A cloud platform digital conference operation and maintenance management system and method

Abstract

The present application relates to the technical field of conference management, and particularly relates to a cloud platform digital conference operation and maintenance management system and method, comprising the following steps: according to the geographic distribution of participating nodes and the conference process time axis, analyzing the space-time correlation, generating a resource allocation strategy identifier carrying a dynamic priority label; establishing a time division multiplexing transmission channel according to the strategy identifier, injecting a state probe carrying space-time characteristics in real time in the audio and video stream transmission process, and generating a channel health degree atlas; analyzing the channel health degree atlas through an abnormal propagation chain analysis algorithm, correcting the space-time correlation parameters and triggering the reconstruction of the strategy identifier, and forming a closed-loop control including network topology adaptation. The present application effectively reduces the bandwidth burden of fixed frequency probes, reduces bandwidth overhead without affecting monitoring accuracy, and balances real-time performance and transmission stability.

Description

Technical Field

[0001] This invention relates to the field of conference management technology, and in particular to a cloud platform digital conference operation and maintenance management system and method. Background Technology

[0002] With the rapid growth in demand for distributed office, remote collaboration, and multi-location video conferencing, cloud-based video conferencing systems have become critical information infrastructure. Traditional meeting operation and maintenance management methods mostly adopt static resource configuration and unified scheduling strategies, which are difficult to meet the actual needs of the current meeting environment, such as the wide geographical distribution of nodes, complex network status, and frequent dynamic changes in meeting load.

[0003] In existing technologies, conferencing systems primarily select links and allocate resources based on predefined bandwidth thresholds or node latency, neglecting the latency differences between participating nodes and the sensitivity of conferencing stages to computing and transmission resources. Furthermore, while some systems introduce monitoring probes for link status detection, they typically employ fixed-interval injection mechanisms, lacking adaptive frequency adjustment mechanisms that correlate with conferencing semantics (such as agenda urgency). This results in high probe bandwidth consumption and delayed anomaly responses.

[0004] In terms of path scheduling, current path priorities are mostly calculated based on static delay or congestion scoring models, lacking comprehensive consideration of factors such as historical transmission success rate, congestion evolution trend, and changes in path structure (route hop count and autonomous system changes), making it difficult to achieve effective channel reconstruction and policy closure under complex network dynamics. Summary of the Invention

[0005] This invention provides a cloud platform digital conference operation and maintenance management system and method, which combines spatiotemporal correlation features, probe-driven state perception and path topology adaptive update digital conference operation and maintenance management method to improve the intelligent scheduling capability and anomaly response efficiency of cloud conference system, and ensure the stability and real-time performance of cross-regional and large-scale conferences.

[0006] A method for the operation and maintenance management of digital meetings on a cloud platform includes the following steps:

[0007] S1. Based on the geographical distribution of participating nodes and the timeline of the meeting process, analyze the spatiotemporal correlation and generate a resource allocation strategy identifier carrying a dynamic priority marker. The strategy identifier includes:

[0008] Bandwidth allocation gradient parameters calculated based on node latency differences;

[0009] Calculate resource elasticity coefficients based on meeting stages;

[0010] Dynamic priority queues for cross-regional transmission paths;

[0011] S2 establishes a time-division multiplexing transmission channel based on the policy identifier, and injects a status probe carrying spatiotemporal characteristics in real time during the audio and video stream transmission process to generate a channel health map including the following dimensions:

[0012] Inter-node transmission delay volatility;

[0013] Priority queue blocking index;

[0014] Spatiotemporal feature drift;

[0015] S3 analyzes the channel health map through the anomaly propagation chain analysis algorithm, corrects the spatiotemporal correlation parameters, and triggers policy identifier reconstruction to form a closed-loop control that includes network topology adaptation.

[0016] Optionally, S1 specifically includes:

[0017] S11. Calculate the communication delay difference matrix between participating nodes based on their geographical coordinates and network topology. Generate bandwidth allocation gradient parameters based on the delay difference matrix, wherein nodes in high-latency regions are allocated bandwidth weights with increasing gradient values.

[0018] S12 divides the meeting process into an initialization phase, a core agenda phase, and a discussion phase. It determines the baseline value of computing resource requirements for each meeting phase through historical load data analysis and calculates the phase elasticity coefficient based on the real-time fluctuation rate of the number of participants.

[0019] S13, establish a priority queue for cross-regional transmission paths and adjust the queue order in real time;

[0020] S14, the policy identifier output is a triplet policy identifier including bandwidth allocation gradient parameters, stage elasticity coefficients, and priority queues.

[0021] Optionally, the stage elasticity coefficient is calculated as follows:

[0022] Among them, E (k) η represents the resource elasticity coefficient for stage k. k N represents the preset urgency factor for each stage. (k) This represents the current real-time number of participants. The baseline number of attendees is fitted to historical data.

[0023] Optionally, in S13, the queue order is adjusted in real time based on the following parameters:

[0024] The deviation between the current latency of the transmission path and the historical average latency;

[0025] Congestion status markers of the autonomous systems traversed by the transmission path;

[0026] The level of transmission timeliness requirement corresponding to the current stage of the meeting process.

[0027] Optionally, S2 specifically includes:

[0028] S21. Based on the bandwidth allocation gradient parameters and dynamic priority queue in the policy identifier, construct virtual channels divided by time slices in the transport layer. The channel bandwidth weight of each time slice is positively correlated with the gradient parameters of the corresponding conference node.

[0029] S22, a probe carrying a spatiotemporal characteristic state is injected during the encapsulation of the audio and video streams.

[0030] An encrypted timestamp generated based on the policy identifier version;

[0031] The geogrid encoding of the sending node;

[0032] The priority queue identifier for the current transmission path;

[0033] S23, the channel health map is generated by reverse parsing the probe data at the receiving end, and the data of each dimension of the channel health map is bound and stored with the policy identifier version number.

[0034] Optionally, the probe injection interval is dynamically adjusted according to the elasticity coefficient of the conference stage, expressed as:

[0035] E (k) ∈[1.2,3.0]; ΔT0 is the set reference probe injection interval, E (k) ∈[1.2,3.0] represents E (k) The range of values ​​for .

[0036] Optionally, the channel health map specifically includes:

[0037] ① Inter-regional transmission delay volatility: Based on the probe round-trip time difference, the sliding window standard deviation algorithm is used to calculate the delay volatility between adjacent geographical regions;

[0038] ② Priority queue blocking index: The blocking index is calculated by combining the average dwell time of data packets in each queue with the theoretical transmission time and the path congestion status markers.

[0039] ③ Spatiotemporal drift: The drift percentage is calculated by comparing the topological differences between the actual transmission path and the preset path in the policy identifier using a path similarity algorithm.

[0040] Optionally, S3 specifically includes:

[0041] S31. Based on the data from each dimension of the channel health map, correct the spatiotemporal correlation parameters according to the following rules:

[0042] When the inter-regional transmission delay fluctuation rate exceeds the first threshold, the bandwidth allocation gradient parameter is adjusted in reverse according to the direction of fluctuation:

[0043] If volatility increases positively, the bandwidth allocation gradient parameter value for the corresponding region will be increased.

[0044] If the volatility decays negatively, then reduce the bandwidth allocation gradient parameter value to a multiple of the original value;

[0045] When the priority queue blocking index remains above the second threshold for two consecutive detection cycles, the feedback correction weight of the priority queue is recalculated.

[0046] When the time-space feature drift is greater than the third threshold, the delay difference matrix of the cross-regional transmission path is reconstructed, and the corresponding items in the inter-node communication delay difference matrix are updated according to the path editing distance between the actual path and the preset path.

[0047] S32 feeds back the corrected bandwidth allocation gradient parameters, priority queue feedback correction weights, and updated delay difference matrix to the spatiotemporal correlation model to regenerate the policy identifier.

[0048] Optionally, S3 further includes adaptively adjusting the closed-loop control frequency according to the latest network topology status:

[0049] When a network node is added or removed or a routing path is changed, the channel health map generation cycle is automatically shortened to 1 / 3 of the original cycle, and the spatiotemporal correlation parameters of S1 are batch calibrated.

[0050] A cloud platform digital conference operation and maintenance management system, used to implement the above-mentioned cloud platform digital conference operation and maintenance management, includes the following modules:

[0051] The spatiotemporal correlation analysis module is used to construct spatiotemporal correlations based on the geographical distribution information of participating nodes and the progress of the meeting, and generate resource allocation strategy identifiers carrying priority tags. The strategy identifiers include: bandwidth allocation gradient parameters calculated based on the communication delay differences between nodes, resource elasticity coefficients divided according to the meeting stages, and priority queues constructed for cross-regional transmission paths.

[0052] The channel construction and probe injection module constructs a time-division multiplexed transmission channel based on the policy identifier, and injects probes carrying spatiotemporal characteristic states in real time during audio and video stream transmission. The probes are used to generate a channel health map including the following dimensions:

[0053] Inter-node transmission delay volatility;

[0054] Priority queue blocking index;

[0055] Spatiotemporal feature drift;

[0056] The anomaly propagation chain analysis module performs anomaly propagation chain analysis based on the channel health map, dynamically corrects the spatiotemporal correlation parameters, updates the bandwidth allocation gradient parameters, calculates the resource elasticity coefficient and priority queue weight, and reconstructs the resource allocation strategy identifier.

[0057] The strategy closed-loop control module adaptively adjusts the model update frequency based on changes in network topology, thereby achieving closed-loop optimization control of resource scheduling strategies.

[0058] The beneficial effects of this invention are:

[0059] This invention proposes a triplet strategy identifier that carries "bandwidth allocation gradient parameters", "elastic resource factors" and "dynamic priority queues" by combining the geographical distribution of conference nodes, network topology status and conference process timeline. Through joint spatiotemporal analysis of latency difference matrix and stage urgency factor, the resource weights of different nodes are dynamically adjusted, realizing deep coupling between resource scheduling and conference semantics. Compared with the existing static allocation method based on absolute node latency, it can improve the available bandwidth guarantee capability of high-latency nodes across regions and improve the latency imbalance problem.

[0060] This invention injects status probes carrying encrypted timestamps, geocoding, and path priority identifiers into audio and video stream transmission channels. The probe injection interval is dynamically adjusted based on the elasticity coefficient corresponding to the conference stage, and a health map including inter-regional delay volatility, blocking index, and path drift is constructed. This effectively reduces the bandwidth burden of fixed-frequency probes, reduces bandwidth overhead without affecting monitoring accuracy, and achieves a balance between real-time performance and transmission stability.

[0061] This invention analyzes the propagation characteristics of various anomalies in the channel health graph, introduces a path structure editing distance function, dynamically corrects the delay difference matrix between nodes, and simultaneously recalculates the priority queue weights by fusing the blocking index and reliability score. Real-time priority self-adjustment is achieved through the product of the weights and the original strategy score. Attached Figure Description

[0062] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only for this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0063] Figure 1This is a schematic diagram of the management method flow according to an embodiment of the present invention;

[0064] Figure 2 This is a schematic diagram of the system functional modules according to an embodiment of the present invention. Detailed Implementation

[0065] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. For some well-known technologies, those skilled in the art may also use other alternative methods to implement the invention. Moreover, the accompanying drawings are only for more specific description of the embodiments and are not intended to specifically limit the present invention.

[0066] like Figure 1 As shown, a cloud platform digital meeting operation and maintenance management method includes the following steps:

[0067] S1, based on the geographical distribution of participating nodes and the timeline of the meeting process, analyze the spatiotemporal correlation and generate resource allocation strategy identifiers carrying dynamic priority markers. The strategy identifiers include:

[0068] Bandwidth allocation gradient parameters calculated based on node latency differences;

[0069] Calculate resource elasticity coefficients based on meeting stages;

[0070] Dynamic priority queues for cross-regional transmission paths;

[0071] S2 establishes a time-division multiplexing transmission channel based on the policy identifier, and injects a status probe carrying spatiotemporal characteristics in real time during the audio and video stream transmission process to generate a channel health map including the following dimensions:

[0072] Inter-node transmission delay volatility;

[0073] Priority queue blocking index;

[0074] Spatiotemporal feature drift;

[0075] S3 analyzes the channel health map through the anomaly propagation chain analysis algorithm, corrects the spatiotemporal correlation parameters, and triggers policy identifier reconstruction to form a closed-loop control that includes network topology adaptation.

[0076] The analysis of spatiotemporal correlations includes:

[0077] Inter-node delay difference matrix → spatial correlation;

[0078] Meeting process division and flexibility coefficient → time correlation;

[0079] Cross-regional path priority sorting → spatiotemporal cross-association.

[0080] S1 specifically includes:

[0081] S11, Calculation of Inter-Node Delay Difference Matrix and Generation of Bandwidth Allocation Gradient Parameters: Based on the geographical coordinates of each participating node and the network topology information, calculate the communication delay difference between any two participating nodes i and j, and construct the delay difference matrix: ΔD ij =|D i -D j |;wherein, D i Let ΔD be the average network latency from node i to the scheduling center or main venue node. ij This represents the relative delay difference between nodes i and j.

[0082] Calculate the bandwidth allocation gradient parameter G based on the delay difference. i :

[0083]

[0084] Where G0 is the initial bandwidth allocation weight, α is the bandwidth adjustment gain coefficient, and N h G represents the total number of participating nodes. i This represents the bandwidth gradient weight assigned to node i.

[0085] S12, Modeling and computational resource elasticity coefficient calculation during the conference phase:

[0086] The meeting process was divided into three phases:

[0087] Initialization phase;

[0088] Core Agenda Phase;

[0089] Discussion phase.

[0090] Each stage is assigned a corresponding urgency factor η. k Where k = 1, 2, 3 correspond to the three stages, and the baseline number of computing resources for each stage is determined based on historical load data. Based on the current real-time number of participants N (k) The elasticity coefficient of the defined stage is:

[0091] Among them, E (k) η represents the resource elasticity coefficient for stage k. k The urgency factor for each stage is represented by the following settings: initialization stage η1 = 0.6, core agenda stage η2 = 1.2, discussion stage η3 = 0.8. (k) This represents the current real-time number of participants. The baseline number of attendees is used to fit the historical data, and the fitting method is as follows:

[0092] In each conference phase k, a baseline number of attendees is calculated by fitting a historical conference dataset (e.g., the most recent 100 similar conferences). Where, μ k σ is the average number of participants in historical conferences during stage k. k λ is the standard deviation, representing volatility, and λb is an empirical adjustment factor (recommended value λb = 0.5~1.0, balance tolerance).

[0093] S13, Construction of dynamic priority queues for cross-regional transmission paths:

[0094] A dynamic priority sorting queue is established for the conference transmission path, with a priority scoring function P. r The definition is as follows:

[0095] P r =β1·δ d +β2·C r +β3·U k ; where δ d Indicates the path delay deviation. D r Delay for the current path, For its historical average, C r This is a congestion marker for the autonomous systems (Autonomous Regions) traversed by the path, where 0 indicates unobstructed access and 1 indicates congestion. k The priority level (U1=1, mainly for access verification and channel testing, low requirement); the core agenda stage (U2=3, for presentations / screen sharing / high-definition audio and video transmission, extremely high requirement); and the discussion stage (U3=2, mainly for real-time interaction, latency is tolerable but stable transmission is required) represents the timeliness requirement of the transmission path in stage k of the meeting (U1=1, mainly for access verification and channel testing, low requirement); the core agenda stage (U2=3, mainly for presentations / screen sharing / high-definition audio and video transmission, extremely high requirement); and the discussion stage (U3=2, mainly for real-time interaction, latency is tolerable but stable transmission is required). β1, β2, and β3 are multi-dimensional weighting coefficients used to balance latency, congestion, and business needs. Paths with higher priority scores rank higher in the queue, and status probes and resource scheduling strategies are preferentially injected and deployed along these paths.

[0096] S14, Combining the above steps, construct the policy identifier for the output triple:

[0097]

[0098] That is, the policy identifier S is a triple containing:

[0099] 1. Set of gradient parameters for bandwidth allocation at each node {G} i};

[0100] 2. Set of flexible computing resource coefficients for each meeting phase {E} (k)};

[0101] 3. Dynamic priority score set for each transmission path {P} r}

[0102] S2 specifically includes:

[0103] S21, Construct time-division multiplexed virtual channels (based on gradient parameters): Allocate gradient parameters G according to the bandwidth in the policy identifier. i and priority queue P r A virtual channel structure based on time-slice partitioning is constructed at the transport layer. Each time slice is allocated a bandwidth weight. Calculated using the following formula: in, This represents the bandwidth weight of node i in time slice t, where B0 is the standard base bandwidth (10Mbps by default for each time slice), and G... i Assign gradient parameters to the bandwidth of the output of S1.

[0104] S22, Injecting a state probe carrying spatiotemporal characteristics: During the audio / video stream encapsulation stage, inject a state probe containing the following fields at dynamic frequencies:

[0105] in, This indicates that the version number is based on the current timestamp t and the policy identifier. Encrypted timestamps, This represents the geohash encoding (QID) of the sending end. r This indicates the priority queue identifier corresponding to the current transmission path, and the probe injection interval ΔT. k Based on the elasticity coefficient E of meeting stage k (k) Dynamic settings:

[0106] ΔT0 is the set reference probe injection interval (value is 2s), E (k) ∈[1.2,3.0] represents E (k) The range of values ​​is determined by an upper / lower bound truncation strategy, where E in S1... (k) If the calculated result is below 1.2, use 1.2; if it is above 3.0, use 3.0, to avoid injection interval ΔT. k Too small an injection interval will overload the network probe and consume bandwidth; too large an injection interval will reduce monitoring accuracy and delay abnormal response.

[0107] S23, Generate Channel Health Map: After receiving the probe data, the receiving end reverse-analyzes the data to generate a multi-dimensional channel health index.

[0108] S231, Inter-node transmission delay volatility: A sliding window standard deviation model is used to analyze the volatility of round-trip time between adjacent nodes.

[0109]

[0110] in, Representative meeting node pair, This represents the time of the probe's t-th round trip between nodes i and j. This represents the average time for n round trips within the sliding window. This represents the inter-node latency volatility;

[0111] S232, Priority Queue Congestion Index: Calculated by the average dwell time T of data packets in each priority queue. stay With ideal transmission time T ideal The ratio, and introduce the number of autonomous system congestion labels C. r Calculate the congestion index:

[0112]

[0113] in, The average time data stays in priority queue r. C represents the theoretical shortest transmission time for path r. r This represents the number of autonomous systems marked as congested along path r (based on BGP message statistics). This is the congestion index.

[0114] S233, Spatiotemporal Feature Drift: The difference between the actual transmission path and the preset strategy path is evaluated using path topology similarity, and the topology edit distance D is defined. edit Then, calculate the percentage of drift:

[0115]

[0116] in, For the actual transmission path, For the preset path in the policy identifier, D edit (·) represents the path edit distance, taking into account route hop count and autonomous system overlap. ref This is the reference path length (average path hops). This represents the path drift rate.

[0117] Path edit distance is represented as:

[0118] in, These are the weighting coefficients. The weighting coefficient for the difference in the number of jumps. This is the weighting coefficient for the overlap difference among autonomous regions. Autonomous System (AS) overlap (the proportion of AS numbers that overlap between two paths). |ΔH (i,j) | represents the difference in hop count between the two paths. This represents the number of hops in the actual path. This represents the number of hops in the strategy path.

[0119] S24, Graph storage and policy binding:

[0120] Generate channel health map Among them, the data of each dimension {σ delay B idx ,δ spatial} and the corresponding policy identifier version number Bind:

[0121] S3 specifically includes:

[0122] S31, Parameter Correction Rule Definition: Based on the dimensional indicators in the channel health graph, the core parameters of the spatiotemporal correlation model are dynamically corrected.

[0123] S311, Delay volatility-guided bandwidth gradient correction: If the transmission delay volatility of a node pair (i,j) is... If the first threshold θ1 is exceeded, the bandwidth allocation gradient parameter G of the corresponding node i is corrected. i .

[0124] If the fluctuation is positive growth (i.e., higher than the historical benchmark):

[0125]

[0126] If the fluctuation is negatively decaying:

[0127] in, The standard deviation of the current node-pair transmission delay is given by E, where σ0 is the set baseline volatility (15%). (k) G represents the elasticity coefficient for the current meeting phase (from S1). i Assign gradient parameters to the original bandwidth.

[0128] S312, Priority queue weight recalculation based on queue blocking: When the blocking index of a certain path r... If the detection cycle exceeds the second threshold θ2 for two consecutive rounds, then the feedback correction weight W in the dynamic priority queue is recalculated. r :

[0129]

[0130] Among them, W r This represents the original priority queue weight of path r (set to 1). Indicates the blocking level of the current path. N represents the path reliability score. succ N fail This represents the number of successful and failed transmissions along the path in the past hour. After calculation, the original priority value P is updated.r The new priority value P is obtained. new P new =W r ·P r ;

[0131] S313, Delay difference matrix correction based on drift amount: If the transmission path drift of a node pair (i,j) is... If the third threshold θ3 is exceeded, then the corresponding term ΔD in the delay difference matrix... ij Perform a topology update:

[0132]

[0133] in, D represents the path drift rate. edit L is the edit distance between the actual path and the strategy path. ref Let f(·) be the reference path length (average hop count), and let f(·) be the mapping function that maps topology changes to delay gain. The path edit distance, also known as "path structure difference," is represented by f(·), which is an estimated mapping of the impact of path structure difference on transmission performance. It converts the topological differences of the transmission path—that is, the structural changes (route hop count, autonomous system coverage) between the actual path and the preset path in the policy identifier—into corrective values ​​in the communication delay difference matrix to reflect the potential impact of network path changes on transmission performance. Specifically:

[0134] If the actual path changes, for example, by hopping two more nodes than the original path or traversing different autonomous systems, it will cause a certain increase in latency.

[0135] The path difference is quantified by "edit distance" and then transformed into a numerical correction term by the mapping function f(·).

[0136] The final delay difference matrix term ΔD used to update node pair (i,j) ij This is used for bandwidth allocation in subsequent policy generation.

[0137] The specific expression for the mapping function f(·) is:

[0138] Among them, D base The original node-pair delay difference baseline value (which can be the previous calculated value or a static estimate), |ΔH| is the hop difference between the preset path and the actual path, and ρ is the hop difference between the preset path and the actual path. AS ∈[0,1], the degree of overlap of autonomous systems (AS) in two paths, the closer to 1, the more similar the path structures.

[0139] In simple terms, the mapping function f(·) essentially maps the original node-to-node communication delay difference base value. Distance from path edit The sums are used to form a new corrected delay difference value, which is then used to update ΔD. ij The role of the mapping function is: It is a performance increment compensation function caused by structural differences.

[0140] The first threshold θ1 is set to 0.15, which means that when the standard deviation of the round-trip delay between regions / nodes exceeds this proportion, the link fluctuation is considered abnormal.

[0141] The second threshold θ2 is set to 0.6, which means that the congestion index is greater than 0.6, that is, the average queue waiting time is more than 60% of the theoretical time, which is a risk of congestion.

[0142] The third threshold θ3 is set to 0.2. If the difference between the actual path and the preset path exceeds 20%, it is determined that a major topology shift has occurred.

[0143] S32, regenerate the policy identifier triplet, feed the revised parameter set back to the original model, and form a new policy identifier: Each parameter directly corresponds to one of the three core components of the model built in S1.

[0144] S33, Adaptively adjust the closed-loop control frequency based on the latest network topology status: When the network topology detection module detects that any of the following conditions are met:

[0145] Condition 1: Addition or deletion of network nodes;

[0146] Condition 2: The routing path has changed (achieved through BGP-LS monitoring);

[0147] The control frequency will be automatically adjusted to compress the health profile generation cycle T. new : T0 is the basic setting cycle;

[0148] And trigger parameter reconstruction:

[0149] If the number of new nodes is N new >0.1×N total Then, a full update of the delay difference matrix ΔD is performed.

[0150] Synchronously recalculate bandwidth gradient and path priority in batches, and enter high-frequency scheduling state.

[0151] like Figure 2 As shown, a cloud platform digital conference operation and maintenance management system, used to implement the above-mentioned operation and maintenance management method, includes the following modules:

[0152] The spatiotemporal correlation analysis module is used to construct spatiotemporal correlations based on the geographical distribution information of participating nodes and the progress of the meeting, and generate resource allocation strategy identifiers carrying priority tags. The strategy identifiers include: bandwidth allocation gradient parameters calculated based on the differences in communication delays between nodes, resource elasticity coefficients divided according to the meeting stages, and priority queues constructed for cross-regional transmission paths.

[0153] The channel construction and probe injection module constructs time-division multiplexed transmission channels based on policy identifiers and injects probes carrying spatiotemporal characteristic states in real time during audio and video stream transmission. The probes are used to generate a channel health map including the following dimensions:

[0154] Inter-node transmission delay volatility;

[0155] Priority queue blocking index;

[0156] Spatiotemporal feature drift;

[0157] The anomaly propagation chain analysis module performs anomaly propagation chain analysis based on the channel health map, dynamically corrects spatiotemporal correlation parameters, updates bandwidth allocation gradient parameters, calculates resource elasticity coefficients and priority queue weights, and reconstructs resource allocation strategy identifiers.

[0158] The strategy closed-loop control module adaptively adjusts the model update frequency based on changes in network topology, thereby achieving closed-loop optimization control of resource scheduling strategies.

[0159] This invention encompasses any substitutions, modifications, equivalent methods, and solutions made within the spirit and scope of this invention. To provide the public with a thorough understanding of this invention, specific details are described in detail in the following preferred embodiments; however, those skilled in the art will fully understand the invention even without these details. Furthermore, to avoid unnecessary misunderstanding of the essence of this invention, well-known methods, processes, procedures, components, and circuits are not described in detail.

[0160] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A cloud platform digitized conference operation and maintenance management method, characterized in that, Comprise the following steps: S1, according to the geographical distribution of the nodes participating in the meeting and the time axis of the meeting process, analyze the space-time correlation, generate resource allocation strategy identifier carrying dynamic priority label, the strategy identifier includes: Bandwidth allocation gradient parameter calculated based on node delay difference; The calculation resource elasticity coefficient is divided according to the meeting stage; Dynamic priority queue of cross-regional transmission path; Among them, the S1 specifically includes: S11, according to the geographical position coordinates of the nodes participating in the meeting and the network topology relationship, the communication delay difference matrix between the nodes participating in the meeting is calculated, and the bandwidth allocation gradient parameter is generated based on the delay difference matrix, wherein the bandwidth weight of the node in the high delay area is increased in the gradient value; S12, the meeting process is divided into initialization stage, core agenda stage, discussion stage, the calculation resource demand baseline value of each meeting stage is determined through historical load data analysis, and the stage elasticity coefficient is calculated based on the real-time number of participants fluctuation rate; S13, priority queue of cross-regional transmission path is established, and queue order is adjusted in real time; S14, the strategy identifier output is a three tuple strategy identifier including bandwidth allocation gradient parameter, stage elasticity coefficient and priority queue; S2, according to the strategy identifier, a time division multiplexing transmission channel is established, a state probe carrying space-time characteristics is injected in real time in the process of audio and video stream transmission, and a channel health degree atlas including the following dimensions is generated: Inter-node transmission delay fluctuation rate; Priority queue blocking index; Space-time characteristic drift amount; Among them, the S2 specifically includes: S21, according to the bandwidth allocation gradient parameter and dynamic priority queue in the strategy identifier, a time slice divided virtual channel is constructed in the transmission layer, and the channel bandwidth weight of each time slice is positively correlated with the gradient parameter of the corresponding meeting node; S22, inject the probe of space-time characteristic state when encapsulating the audio and video stream, the probe carries: Encrypted timestamp generated based on the strategy identifier version; Geographical grid encoding of the sending end node; Priority queue identification of the current transmission path; S23, the channel health degree atlas is calculated and generated by receiving end reverse analysis of probe data, and the dimension data of the channel health degree atlas is bound and stored with the version number of the strategy identifier; S3, the channel health degree atlas is analyzed by abnormal propagation chain analysis algorithm, the space-time correlation parameter is corrected, and the strategy identifier is reconstructed, forming a closed loop control including network topology adaptation; Among them, the S3 specifically includes: S31, according to the dimension data of the channel health degree atlas, correct the space-time correlation parameter according to the following rules: When the inter-regional transmission delay fluctuation rate exceeds the first threshold value, the bandwidth allocation gradient parameter is adjusted in the reverse direction according to the fluctuation direction: If the fluctuation rate is positive, the bandwidth allocation gradient parameter value of the corresponding region is increased; If the fluctuation rate is negative, the bandwidth allocation gradient parameter value is reduced to a multiple of the original value; When the priority queue blocking index is higher than the second threshold value for two rounds of detection period, the feedback correction weight of the priority queue is recalculated: When the space-time feature drift is greater than a third threshold, reconstructing a delay difference matrix of the cross-region transmission path, and updating a corresponding item in the inter-node communication delay difference matrix according to a path edit distance between the actual path and the preset path; S32, feeding back the corrected bandwidth allocation gradient parameter, priority queue feedback correction weight and updated delay difference matrix to the space-time correlation model to regenerate the policy identifier. 2.The cloud platform digital conference operation and maintenance management method of claim 1, wherein, The stage elasticity coefficient is calculated as: ; wherein, represents the stage elasticity coefficient of elastic computing resources, represents the preset urgency factor of each stage, is the real-time number of participants in the current stage, is the benchmark number of participants fitted from historical data. 3.The cloud platform digital conference operation and maintenance management method of claim 1, wherein, The queue order is adjusted in real time in S13 based on the following parameters: Deviation of the current delay of the transmission path from the mean of the historical delay; Congestion state label of the network autonomous domain through which the transmission path passes; Grade of the transmission timeliness requirement corresponding to the stage of the conference process.

4. The cloud platform digitized conference operation and maintenance management method of claim 1, wherein, The probe injection interval is dynamically adjusted according to the conference stage elasticity coefficient and is expressed as: ; injecting intervals for the set reference probes, indicates the value range.

5. The cloud platform digitized conference operation and maintenance management method of claim 1, wherein, The channel health degree atlas specifically includes: ① Inter-regional transmission delay fluctuation rate: based on the round-trip time difference of the probe, the sliding window standard deviation algorithm is used to calculate the delay fluctuation rate between adjacent geographical regions; ② Priority queue blocking index: the ratio of the average residence time of data packets in each queue to the theoretical transmission time is calculated, and the blocking index is calculated in combination with the path congestion state label; ③ Space-time feature drift: the topology difference between the actual transmission path and the preset path in the policy identifier is compared, and the path similarity algorithm is used to calculate the drift percentage.

6. The cloud platform digitized conference operation and maintenance management method of claim 1, wherein, S3 further includes adaptively adjusting the closed-loop control frequency according to the latest network topology state: When network node addition or deletion or routing path change is detected, the channel health degree atlas generation period is automatically shortened to 1 / 3 of the original period, and the batch calibration of the space-time correlation parameters in S1 is triggered.

7. A cloud platform digital conference operation and maintenance management system for implementing the cloud platform digital conference operation and maintenance management method according to any one of claims 1-6, characterized in that, The following modules are included: A space-time correlation analysis module for constructing a space-time correlation relationship according to the geographical distribution information of the participating nodes and the conference process, generating a resource allocation strategy identifier carrying a priority label, the strategy identifier including a bandwidth allocation gradient parameter calculated based on the inter-node communication delay difference, a computing resource elasticity coefficient divided according to the conference stage, and a priority queue constructed for a cross-region transmission path; A channel construction and probe injection module for constructing a time-division multiplexing transmission channel according to the strategy identifier, and injecting a probe carrying a space-time feature state in real time during the audio and video stream transmission process, the probe being used to generate a channel health degree atlas including the following dimensions: Inter-node transmission delay fluctuation rate; Priority queue blocking index; Space-time feature drift; An abnormal propagation chain analysis module for analyzing abnormal propagation chains based on the channel health degree atlas, dynamically correcting the space-time correlation parameters, updating the bandwidth allocation gradient parameter, the computing resource elasticity coefficient and the priority queue weight, and reconstructing the resource allocation strategy identifier; A strategy closed-loop control module for adaptively adjusting the model update frequency based on the change of the network topology state, to realize closed-loop optimization control of the resource scheduling strategy.

Images