time step in the continuous sequence, when the change direction changes alternately within three consecutive time steps, the corresponding sequence segment is marked as a discontinuous segment and removed, thus obtaining the delay performance sequence, jitter performance sequence, packet loss performance sequence and bandwidth usage performance sequence.

[0030] Based on the latency performance sequence, jitter performance sequence, packet loss performance sequence, and bandwidth usage performance sequence, the magnitude, direction, and degree of change within a continuous time window are calculated respectively. Then, the terminal-side network quality description corresponding to each game terminal is generated in the order of latency performance first, jitter performance second best, packet loss performance third best, and bandwidth usage performance third best. Based on the latency performance sequence, jitter performance sequence, packet loss performance sequence, and bandwidth usage performance sequence, calculate the magnitude, direction, and degree of change within a continuous time window. The specific calculation steps are as follows: The difference between any two adjacent time points is calculated based on the delay performance sequence to form a delay difference sequence. The absolute values ​​of the delay difference sequences are accumulated and divided by the time window length to obtain the delay change amplitude. Based on the delay difference sequence, the positive and negative values ​​of the difference are determined to generate a delay change direction sequence, and the difference between the number of occurrences of the positive direction and the number of occurrences of the negative direction is used as the delay direction determination value. The degree of deviation between the absolute value of the delay difference and the magnitude of the delay change is calculated based on the delay difference sequence, and the summation is used to form the degree of delay fluctuation; Based on the jitter performance sequence, packet loss performance sequence, and bandwidth usage performance sequence, repeat the above calculation steps to obtain the corresponding change amplitude, change direction, and fluctuation degree.

[0031] The steps for constructing a node-side network quality description are as follows: The node load status and node queuing status of each edge node are time-aligned using a unified time base, and the node load status and node queuing status of each edge node are updated in a continuous time window according to a preset sampling period to form a corresponding node running subsequence. Outlier removal and continuity verification are performed on the node load status and node queuing status in the node running subsequence, respectively, to obtain the processed node load status sequence and node queuing status sequence. The magnitude, direction, and degree of change of node load state sequence and node queuing state sequence within a continuous time window are calculated respectively. The node load state sequence and node queuing state sequence are then filtered and integrated step by step in the order of priority of node queuing state and second priority of node load state to generate a node-side network quality description corresponding to each edge node.

[0032] The magnitude, direction, and degree of change of node load state sequences and node queuing state sequences within a continuous time window are calculated, respectively. The specific calculation method is as follows: The load difference between adjacent time points is calculated based on the node load state sequence to form a load difference sequence. The absolute values ​​of the load difference sequence are summed and divided by the time window length to obtain the load change amplitude. The direction of load change is determined according to the sign of the difference. The queue length difference is calculated based on the node queue state sequence to form a queue difference sequence. The absolute value of the queue difference sequence is accumulated and divided by the time window length to obtain the queue change amplitude and determine the queue change direction. The degree of deviation between the load difference sequence and the queue difference sequence and their respective change ranges is calculated and accumulated to obtain the corresponding fluctuation degree.

[0033] In one implementation, the link set is formed as follows: Based on the terminal-side network quality description and the node-side network quality description, initial link pairs are generated according to the correspondence between each game terminal and each edge node, and the terminal-side network quality description and the node-side network quality description are mapped to the corresponding initial link pairs to form an initial link set; The steps for generating the initial link pair are as follows: A terminal node mapping table is established based on the connection relationship between the game terminal and the edge node, and an initial link pair is formed by a one-to-one combination of terminal identifier and node identifier based on the terminal node mapping table. Based on the initial link pair, extract the corresponding terminal-side network quality description and node-side network quality description, and combine them to form a link description pair; Based on the link description, the corresponding time series are aligned according to a unified time base and a synchronous link sequence is constructed. An initial set of links is generated based on the synchronous link sequence, including latency performance, jitter performance, packet loss performance, and bandwidth usage performance.

[0034] Based on the initial link set, the synchronization data of each initial link pair within a continuous time window is extracted according to a unified time base, and the synchronization data is time-aligned and combined to generate the basic link description of each initial link pair. Based on the basic description of the links, the latency, jitter, packet loss, bandwidth usage, node load status, and node queuing status of each initial link pair are compared. The links are then selected in the following order: latency and node queuing status are prioritized, jitter and node load status are second best, and packet loss and bandwidth usage are third best, to obtain a set of candidate links. Based on the candidate link set, the consistency of the transmission performance of each initial link pair within a continuous time window is determined. Specifically, the direction of change of the delay performance, jitter performance, packet loss performance and bandwidth usage performance at adjacent time points is compared. When the direction of change remains consistent within a preset number of consecutive times and the change amplitude is less than a preset range, the corresponding initial link pair is retained. When the direction of change does not remain consistent within a preset number of consecutive times or the change amplitude is greater than or equal to a preset range, the corresponding initial link pair is removed, thus obtaining the link set.

[0035] Preset number of times: The preferred value is 3 to 5 times. The basis is to ensure the continuity of the change direction and avoid misjudgment caused by short-term jitter. When the number of continuous changes reaches more than 3 times, the trend can be judged to be stable.

[0036] Preset range: preferably 0.5 to 1.5 times the average value of the change amplitude, based on the principle that limiting the range of change amplitude fluctuations can filter out abnormal fluctuations and retain normal fluctuation trends.

[0037] The performance of latency, jitter, packet loss, and bandwidth usage at adjacent time points is compared in terms of direction of change. The direction of change comparison includes: calculating the difference based on the performance values ​​at adjacent time points and determining the direction as positive if the difference is greater than 0 and negative if the difference is less than 0; constructing a direction sequence based on the positive and negative directions and performing continuity statistics on the direction sequence; determining whether the continuous directions are consistent based on the direction sequence, marking them as stable directions when they are consistent, and otherwise marking them as changing directions.

[0038] In one implementation, the method for generating the effective link set is as follows: Based on the link set, the node load status and node queuing status of each initial link pair in the link set are extracted according to the basic link description. The node load status and node queuing status of each initial link pair are compared within the same time window. When multiple initial link pairs correspond to the same edge node, the multiple initial link pairs are sorted according to the value of the node load status. If the node load status values ​​are the same, they are sorted according to the value of the node queuing status. The initial link pairs that are in the preset later order position are marked as competing link pairs and removed. For the remaining initial link pairs, the initial link pairs that are in the preset earlier order position are retained as valid link pairs according to the sorting results of the node load status and node queuing status, thus generating a valid link set.

[0039] Preset subsequent position: preferably in the last 30% of the sorting results, based on the elimination of links with poor performance; Preset preceding position: preferably in the first 50% of the sorting results, based on the principle of reserving stable links for subsequent matching.

[0040] In one implementation, the current game data is divided into data types, and the data type division results include control data, state synchronization data, and image rendering data. Control data specifically includes: key command data, movement control data, and operation response data; The status synchronization data specifically includes: character location data, status update data, and physical synchronization data; Image rendering data specifically includes: texture data, frame rendering data, and video stream data.

[0041] The rules for link matching include: prioritizing control data, followed by state synchronization data, and then image rendering data; The specific steps for generating the candidate link subset are as follows: First, select the first link that satisfies the continuous interaction of control data from the set of valid links. For the links that do not match the control data, further filter the second link that satisfies the continuous transmission of state synchronization data. Then, filter the third link that can carry the transmission of image rendering data from the remaining links. Integrate the first link, the second link and the third link to generate a candidate link subset.

[0042] The requirement for continuous interaction of control data specifically means that the difference between adjacent time points in the delay performance sequence is not greater than the delay change amplitude and the number of packet loss changes in the packet loss performance sequence does not exceed once. Satisfying continuous transmission of state synchronization data specifically means that the packet loss performance sequence changes in a continuous and consistent direction, and the bandwidth usage performance sequence does not show a decreasing trend within the time window; The ability to carry image rendering data transmission specifically means that the average value of the bandwidth usage sequence within a continuous time window is greater than the historical average bandwidth of the corresponding link, and the degree of fluctuation does not show a continuous increase.

[0043] In one specific embodiment, by performing the above determination on the three types of links respectively, the first link that satisfies the continuous interaction of control data is selected first, the second link that satisfies the continuous transmission of state synchronization data is selected from the remaining links, and finally the third link that satisfies the transmission of image rendering data is selected from the remaining links to form a candidate link subset.

[0044] In one implementation, the method for selecting the target transmission link includes: Based on the candidate link subset, the corresponding transmission performance sequences are extracted for the first link, the second link and the third link respectively, and the transmission performance sequences are aligned under a unified time reference to form a standard link sequence for comparison. The corresponding transmission performance sequences are extracted for the first link, the second link, and the third link respectively. The method for extracting the transmission performance sequences is as follows: based on the link basic descriptions corresponding to the first link, the second link, and the third link, the delay performance, jitter performance, packet loss performance, and bandwidth usage performance are extracted and sorted according to a unified time base to form the transmission performance sequences.

[0045] Based on the standard link sequence, the latency, jitter, packet loss, and bandwidth usage of the first, second, and third links within a continuous time window are compared. Based on the matching results of control data, state synchronization data, and image rendering data, the latency and packet loss of the first link are used as the first judgment indicator, the packet loss and bandwidth usage of the second link are used as the second judgment indicator, and the bandwidth usage of the third link is used as the third judgment indicator. The first criterion is a combination sequence of latency performance and packet loss performance, which is calculated as the sum of the latency change magnitude and the number of packet loss changes. The second criterion is a combination sequence of packet loss performance and bandwidth usage performance, which is calculated as the ratio of the number of packet loss changes to the bandwidth change magnitude. The third criterion is the bandwidth usage performance sequence, which is calculated as the difference between the bandwidth change magnitude and the degree of fluctuation.

[0046] In one specific embodiment, the ranking result is determined by calculating three types of judgment indicators and comparing their corresponding link performance.

[0047] Based on the first, second, and third judgment indicators, the comprehensive performance values ​​of the first link, the second link, and the third link are calculated within a continuous time window. The comprehensive performance value is determined by the change magnitude and direction of the corresponding judgment indicator, and the first link, the second link, and the third link are sorted according to the priority order of the comprehensive performance values. The calculation steps for the comprehensive performance value are as follows: Comprehensive performance value = magnitude of change in the judgment indicator + number of consecutive changes in the direction of the judgment indicator - degree of fluctuation in the judgment indicator.

[0048] The ranking results are corrected by combining the historical stability of the first, second, and third links, and the link in the first position in the corrected ranking results is selected as the target transmission link.

[0049] In a preferred embodiment, the historical stability of the link is determined based on the fluctuation of the transmission performance sequence of the corresponding link within a preset time range and the occurrence of link switching. When there are multiple links with the same ranking result, the link with higher historical stability and lower node load status is selected as the target transmission link.

[0050] In one implementation, after the target transmission link is generated, an alternative link needs to be selected synchronously. The method for selecting the alternative link includes: Based on the corrected sorting results, after selecting the link in the first position as the target transmission link, the remaining links in the sorting results other than the target transmission link are used as a candidate alternative link set, and the corresponding transmission performance sequence is extracted from each link in the candidate alternative link set to form a candidate alternative link sequence. Based on the candidate alternative link sequence, each candidate alternative link is input into a preset edge computing simulation node, and the transmission performance sequence of the target transmission link during data transmission is used as the simulation input to perform synchronous transmission process simulation on each candidate alternative link, thereby obtaining the simulation transmission performance sequence corresponding to each candidate alternative link. The pre-configured edge computing simulation node is specifically a network simulation module deployed in an edge server or cloud server, used to simulate the link transmission process based on historical transmission performance data.

[0051] Based on the simulated transmission performance sequence, the delay performance, jitter performance, packet loss performance and bandwidth usage performance of each candidate alternative link are compared with the corresponding transmission performance of the target transmission link. When the candidate alternative link has at least one of the delay performance, jitter performance or packet loss performance with a greater change in the corresponding performance of the target transmission link within a continuous time window, the corresponding candidate alternative link is marked as a link to be reassembled. Based on the link to be reconstructed, path reconstruction processing is performed on its corresponding link basic description. Specifically, at least one intermediate edge node is inserted between the game terminal and the edge node corresponding to the link to be reconstructed, and the inserted intermediate edge nodes are combined according to the bandwidth usage and node load status in the link basic description to generate at least one reconstruction path, and the reconstruction path is used as the reconstruction link. Based on the reconstructed link, the transmission process of the reconstructed link is simulated in a preset edge computing simulation node under the same input conditions as the target transmission link, and the reconstructed transmission performance sequence corresponding to the reconstructed link is obtained. Based on the reconstructed transmission performance sequence, the delay performance, jitter performance, and packet loss performance of the reconstructed link are compared with the corresponding transmission performance of the target transmission link item by item. When the change amplitude of each item in the delay performance, jitter performance, and packet loss performance of the reconstructed link within the continuous time window is not greater than the change amplitude of the corresponding performance of the target transmission link, the reconstructed link is marked as an alternative link. Alternative links are associated with the target transmission link and stored for use in performing link switching when the target transmission link is marked as a replaceable link.

[0052] In one implementation, the method for generating the reconstructed path is as follows: Based on the link to be reassembled, the original transmission path between the game terminal and the edge node corresponding to the link to be reassembled is split. According to the basic description of the link, the bandwidth usage and node load status of each transmission segment in the original transmission path are extracted. Based on the bandwidth usage and node load status, the replaceable transmission segments are determined to form a set of path segments to be replaced. Based on the set of path segments to be replaced, an extended search is performed on the edge nodes corresponding to each path segment to be replaced. A set of candidate intermediate edge nodes is established among the edge nodes involved in the set of path segments to be replaced. The set of candidate intermediate edge nodes is then filtered according to the node load status and node queuing status of each candidate intermediate edge node. Candidate intermediate edge nodes whose node load status changes in the same direction as the corresponding node load status changes in the path segment to be replaced within a continuous time window and whose node queuing status has not shown continuous growth are retained, thus generating an intermediate node set. The method for establishing the candidate intermediate edge node set is as follows: Based on the set of path segments to be replaced, extract the relevant edge nodes to form the original node set; The candidate node set is obtained by expanding the adjacent edge nodes based on the original node set; Based on the candidate node set, extract the node load status and node queuing status, and filter nodes that meet the condition of consistent change direction to form a candidate intermediate edge node set.

[0053] Based on the intermediate node set, each path segment to be replaced in the set of path segments to be replaced is processed by path combination. According to the connection relationship between each intermediate edge node in the intermediate node set and the game terminal and edge nodes, at least one candidate path is generated from the game terminal to the edge node via one or more intermediate edge nodes. Each candidate path is represented as a path sequence composed of multiple continuous transmission segments to form a candidate path set. Based on the candidate path set, the continuity of the path sequences in each candidate path is verified. Specifically, the bandwidth occupancy and node load status of adjacent transmission segments in the path sequence are compared segment by segment. When the bandwidth occupancy of adjacent transmission segments changes in the same direction and the node load status does not change in the opposite direction, the corresponding path sequence is marked as a continuous path sequence. Otherwise, the corresponding path sequence is removed, and a continuous path set is generated. Here, "otherwise" means that the bandwidth occupancy of adjacent transmission segments does not change in the same direction or the node load status changes in the opposite direction.

[0054] Based on the continuous path set, the number of transmission segments in each continuous path sequence and the corresponding transmission performance sequence of each transmission segment are compared. When the number of transmission segments in the continuous path sequence is less than the number of transmission segments in the corresponding original transmission path and the change amplitude of the transmission performance sequence of each transmission segment within the continuous time window does not increase synchronously, the corresponding continuous path sequence is determined as the reconstructed path.

[0055] In one implementation, the preset rules include: based on the transmission performance sequence of the target transmission link, comparing the change direction of delay performance, jitter performance and packet loss performance at each time step within a continuous time window, and counting the number of consecutive occurrences of the same change direction and the cumulative change amplitude. When at least two of the delay performance, jitter performance and packet loss performance show a continuous consistent change direction within the continuous time window and the corresponding change amplitude shows a continuous increasing trend during the continuous comparison, the target transmission link is marked as a replaceable state. After the target transmission link is marked as a replaceable state, continuous monitoring of the transmission performance sequence continues. When the number of consecutive occurrences of the same change direction does not stop in the subsequent time window and the corresponding change amplitude does not change in the opposite direction, the replaceable state is determined to enter a preset stable stage. When at least two of the latency, jitter, and packet loss performance parameters show a consistent direction of change within a continuous time window, and the corresponding magnitude of change shows a continuously increasing trend during continuous comparison, specifically: Calculate the sequence of changing directions based on latency performance, jitter performance, and packet loss performance, and count the number of consecutive consistent sequences. The condition is met when at least two changes in the same direction are consecutively repeated a preset number of times and the corresponding change amplitudes show an increasing relationship in consecutive time intervals. The target transmission link is marked as replaceable based on the results of meeting the conditions.

[0056] The preset stable phase is the time interval in which the direction of change remains consistent for two consecutive time windows under the replaceable state, and the magnitude of change does not reverse.

[0057] After link replacement is performed, the transmission performance sequence of the replacement link is continuously monitored. If the direction of change of each of the replacement link's delay performance, jitter performance and packet loss performance within a continuous time window is consistent with the direction of change of the target transmission link before replacement and the corresponding change amplitude does not show a continuous increase, the replacement link is marked as a stable link and its current transmission state is maintained. Otherwise, the replacement link is marked as an abnormal link and the target transmission link is reselected based on the candidate link subset.

[0058] In a preferred embodiment, after the alternative link is used for data transmission, the method further includes: continuously updating the transmission performance sequence corresponding to the target transmission link based on the standard network state sequence, and comparing the change direction of the delay performance, jitter performance, and packet loss performance of the target transmission link at each time step within a continuous time window, and counting the number of consecutive occurrences of the same change direction and the corresponding change amplitude. When the change direction of each item in the delay performance, jitter performance, and packet loss performance of the target transmission link within the continuous time window is consistent with the change direction corresponding to the current transmission state of the alternative link and the corresponding change amplitude does not show a continuous increase, the target transmission link is marked as a back-switch candidate link, and continuous monitoring is performed on the back-switch candidate link in subsequent time windows. When the number of consecutive occurrences of the same change direction does not stop in subsequent time windows and the corresponding change amplitude does not show a reverse change, the target transmission link is re-determined as the target transmission link and the link back-switch is performed; otherwise, the alternative link continues to be used for data transmission.

[0059] In one specific embodiment, network status data of three game terminals and node operation data of two edge nodes are acquired, and standard network status sequence is generated by collecting and preprocessing according to a preset sampling period of 50 milliseconds. Under the condition of a continuous time window of 500 milliseconds, terminal-side network quality description and node-side network quality description are constructed and a link set is formed. Based on the link set, the transmission performance sequences of the first link, the second link and the third link are generated and the transmission performance sequences are aligned to form a standard link sequence. Within a continuous time window, the latency variation of the first link is calculated to be 8 milliseconds and the number of packet loss changes is 2, resulting in a first judgment index of 10. The number of packet loss changes of the second link is calculated to be 1 and the bandwidth variation is 5, resulting in a second judgment index of 0.2. The bandwidth variation of the third link is calculated to be 6 and the fluctuation level is 3, resulting in a third judgment index of 3. Further, based on the comprehensive performance value calculation rules, the comprehensive performance value of the first link is 10 plus the number of consecutive consistent times 4 minus the fluctuation level 2, which equals 12. The comprehensive performance value of the second link is 0.2 plus the number of consecutive consistent times 5 minus the fluctuation level 1, which equals 4.2. The comprehensive performance value of the third link is 3 plus the number of consecutive consistent times 3 minus the fluctuation level 1, which equals 5. Thus, the first link is determined to be the target transmission link. Based on the ranking results, the second and third links are used as candidate alternative links and input into the edge computing simulation node for simulation to obtain a simulated transmission performance sequence. In the comparison, when the latency variation of the second link is 12 milliseconds, which is greater than the 8 milliseconds corresponding to the target transmission link, the second link is marked as the link to be reconstructed. By inserting an intermediate edge node, a reconstruction path containing two transmission segments is generated, and a reconstruction link is formed. After simulating the reconstructed link, the latency variation was 7 milliseconds, the jitter variation was 2, and the number of packet loss changes was 1, all of which were no greater than the corresponding performance of the target transmission link. Therefore, the reconstructed link was marked as an alternative link and associated with it. In subsequent monitoring, when the target transmission link's latency and packet loss performance changed in the same direction for 3 consecutive time windows and the change amplitude continued to increase, it was marked as replaceable. After remaining unchanged for two consecutive time windows and entering a preset stable stage, the link was replaced and switched to the alternative link for data transmission. At the same time, the alternative link was continuously monitored, and the current transmission state was maintained when the change direction was consistent and the change amplitude did not increase. Otherwise, the target transmission link was reselected based on the candidate link subset.

[0060] By constructing terminal-side and node-side network quality descriptions based on standard network state sequences and combining judgment indicators and comprehensive performance values, dynamic selection and path reconstruction of target transmission links and alternative links are achieved, thereby improving the determinism of link switching and enhancing the continuity and stability of the transmission process.

[0061] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A game network acceleration method for multi-terminal collaboration, characterized in that: Includes the following steps: Acquire network status data from multiple game terminals and corresponding edge node operation data, preprocess the network status data and the edge node operation data, and generate a standard network status sequence. Based on the standard network state sequence, construct terminal-side network quality descriptions and node-side network quality descriptions respectively. Based on the terminal-side network quality description and the node-side network quality description, a set of links between each game terminal and each edge node is constructed, and the terminal-side network quality description and the node-side network quality description are respectively associated with the corresponding links to form a transmission performance sequence of each link. Sequence analysis is performed on the transmission performance sequence of each link based on a continuous preset time window to generate an effective link set. Based on the set of valid links, the current game data is divided into data types, and link matching is performed according to the data type division results to form a candidate link subset; From the candidate link subset, a target transmission link is selected, and data transmission is performed. During data transmission, the standard network state sequence is continuously updated, and the transmission performance sequence of the target transmission link is continuously monitored. The target transmission link is marked as replaceable according to preset rules, and the link replacement is performed after the replaceable state continues to a preset stable stage. Otherwise, the current target transmission link remains unchanged, thus completing the game network acceleration for multi-terminal collaboration.

2. The game network acceleration method for multi-terminal collaboration according to claim 1, characterized in that: The method for obtaining the network status data and the edge node operation data is as follows: by deploying a network status acquisition module on each game terminal side to monitor the current communication link of the terminal in real time, and to obtain latency performance, jitter performance, packet loss performance and bandwidth usage performance; by deploying an operation status acquisition module on the edge node side to monitor the resource usage and data processing queue of the node to obtain the node load status and node queuing status; and by synchronously collecting and uploading the network status data and the edge node operation data according to a unified time base. The preprocessing steps include time alignment, source identification, and order reordering of the network status data and the edge node operation data.

3. The game network acceleration method for multi-terminal collaboration according to claim 2, characterized in that: The method for constructing the terminal-side network quality description includes: The latency, jitter, packet loss, and bandwidth usage of each game terminal within a continuous time window are updated using a preset sampling period to form a corresponding transmission performance subsequence. Outlier removal and continuity verification are performed on the delay performance, jitter performance, packet loss performance and bandwidth usage performance in the transmission performance subsequence, respectively, to obtain the processed delay performance sequence, jitter performance sequence, packet loss performance sequence and bandwidth usage performance sequence; Based on the latency performance sequence, jitter performance sequence, packet loss performance sequence, and bandwidth usage performance sequence, the magnitude, direction, and degree of change within the continuous time window are calculated respectively, and a terminal-side network quality description corresponding to each game terminal is generated in the order of latency performance first, jitter performance second best, packet loss performance third best, and bandwidth usage performance third best. The steps for constructing the node-side network quality description are as follows: The node load status and node queuing status of each edge node are time-aligned using a unified time base, and the node load status and node queuing status of each edge node are updated in a continuous time window according to the preset sampling period to form a corresponding node running subsequence. Outlier removal and continuity verification are performed on the node load status and node queuing status in the node running sub-sequence, respectively, to obtain the processed node load status sequence and node queuing status sequence. Based on the node load state sequence and the node queuing state sequence, the magnitude, direction, and degree of change within a continuous time window are calculated respectively. The node load state sequence and the node queuing state sequence are then filtered and integrated step by step in the order of priority for node queuing state and second priority for node load state to generate a node-side network quality description corresponding to each edge node.

4. The game network acceleration method for multi-terminal collaboration according to claim 3, characterized in that: The method for forming the link set is as follows: Based on the terminal-side network quality description and the node-side network quality description, initial link pairs are generated according to the correspondence between each game terminal and each edge node, and the terminal-side network quality description and the node-side network quality description are mapped to the corresponding initial link pairs to form an initial link set; Based on the initial link set, the synchronization data of each initial link pair within a continuous time window is extracted according to a unified time base, and the synchronization data is time-aligned and combined to generate the basic link description of each initial link pair. Based on the link basic description, the latency performance, jitter performance, packet loss performance, bandwidth utilization performance, node load status and node queuing status of each initial link pair are compared, and the links are filtered in the order of priority of latency performance and node queuing status, second best of jitter performance and node load status, and third best of packet loss performance and bandwidth utilization performance to obtain a candidate link set. Based on the candidate link set, the consistency of the transmission performance of each initial link pair within a continuous time window is determined. Specifically, the direction of change of the delay performance, jitter performance, packet loss performance and bandwidth usage performance at adjacent times is compared. When the direction of change remains consistent within a preset number of consecutive times and the change amplitude is less than a preset range, the corresponding initial link pair is retained. When the direction of change does not remain consistent within a preset number of consecutive times or the change amplitude is greater than or equal to a preset range, the corresponding initial link pair is removed, thus obtaining the link set.

5. A game network acceleration method for multi-terminal collaboration according to claim 4, characterized in that: The method for generating the set of valid links is as follows: Based on the link set, the node load status and node queuing status of each initial link pair in the link set are extracted according to the link basic description. The node load status and node queuing status of each initial link pair are compared within the same time window. When multiple initial link pairs correspond to the same edge node, the multiple initial link pairs are sorted according to the value of the node load status. If the node load status values ​​are the same, they are sorted according to the value of the node queuing status. The initial link pairs whose sorting results are in a preset later position are marked as competing link pairs and removed. For the remaining initial link pairs, the initial link pairs whose sorting results are in a preset earlier position are retained as valid link pairs, thus generating the valid link set.

6. A game network acceleration method for multi-terminal collaboration according to claim 5, characterized in that: The current game data is classified into data types, and the classification results include control data, state synchronization data, and image rendering data. The link matching rules include: executing in the following order: control data first, state synchronization data second, and image rendering data last; The specific steps for generating the candidate link subset are as follows: Firstly, select the first link that satisfies the continuous interaction of the control data from the set of valid links. Then, further filter the links that do not match the control data to ensure the continuous transmission of state synchronization data. Finally, filter the remaining links to ensure that the third link can carry the transmission of image rendering data. Integrate the first link, the second link, and the third link to generate a candidate link subset.

7. A game network acceleration method for multi-terminal collaboration according to claim 6, characterized in that: The method for selecting the target transmission link includes: Based on the candidate link subset, the corresponding transmission performance sequences are extracted for the first link, the second link and the third link respectively, and the transmission performance sequences are aligned under a unified time reference to form a standard link sequence for comparison. Based on the standard link sequence, the latency, jitter, packet loss, and bandwidth usage of the first link, the second link, and the third link are compared within a continuous time window. Based on the matching results of the control data, the state synchronization data, and the image rendering data, the latency and packet loss of the first link are used as the first judgment indicator, the packet loss and bandwidth usage of the second link are used as the second judgment indicator, and the bandwidth usage of the third link is used as the third judgment indicator. Based on the first judgment index, the second judgment index, and the third judgment index, the comprehensive performance value of the first link, the second link, and the third link within a continuous time window is calculated, wherein the comprehensive performance value is determined by the change magnitude and change direction of the corresponding judgment index, and the first link, the second link, and the third link are sorted according to the priority order of the comprehensive performance value. The ranking results are corrected by combining the historical stability of the first link, the second link and the third link, and the link in the first position in the corrected ranking results is selected as the target transmission link.

8. A game network acceleration method for multi-terminal collaboration according to claim 7, characterized in that: After the target transmission link is generated, an alternative link needs to be selected synchronously. The selection method for the alternative link includes: Based on the corrected sorting results, after selecting the link in the first position as the target transmission link, the remaining links in the sorting results other than the target transmission link are taken as a set of candidate alternative links, and the corresponding transmission performance sequence is extracted for each link in the set of candidate alternative links to form a candidate alternative link sequence. Based on the candidate alternative link sequence, each candidate alternative link is input into a preset edge computing simulation node, and the transmission performance sequence of the target transmission link in the data transmission process is used as the simulation input to perform synchronous transmission process simulation on each candidate alternative link, thereby obtaining the simulation transmission performance sequence corresponding to each candidate alternative link. Based on the simulated transmission performance sequence, the delay performance, jitter performance, packet loss performance, and bandwidth usage performance of each candidate alternative link are compared with the corresponding transmission performance of the target transmission link. When the candidate alternative link has at least one of the delay performance, jitter performance, or packet loss performance that changes more significantly within a continuous time window than the corresponding performance of the target transmission link, the corresponding candidate alternative link is marked as a link to be reassembled. Based on the link to be reconstructed, a path reconstruction process is performed on the corresponding link basic description. Specifically, at least one intermediate edge node is inserted between the game terminal and the edge node corresponding to the link to be reconstructed, and the inserted intermediate edge nodes are combined according to the bandwidth usage and node load status in the link basic description to generate at least one reconstruction path, and the reconstruction path is used as the reconstruction link. Based on the reconstructed link, a transmission process simulation is performed on it in a preset edge computing simulation node under the same input conditions as the target transmission link to obtain the reconstructed transmission performance sequence corresponding to the reconstructed link. Based on the reconstructed transmission performance sequence, the delay performance, jitter performance, and packet loss performance of the reconstructed link are compared item by item with the corresponding transmission performance of the target transmission link. When the change amplitude of each of the delay performance, jitter performance, and packet loss performance of the reconstructed link within a continuous time window is not greater than the change amplitude of the corresponding performance of the target transmission link, the reconstructed link is marked as an alternative link. The alternative link is associated with and stored with the target transmission link.

9. A game network acceleration method for multi-terminal collaboration according to claim 8, characterized in that: The method for generating the reconstructed path is as follows: Based on the link to be reassembled, the original transmission path between the game terminal and the edge node corresponding to the link to be reassembled is split. According to the basic description of the link, the bandwidth usage and node load status of each transmission segment in the original transmission path are extracted. Based on the bandwidth usage and node load status, replaceable transmission segments are determined to form a set of path segments to be replaced. Based on the set of path segments to be replaced, an extended search is performed on the edge nodes corresponding to each path segment to be replaced. A set of candidate intermediate edge nodes is established among the edge nodes involved in the set of path segments to be replaced. The set of candidate intermediate edge nodes is then filtered according to the node load status and node queuing status corresponding to each candidate intermediate edge node. Candidate intermediate edge nodes whose node load status change direction is consistent with the node load status change direction of the corresponding node in the path segment to be replaced within a continuous time window and whose node queuing status has not shown continuous growth are retained, thus generating an intermediate node set. Based on the set of intermediate nodes, path combination processing is performed on each path segment to be replaced in the set of path segments to be replaced. According to the connection relationship between each intermediate edge node in the set of intermediate nodes and the game terminal and the edge node, at least one candidate path is generated from the game terminal to the edge node via one or more intermediate edge nodes. Each candidate path is represented as a path sequence composed of multiple consecutive transmission segments to form a candidate path set. Based on the candidate path set, the continuity of the path sequence in each candidate path is checked. Specifically, the bandwidth usage and node load status of adjacent transmission segments in the path sequence are compared segment by segment. When the bandwidth usage of adjacent transmission segments changes in the same direction and the node load status does not change in the opposite direction, the corresponding path sequence is marked as a continuous path sequence. Otherwise, the corresponding path sequence is removed to generate a continuous path set. Based on the set of continuous paths, the number of transmission segments in each continuous path sequence and the transmission performance sequence corresponding to each transmission segment are compared. When the number of transmission segments in the continuous path sequence is less than the number of transmission segments in the corresponding original transmission path and the change amplitude of the transmission performance sequence corresponding to each transmission segment does not increase synchronously within the continuous time window, the corresponding continuous path sequence is determined as the reconstructed path.

10. A game network acceleration method for multi-terminal collaboration according to claim 9, characterized in that: The preset rules include: based on the transmission performance sequence of the target transmission link, comparing the change direction of latency performance, jitter performance and packet loss performance at each time step within a continuous time window, and counting the number of consecutive occurrences of the same change direction and the cumulative change amplitude. When at least two of the latency performance, jitter performance and packet loss performance show a continuous consistent change direction within the continuous time window and the corresponding change amplitude shows a continuous increasing trend during the continuous comparison, the target transmission link is marked as a replaceable state. After the target transmission link is marked as a replaceable state, continuous monitoring of the transmission performance sequence continues. When the number of consecutive occurrences of the same change direction does not stop in subsequent time windows and the corresponding change amplitude does not show a reverse change, the replaceable state is determined to have entered the preset stable stage. After link replacement is performed, the transmission performance sequence of the replacement link is continuously monitored. If the direction of change of each of the delay performance, jitter performance and packet loss performance of the replacement link within a continuous time window is consistent with the direction of change of the target transmission link before replacement and the corresponding change amplitude does not increase continuously, the replacement link is marked as a stable link and the current transmission state is maintained. Otherwise, the replacement link is marked as an abnormal link and the target transmission link is reselected based on the candidate link subset.

Images