Cross-cloud network settlement method based on cost and performance dynamic optimization and related equipment

Abstract

The embodiment of the present application relates to the field of computer network and cloud computing technology, and discloses a cross-cloud network settlement method based on cost and performance dynamic optimization and related equipment, the method comprises the following steps: collecting performance parameters and charging parameters, converting the charging parameters into equivalent traffic costs; identifying the business type, obtaining the weight strategy; determining the comprehensive score of each traffic path based on the performance parameters, the equivalent traffic costs and the weight strategy; determining whether to switch the path according to the comprehensive score, the continuous use time and the path switching frequency of the data stream to be transmitted; if yes, switching to the candidate traffic path with the highest comprehensive score; determining the quality coefficient based on the performance parameters, and generating a traffic settlement bill based on the equivalent traffic costs, the path switching time and the quality coefficient. Through the above method, the embodiment of the present application can dynamically select the network traffic path with the optimal cost performance, adapt to the dynamic changes of the network state, and take into account the cost fluctuation and the performance state.

Description

Technical Field

[0001] This invention relates to the fields of computer network and cloud computing technology, specifically to a cross-cloud network settlement method and related equipment based on dynamic optimization of cost and performance. Background Technology

[0002] As enterprises deepen their digital transformation, their IT architecture commonly adopts a hybrid multi-cloud model. Enterprises typically lease resources from multiple public cloud service providers (e.g., Alibaba Cloud, Tencent Cloud, AWS, and Azure) and multiple network operators (e.g., China Telecom, China Unicom, PCCW, and NTT). While this model offers flexibility, its shortcomings are becoming increasingly apparent: On the one hand, traditional network path selection and billing systems are independent. Path selection is usually based solely on network performance metrics (e.g., latency, packet loss), while billing relies on static bills provided by operators (e.g., 95th percentile billing, peak bandwidth billing), meaning scheduling and billing are disconnected. On the other hand, traditional routing strategies are often static or semi-static. For example, high-cost leased lines are fixed for critical services, while low-cost internet links are configured for non-critical services. This disconnect and routing strategy cannot adapt to dynamic changes in network conditions or balance cost fluctuations with performance status. In particular, low-cost paths may degrade in performance during peak hours, or high-priced leased line resources may become idle during off-peak hours. Furthermore, the billing model (e.g., traffic-based billing) is decoupled from the actual service quality provided by the path. Customers may pay the same price for a high-priced link experiencing severe congestion during transmission as they do for a stable, high-quality link, resulting in an unfair billing model. Once a bill is generated, customers find it difficult to verify the correspondence between cost and service quality, leading to difficulties in dispute resolution and auditing. Summary of the Invention

[0003] In view of the above problems, embodiments of the present invention provide a cross-cloud network settlement method and related equipment based on dynamic optimization of cost and performance, which are used to solve the problems existing in the prior art.

[0004] According to one aspect of the present invention, a cross-cloud network settlement method based on dynamic cost and performance optimization is provided, the method comprising: Within a preset time window, collect the performance parameters and billing parameters corresponding to the current traffic path and multiple candidate traffic paths respectively, and convert the billing parameters into equivalent traffic costs based on predetermined billing rules. Identify the service type of the data stream to be transmitted in the current traffic path, and obtain the weight strategy corresponding to the service type from the preset strategy library; The performance parameters and the equivalent traffic cost are normalized, and the comprehensive score of each traffic path is determined based on the normalized performance parameters, equivalent traffic cost and the corresponding weighting strategy. Based on the comprehensive score of the current traffic path, the comprehensive score of multiple candidate traffic paths, the continuous usage time of the current traffic path, and the number of path switching times of the data stream to be transmitted within the time window, it is determined whether to switch the path of the data stream to be transmitted. If so, a path switching instruction is sent to the network execution module, which then switches the current traffic path of the data stream to be transmitted to the candidate traffic path with the highest comprehensive score based on the path switching instruction. During the transmission period of the data stream to be transmitted, a service quality compliance rate is determined based on the performance parameters, and a continuously changing quality coefficient is generated based on the compliance rate. The higher the compliance rate, the larger the quality coefficient. A traffic settlement bill is generated based on the equivalent traffic cost of the current traffic path, the equivalent traffic cost of the candidate traffic path with the highest comprehensive score, the path switching time, and the quality coefficient.

[0005] In one optional approach, determining whether to switch the path of the data stream to be transmitted based on the comprehensive score of the current traffic path, the comprehensive score of multiple candidate traffic paths, the continuous usage time of the current traffic path, and the number of path switching times of the data stream to be transmitted within the time window includes: The system analyzes whether the comprehensive score of a candidate traffic path is significantly higher than that of the current traffic path, whether the continuous usage time of the current traffic path is greater than or equal to the predetermined minimum dwell time, and whether the number of path switching times of the data stream to be transmitted within the time window is less than the predetermined number of switching times. If the analysis results are all yes, then the path to switch the data stream to be transmitted is determined.

[0006] In one alternative, the strategy library includes a dynamic optimization engine that dynamically optimizes the weights in the weight strategy for each of the business types according to a predetermined optimization algorithm, including a value-based iteration optimization algorithm, a strategy gradient algorithm, and a multi-armed gambling machine algorithm.

[0007] In one optional approach, generating a traffic bill based on the equivalent traffic cost of the current traffic path, the equivalent traffic cost of the candidate traffic path with the highest comprehensive score, the path switching time, and the quality coefficient includes: Based on the equivalent traffic cost of the current traffic path, the equivalent traffic cost of the candidate traffic path with the highest comprehensive score, the path switching time, and the quality coefficient, a sharding bill with multiple predetermined dimensions is generated. The multiple predetermined dimensions include region, cloud vendor, operator, and tenant. The traffic settlement bill is generated based on the segmented bill.

[0008] In an alternative approach, the method further includes: Write the traffic settlement bill, the performance parameters, the billing parameters, the equivalent traffic cost, and the weighting strategy into a blockchain or an immutable storage system; If a review request is received within the preset dispute period, the traffic settlement bill, performance parameters, billing parameters, equivalent traffic cost, and weighting strategy are obtained from the blockchain or the storage system. The review is performed based on the performance parameters, billing parameters, equivalent traffic cost, and weighting strategy, and a traffic settlement bill to be reviewed is generated. The obtained traffic settlement bill is then compared with the traffic settlement bill to be reviewed.

[0009] In one optional approach, after collecting the performance parameters and billing parameters corresponding to the current traffic path and multiple candidate traffic paths within a preset time window, the method further includes: The performance parameters are input into a pre-trained prediction model for prediction, and the future performance trends of the current traffic path and multiple candidate traffic paths output by the prediction model are obtained. If the future performance trend of the current traffic path is deterioration, then the overall score of the current traffic path will be reduced according to the deterioration state. If the future performance trend of any candidate traffic path is deteriorating, then after determining the path to switch the data stream to be transmitted, the candidate traffic path with a good future performance trend is selected as the target path for switching.

[0010] In one alternative approach, the service type includes at least real-time interaction, live streaming, and large file transfer, and the service type for identifying the data stream to be transmitted in the current traffic path includes: The service type of the data stream to be transmitted can be identified based on the packet size, packet interval, traffic rate, and protocol type of the current traffic path; or based on a pre-trained multi-feature fusion classification model.

[0011] According to another aspect of the present invention, a cross-cloud network settlement system based on dynamic cost and performance optimization is provided, the system comprising: The data acquisition and cost calculation module is used to collect the performance parameters and billing parameters corresponding to the current traffic path and multiple candidate traffic paths within a preset time window, and to convert the billing parameters into equivalent traffic costs based on predetermined billing rules. The identification module is used to identify the service type of the data stream to be transmitted in the current traffic path. A strategy library is used to obtain the weight strategy corresponding to the business type from a preset strategy library; A dynamic scheduling engine is used to normalize the performance parameters and the equivalent traffic cost, and determine the comprehensive score of each traffic path based on the normalized performance parameters, equivalent traffic cost, and corresponding weighting strategy. Based on the comprehensive score of the current traffic path, the comprehensive scores of multiple candidate traffic paths, the continuous usage time of the current traffic path, and the number of path switching times of the data stream to be transmitted within the time window, it determines whether to switch the path of the data stream to be transmitted; if so, it issues a path switching command to the network execution module. The network execution module is used to switch the current traffic path of the data stream to be transmitted to the candidate traffic path with the highest comprehensive score based on the path switching instruction. The billing and settlement module is used to determine the service quality compliance rate based on the performance parameters during the transmission period of the data stream to be transmitted, generate a continuously changing quality coefficient based on the compliance rate, and the higher the compliance rate, the larger the quality coefficient; and generate a traffic settlement bill based on the equivalent traffic cost of the current traffic path, the equivalent traffic cost of the candidate traffic path with the highest comprehensive score, the path switching time, and the quality coefficient.

[0012] According to another aspect of the present invention, a computer device is provided, comprising: a processor, a memory, a communication interface, and a communication bus, wherein the processor, the memory, and the communication interface communicate with each other via the communication bus; the memory is used to store at least one executable instruction, wherein the executable instruction causes the processor to perform the method described above.

[0013] According to another aspect of the present invention, a computer-readable storage medium is provided, the storage medium storing at least one executable instruction, which, when executed on a computer device, causes the computer device to perform the method described above.

[0014] In a multi-cloud and multi-carrier hybrid network environment, this invention collects performance and billing parameters in real time, converts the billing parameters into equivalent traffic costs, and comprehensively scores all traffic paths based on the performance parameters and equivalent traffic costs. It then determines whether to switch paths based on the comprehensive score, selecting the path with the highest comprehensive score for switching. Subsequently, it generates a continuously changing quality coefficient based on the service quality compliance rate during transmission to dynamically adjust billing costs. This invention deeply integrates scheduling and settlement, dynamically selecting the most cost-effective network traffic path based on real-time collected performance and equivalent traffic costs. It adapts to dynamic changes in network conditions while considering cost fluctuations and performance status, significantly reducing network costs while ensuring service quality. Furthermore, the settlement scheme is fair and transparent.

[0015] The above description is merely an overview of the technical solutions of the embodiments of the present invention. In order to better understand the technical means of the embodiments of the present invention and to implement them in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the embodiments of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0016] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 A flowchart illustrating the cross-cloud network settlement method based on dynamic cost and performance optimization provided in an embodiment of the present invention is shown. Figure 2 A schematic diagram is shown illustrating the search for the optimal flow path using the Pareto optimal search method; Figure 3 A schematic diagram of the structure of a cross-cloud network settlement system based on dynamic cost and performance optimization provided in an embodiment of the present invention is shown. Figure 4 A schematic diagram of the structure of a computer device provided in an embodiment of the present invention is shown. Detailed Implementation

[0017] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0018] Figure 1 A flowchart of a cross-cloud network settlement method based on dynamic cost and performance optimization according to an embodiment of the present invention is shown, as follows: Figure 1 As shown, the method includes the following steps: Step S10: Collect the performance parameters and billing parameters corresponding to the current traffic path and multiple candidate traffic paths within a preset time window, and convert the billing parameters into equivalent traffic costs based on the predetermined billing rules.

[0019] In this embodiment, a time window can be used to limit the number of path switching operations. Within this time window, the number of path switching operations shall not exceed a predetermined number.

[0020] In a multi-cloud and multi-carrier hybrid network environment, performance and billing parameters of multiple cross-cloud and cross-carrier network traffic paths are collected in real time, including the current traffic path and multiple candidate traffic paths. Specifically, real-time performance data collection and aggregation are performed through a distributed stream processing framework (e.g., Apache Flink) to ensure low latency and high throughput.

[0021] By integrating cloud vendor APIs and carrier APIs, contract rules can be obtained from cloud vendor and carrier systems through the integrated APIs, the contract rules can be automatically parsed, billing parameters can be collected from them, and the conversion logic can be dynamically executed through a rule engine (e.g., Drools) to support complex billing scenarios.

[0022] In this embodiment, performance parameters are collected at fixed intervals using cloud server nodes and edge gateway probes deployed globally. This allows for the acquisition of performance parameters for multiple cross-cloud and cross-carrier traffic paths, including latency, jitter, packet loss rate, and bandwidth. Latency: Calculated by sending probe messages and measuring the round-trip time; Jitter: Calculated by the difference in time delay between multiple samplings; Packet loss rate: Calculated by the difference between the number of packets sent and received; Available bandwidth: The remaining bandwidth is estimated using a bandwidth probing algorithm.

[0023] The billing parameters are the corresponding traffic costs under various billing rules, including 95th percentile billing, tiered pricing, monthly overage fees, and cross-domain surcharges. Among them, 95th percentile billing calculates the marginal impact of new traffic on the percentile value by statistically analyzing bandwidth usage through a sliding window; tiered pricing determines the applicable unit price for new traffic based on the traffic range; monthly overage fees convert the excess amount to a unit price when it exceeds the fixed limit; and cross-domain surcharges add extra costs when transmitting data across borders or across carriers.

[0024] Ultimately, all billing parameters are uniformly converted into equivalent traffic cost, which is the cost of transmitting one unit of traffic within the current time window, thereby achieving comparability between different billing models.

[0025] Furthermore, after the billing parameters are converted into equivalent traffic costs, the historical equivalent traffic costs are input into the pre-trained time series prediction model (such as ARIMA or LSTM) to predict the cost trend of future time windows. This provides a prerequisite for dynamically adjusting the current cost calculation and avoids lag caused by not predicting future cost trends.

[0026] Step S20: Identify the service type of the data stream to be transmitted in the current traffic path, and obtain the weight strategy corresponding to the service type from the preset strategy library.

[0027] In this embodiment, the service type of the data stream to be transmitted on the current traffic path is first identified. For example, the service type of the data stream to be transmitted is identified by characteristics such as packet size, packet interval, traffic rate, and protocol type of the current traffic path. Service types include real-time interaction, live streaming, and large file transfer, etc. For example, if the packets are small and uniform, the packet interval is fixed and has low jitter, and the traffic is consistently stable and low, then the service type is identified as real-time interaction; if there are usually large and relatively consistent data packets (video frames), uniform packet intervals, and consistently stable and high traffic, then the service type is identified as live streaming; if the data packets are very large, the packet intervals are irregular, and the traffic rate is high, then the service type is identified as large file transfer.

[0028] Furthermore, this embodiment uses deep packet inspection (DPI) technology to extract application layer features (e.g., HTTP headers, payload content) to identify service types. It can also be further combined with behavioral analysis (e.g., traffic bursts, duration) to improve the accuracy of service type identification.

[0029] Furthermore, to improve the accuracy of service type identification, a multi-feature fusion classification model can be used to identify the service type of the data stream to be transmitted in the current traffic path. This multi-feature fusion classification model can be a deep learning model, such as a convolutional neural network (CNN) or a recurrent neural network (RNN). The model input includes multiple features such as packet size, packet interval, traffic rate, and protocol type. The model outputs a probability distribution of the service type through multi-feature fusion analysis. The model is trained based on historical traffic data and is periodically updated online to adapt to new service types.

[0030] In this embodiment, the policy library stores weight policies corresponding to various service types, with different weight policies for different service types. Each service type's weight policy includes a weight combination of latency weight, jitter weight, packet loss rate weight, bandwidth weight, and cost weight. The weight can be obtained by querying the weight policy corresponding to the service type. For real-time interaction, latency is the most important influencing factor, therefore, latency has the largest weight; for large file transfer, cost is the most important influencing factor, therefore, cost has the largest weight, as shown in Table 1 below. Real-time interaction 0.35 0.25 0.20 0.10 0.10 Live streaming 0.25 0.20 0.25 0.20 0.10 Large file transfer 0.10 0.05 0.10 0.25 0.50 Table 1 Furthermore, the policy library includes an adaptive dynamic optimization engine based on reinforcement learning. The weights in the weight policy for each business type can be dynamically optimized through Q-learning or policy gradient methods. For example, when network conditions change, dynamically optimized weights can more accurately evaluate the optimal traffic path. The policy library supports multi-tenant isolation, with each tenant having its own weight policy instance, and sharing weight policy updates under privacy protection through federated learning. In one embodiment, a multi-armed bandit algorithm can also be used to select the weight combination corresponding to the business type in real time to dynamically adjust and ultimately find the weight configuration with the best long-term performance. In this algorithm, arms are first defined, each arm representing a weight combination for a certain business type. Then, a reward is defined, i.e., a quantifiable metric is determined to measure the performance after each use of a specific weight combination (pulling an "arm"), and an appropriate algorithm is selected based on the actual situation to achieve a balance between exploration and exploitation. The algorithms include the ε-Greedy algorithm, the upper confidence interval algorithm, and the Thompson sampling algorithm. The weight combination (arm) to be tried is selected according to the policy of the selected algorithm, the system or model is run with the new weights, and the reward is observed. The value estimate of the weight combination is updated based on the obtained reward.

[0031] Step S30: Normalize the performance parameters and the equivalent traffic cost, and determine the comprehensive score of each traffic path based on the normalized performance parameters, equivalent traffic cost and the corresponding weighting strategy.

[0032] Performance parameters and equivalent traffic costs are normalized to eliminate dimensional differences. Adaptive normalization algorithms are used, such as dynamic minimum and maximum normalization based on historical data, or Z-score normalization, where the mean and standard deviation are updated in real-time through a sliding window to adapt to changes in data distribution. A comprehensive score for each traffic path is calculated based on the normalized performance parameters, equivalent traffic costs, and weights.

[0033] Normalization transforms the raw values ​​of performance parameters and equivalent flow costs to a value between 0 and 1, where 1 represents best and 0 represents worst. Normalization includes: For parameters where smaller values ​​are better, including latency, jitter, packet loss rate, and cost, the normalization method is as follows: subtract the parameter value Rcur of the current traffic path from the maximum value Rmax of the parameter across all candidate traffic paths, and then divide the difference by the difference between the maximum value Rmax and the minimum value Rmin of the parameter across all candidate traffic paths. .

[0034] For parameters where larger values ​​are generally better, including bandwidth, the normalization method is as follows: subtract the minimum value Rmin of the parameter among all candidate traffic paths from the parameter value Rcur of the current traffic path, and then divide the difference by the difference between the maximum value Rmax and the minimum value Rmin of the parameter among all candidate traffic paths. .

[0035] The comprehensive score for each traffic path is determined based on the normalized performance parameters, equivalent traffic cost, and corresponding weighting strategy. The parameters are normalized latency, jitter, packet loss rate, bandwidth, and equivalent traffic cost, respectively. w1, w2, w3, w4, and w5 represent the latency weight, jitter weight, packet loss rate weight, bandwidth weight, and cost weight, respectively. After calculating the comprehensive score for each traffic path, the candidate traffic paths are sorted by their comprehensive scores to determine the candidate traffic paths with higher comprehensive scores. The comprehensive score considers both performance and cost; a higher comprehensive score indicates a higher cost-effectiveness.

[0036] For example, for real-time interactive services, a comprehensive score is calculated for a certain traffic path: After normalization, the latency, jitter, packet loss rate, bandwidth, and equivalent traffic cost are 0.9, 0.8, 0.95, 0.7, and 0.4, respectively. The latency weight, jitter weight, packet loss rate weight, bandwidth weight, and cost weight for real-time interactive services are 0.35, 0.25, 0.2, 0.1, and 0.1, respectively. Therefore, the comprehensive score for this traffic path is... .

[0037] Step S40: Based on the comprehensive score of the current traffic path, the comprehensive score of multiple candidate traffic paths, the continuous usage time of the current traffic path, and the number of path switching times of the data stream to be transmitted within the time window, determine whether to switch the path of the data stream to be transmitted.

[0038] Further, based on the comprehensive score of the current traffic path, the comprehensive score of multiple candidate traffic paths, the continuous usage time of the current traffic path, and the number of path switching times of the data stream to be transmitted within the time window, it is determined whether to switch the path of the data stream to be transmitted, including: The system analyzes whether the comprehensive score of a candidate traffic path is significantly higher than that of the current traffic path, whether the continuous usage time of the current traffic path is greater than or equal to the predetermined minimum dwell time, and whether the number of path switching times of the data stream to be transmitted within the time window is less than the predetermined number of switching times. If the analysis results are all yes, then the path to switch the data stream to be transmitted is determined.

[0039] To ensure stable path switching, this embodiment does not immediately switch when a traffic path's overall score is slightly high. Instead, it employs a triple-judgment mechanism: The overall score advantage condition is that the overall score of the candidate traffic path must be significantly higher than that of the current traffic path. This is determined by multiplying the overall score of the candidate traffic path by (one plus a lag threshold). For example, if the lag threshold is 5% and the current traffic path's overall score is 0.8, then the candidate traffic path's overall score must be greater than... Only when the score is greater than 0.84 does it meet the condition for a comprehensive score advantage. Time stability condition: The current traffic path has been used continuously for a period of time. The determination method is: the continuous use time of the current traffic path must be greater than or equal to the preset minimum dwell time (e.g., 60 seconds). Meeting the time stability condition can avoid frequent oscillations. Global frequency condition: Within the set time window, the number of switching times does not reach the upper limit. The judgment method is: the cumulative number of switching of the data stream to be transmitted within the time window must be less than the maximum allowed number of switching times (for example, two). Meeting the global frequency condition can avoid frequent oscillations.

[0040] Only when the conditions of the above triple judgment mechanism are met can the path of the data stream to be transmitted be switched, so as to achieve stable path switching, avoid service jitter caused by frequent switching, and ensure maximum utilization of path resources.

[0041] To further improve stability, the hysteresis threshold and minimum dwell time can be dynamically adjusted, so that these are not fixed values ​​but dynamically adjusted through control theory (such as a PID controller). For example, the hysteresis threshold can be automatically increased based on the degree of network jitter to improve stability.

[0042] In other embodiments, Markov Decision Process (MDP) modeling can be used to decide on path switching. The states include the overall score of the current traffic path, the overall score of the candidate traffic paths, the duration of use of the current traffic path, network load, and switching history, with the corresponding action being to switch or not switch. Based on the initial estimation of the state transition probability P from the state set and the design of the reward function R, the state transition probability P describes the probability that the network environment will transition to a new state s' after a path switching action is performed in state s. The reward function R is used to evaluate the immediate effect of performing a path switching action in state s and reaching the new state s'. Value iteration / policy iteration or reinforcement learning algorithm (such as DQN) is selected, and the model is trained until the policy converges to obtain the path switching model. The path switching model is used to determine whether to switch the path of the data stream to be transmitted.

[0043] Step S50: If yes, a path switching instruction is sent to the network execution module, so that the network execution module can switch the current traffic path of the data stream to be transmitted to the candidate traffic path with the highest comprehensive score based on the path switching instruction.

[0044] In this embodiment, the network execution module can be a software-defined network controller or a smart gateway. After receiving the path switching instruction, the software-defined network controller or smart gateway switches the current traffic path of the data stream to be transmitted (e.g., path A) to a candidate traffic path that has been evaluated as better (e.g., path B).

[0045] Furthermore, when switching paths, a Pareto-optimal search method can be used, such as... Figure 2 As shown, a dashed "lag region" is drawn around the current traffic path point, labeled as the "lag threshold range". When a candidate traffic path point falls outside the lag region and has a higher overall score, a path switching is triggered, and the "Pareto front" is outlined with a dashed line, representing the set of cost-performance optimal traffic paths.

[0046] Step S60: During the transmission period of the data stream to be transmitted, a service quality compliance rate is determined based on the performance parameters, and a continuously changing quality coefficient is generated based on the compliance rate. The higher the compliance rate, the larger the quality coefficient.

[0047] In this embodiment, the service quality compliance rate is calculated after the data stream to be transmitted is completed. Specifically, the service quality compliance rate refers to the proportion of time during which performance parameters (such as latency, packet loss rate, jitter, and bandwidth) meet preset thresholds throughout the entire transmission cycle. For example, the proportion of time with data latency less than 50ms is over 90%, the proportion of time with packet loss rate less than 0.1% is over 95%, the proportion of time with jitter less than 30ms is over 95%, and the proportion of time where actual available bandwidth equals promised bandwidth is over 85%. The service quality compliance rate can be used to comprehensively evaluate the quality of network transmission; a higher compliance rate indicates better network transmission quality. A continuously varying quality coefficient is generated based on the compliance rate. The quality coefficient and the compliance rate have a continuous monotonic relationship; a higher compliance rate results in a larger quality coefficient.

[0048] Furthermore, the quality coefficient is obtained by mapping the compliance rate. This mapping can be achieved using fuzzy logic systems or neural networks, taking into account the nonlinear relationships between multiple indicators. The quality coefficient can be dynamically adjusted or predicted based on SLA default risk.

[0049] Step S70: Generate a traffic settlement bill based on the equivalent traffic cost of the current traffic path, the equivalent traffic cost of the candidate traffic path with the highest comprehensive score, the path switching time, and the quality coefficient.

[0050] This embodiment determines the transmission time of traffic on the corresponding path by using the path switching time. An event-driven architecture can be adopted to aggregate traffic and cost data in real time, and billing and settlement are performed based on the traffic, cost, and transmission time data. Billing costs are dynamically adjusted according to the quality coefficient. A penalty mechanism can be introduced into the settlement process, automatically applying discounts or fines when the quality coefficient falls below a threshold.

[0051] This embodiment links billing costs to actual network traffic quality through a continuous quality coefficient, avoiding the phenomenon of charging high prices for poor service. After the bill is generated, customers can verify the correspondence between cost and service quality through the quality coefficient, resolve disputes, and facilitate auditing.

[0052] This embodiment differs from traditional segmented discounts. This embodiment uses a continuous quality coefficient to measure billing; the higher the quality coefficient, the closer the cost is to the standard billing. When the quality coefficient is close to or equal to the target quality coefficient, the billing cost is close to or equal to the standard cost value. When the quality rating decreases, the billing cost decreases proportionally. When the quality score is significantly insufficient, the billing cost decreases significantly.

[0053] For example, the quality coefficient, after normalization, is a value between 0 and 1: When the quality coefficient is greater than or equal to 90%, the billing cost is equal to the standard cost value; When the compliance rate is between 80% and 90%, the billing cost drops to 85% of the standard cost value; When the compliance rate is below 80%, the billing cost drops to below 70% of the standard cost value.

[0054] This embodiment can use a distributed computing framework (such as Apache Spark) for batch settlement processing, and use stream processing for the real-time part to improve settlement efficiency.

[0055] In this embodiment, after the traffic path is switched, the entity responsible for traffic cost billing changes, and the newly generated data traffic will be billed according to the equivalent traffic cost of the new path. This embodiment records the time when the path switch occurs, thereby accurately dividing the total traffic into time periods of different traffic paths, and applying the respective costs for calculation.

[0056] In addition, the performance and billing parameters of the current traffic path are archived as a basis for evaluating whether the network service quality meets the standards within the corresponding time period, and the collection of performance and billing parameters of the new path is started.

[0057] Taking e-commerce live streaming as an example: Initial stage: Using high-priced dedicated line route A, the overall score is 0.88.

[0058] Changes: Path A's performance declined during peak hours, with its overall score dropping to 0.82; Path B from cloud vendors saw its overall score rise to 0.87.

[0059] Decision-making stage: Score advantage, determine whether 0.87 is greater than 0.861 (i.e. 0.82 * 1.05), the result is yes; Dwell time: Determine whether the 300 seconds used by path A is greater than or equal to a dwell time of 60 seconds; the result is yes. The number of switching times is 0 within the current time window. We check if 0 is less than the preset number of switching times of 2, and the result is yes.

[0060] Switching action: Triggers a path switching command to switch the live stream traffic from path A to path B.

[0061] Cost impact: Subsequent traffic will be billed at the lower cost of path B, and network service quality will begin to be calculated based on the performance parameters of path B.

[0062] Further, a traffic bill is generated based on the equivalent traffic cost of the current traffic path, the equivalent traffic cost of the candidate traffic path with the highest comprehensive score, the path switching time, and the quality coefficient, including: Based on the equivalent traffic cost of the current traffic path, the equivalent traffic cost of the candidate traffic path with the highest comprehensive score, the path switching time, and the quality coefficient, a sharding bill with multiple predetermined dimensions is generated. The multiple predetermined dimensions include region, cloud vendor, operator, and tenant. The traffic settlement bill is generated based on the segmented bill.

[0063] The data is categorized into several dimensions: geographical region (e.g., China, North America, Europe); cloud vendor (e.g., AWS, Alibaba Cloud, Tencent Cloud, Azure); carrier (e.g., China Telecom, China Unicom, China Mobile, and overseas operators); and tenant (e.g., different companies or sub-accounts in multi-tenant scenarios). Based on these dimensions, corresponding sharding bills are generated. For example, sharding bills are generated for China, North America, and Europe, respectively, based on the geographical dimension. Tenants can view their costs and performance through these tenant-level sharding bills. Finally, the sharding bills are globally aggregated to obtain a traffic settlement bill, which includes total costs and overall network performance information.

[0064] In a multi-cloud and multi-carrier hybrid network environment, this invention collects performance and billing parameters in real time, converts the billing parameters into equivalent traffic costs, and comprehensively scores all traffic paths based on the performance parameters and equivalent traffic costs. It then determines whether to switch paths based on the comprehensive score, selecting the path with the highest comprehensive score for switching. Subsequently, it generates a continuously changing quality coefficient based on the service quality compliance rate during transmission to dynamically adjust billing costs. This invention deeply integrates scheduling and settlement, dynamically selecting the most cost-effective network traffic path based on real-time collected performance and equivalent traffic costs. It adapts to dynamic changes in network conditions while considering cost fluctuations and performance status, significantly reducing network costs while ensuring service quality. Furthermore, the settlement scheme is fair and transparent.

[0065] In one embodiment, after collecting the performance parameters and billing parameters corresponding to the current traffic path and multiple candidate traffic paths within a preset time window, the method further includes: The performance parameters are input into a pre-trained prediction model for prediction, and the future performance trends of the current traffic path and multiple candidate traffic paths output by the prediction model are obtained. If the future performance trend of the current traffic path is deterioration, then the overall score of the current traffic path will be reduced according to the deterioration state. If the future performance trend of any candidate traffic path is deteriorating, then after determining the path to switch the data stream to be transmitted, the candidate traffic path with a good future performance trend is selected as the target path for switching.

[0066] The predictive model employs time series modeling or machine learning algorithms, trained on extensive historical data. It forecasts the future performance of various traffic paths, identifying potential network bottlenecks. When the future performance trend of a current traffic path is deteriorating, its overall score is reduced to prepare for path switching and avoid sudden congestion. For example, some traffic is gradually redirected to candidate traffic paths. When a candidate traffic path is predicted to have a deteriorating future performance trend, it is not prioritized during path switching; instead, candidate traffic paths with good future performance trends are given priority to avoid frequent path switching. For instance, before a large e-commerce promotion, if it is predicted that the main path may experience increased latency during peak periods, some traffic is pre-scheduled to backup links to avoid sudden congestion.

[0067] In one embodiment, the cross-cloud network settlement method based on dynamic cost and performance optimization further includes: Write the traffic settlement bill, the performance parameters, the billing parameters, the equivalent traffic cost, and the weighting strategy into a blockchain or an immutable storage system; If a review request is received within the preset dispute period, the traffic settlement bill, performance parameters, billing parameters, equivalent traffic cost, and weighting strategy are obtained from the blockchain or the storage system. The review is performed based on the performance parameters, billing parameters, equivalent traffic cost, and weighting strategy, and a traffic settlement bill to be reviewed is generated. The obtained traffic settlement bill is then compared with the traffic settlement bill to be reviewed.

[0068] In this embodiment, a billing dispute period and recalculation mechanism are introduced to enhance the transparency and credibility of settlement. Specifically, data notarization involves hashing the traffic settlement bill, performance parameters, billing parameters, equivalent traffic cost, weighting strategy, and abnormal data processing records after the settlement bill is generated (e.g., SHA-256 hash signature) and writing it to a blockchain (such as an Ethereum private chain) or other equivalent immutable storage system, ensuring the authenticity and reliability of the settlement bill and its related data.

[0069] A dispute period is set for each invoice, such as seven days, during which tenants can apply for a review. Invoice recalculation: Upon receiving a customer's review request, the invoice generation process is fully reproduced based on data stored on the blockchain or other storage systems, and the recalculated result is output. Smart contracts automatically handle disputed recalculations, providing a verifiable audit trail. This reduces settlement disputes and makes auditing more convenient. Furthermore, if the tenant still has doubts, the settlement invoice and related data can be submitted to a third-party auditing firm for independent verification.

[0070] Figure 3 A schematic diagram of the cross-cloud network settlement system based on dynamic cost and performance optimization according to an embodiment of the present invention is shown. Figure 3 As shown, the system includes: The data acquisition and cost calculation module 100 is used to collect the performance parameters and billing parameters corresponding to the current traffic path and multiple candidate traffic paths within a preset time window, and to convert the billing parameters into equivalent traffic costs based on predetermined billing rules. The identification module 200 is used to identify the service type of the data stream to be transmitted in the current traffic path. Strategy library 300 is used to obtain the weight strategy corresponding to the business type from a preset strategy library; The dynamic scheduling engine 400 is used to normalize the performance parameters and the equivalent traffic cost, and determine the comprehensive score of each traffic path based on the normalized performance parameters, equivalent traffic cost, and corresponding weighting strategy; based on the comprehensive score of the current traffic path, the comprehensive scores of multiple candidate traffic paths, the continuous usage time of the current traffic path, and the number of path switching times of the data stream to be transmitted within the time window, it determines whether to switch the path of the data stream to be transmitted; if so, it issues a path switching instruction to the network execution module. The network execution module 500 is used to switch the current traffic path of the data stream to be transmitted to the candidate traffic path with the highest comprehensive score based on the path switching instruction. The billing and settlement module 600 is used to determine the service quality compliance rate based on the performance parameters during the transmission period of the data stream to be transmitted, generate a continuously changing quality coefficient based on the compliance rate, and the higher the compliance rate, the larger the quality coefficient; and generate a traffic settlement bill based on the equivalent traffic cost of the current traffic path, the equivalent traffic cost of the candidate traffic path with the highest comprehensive score, the path switching time, and the quality coefficient.

[0071] The implementation of the cross-cloud network settlement system based on dynamic cost and performance optimization is basically the same as the implementation of the cross-cloud network settlement method based on dynamic cost and performance optimization described above, and can be referred to the above implementation.

[0072] Figure 4 The diagram shows a structural schematic of an embodiment of the computer device of the present invention. The specific embodiments of the present invention do not limit the specific implementation of the computer device.

[0073] like Figure 4 As shown, the computer device may include: a processor 402, a communications interface 404, a memory 406, and a communications bus 408.

[0074] The processor 402, communication interface 404, and memory 406 communicate with each other via communication bus 408. Communication interface 404 is used to communicate with other computer devices, such as clients or other server network elements. The processor 402 executes program 410, specifically performing the relevant steps described above in the computer device embodiment.

[0075] Specifically, program 410 may include program code, which includes computer-executable instructions.

[0076] Processor 402 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention. The computer device includes one or more processors, which may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0077] Memory 406 is used to store program 410. Memory 406 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0078] Specifically, program 410 can be called by processor 402 to cause the computer device to perform the following operations: Within a preset time window, collect the performance parameters and billing parameters corresponding to the current traffic path and multiple candidate traffic paths respectively, and convert the billing parameters into equivalent traffic costs based on predetermined billing rules. Identify the service type of the data stream to be transmitted in the current traffic path, and obtain the weight strategy corresponding to the service type from the preset strategy library; The performance parameters and the equivalent traffic cost are normalized, and the comprehensive score of each traffic path is determined based on the normalized performance parameters, equivalent traffic cost and the corresponding weighting strategy. Based on the comprehensive score of the current traffic path, the comprehensive score of multiple candidate traffic paths, the continuous usage time of the current traffic path, and the number of path switching times of the data stream to be transmitted within the time window, it is determined whether to switch the path of the data stream to be transmitted. If so, a path switching instruction is sent to the network execution module, which then switches the current traffic path of the data stream to be transmitted to the candidate traffic path with the highest comprehensive score based on the path switching instruction. During the transmission period of the data stream to be transmitted, a service quality compliance rate is determined based on the performance parameters, and a continuously changing quality coefficient is generated based on the compliance rate. The higher the compliance rate, the larger the quality coefficient. A traffic settlement bill is generated based on the equivalent traffic cost of the current traffic path, the equivalent traffic cost of the candidate traffic path with the highest comprehensive score, the path switching time, and the quality coefficient.

[0079] This invention provides a computer-readable storage medium storing at least one executable instruction that, when executed on a computer device, causes the computer device to perform any of the above-described method embodiments.

[0080] This invention provides a computer program that can be invoked by a processor to cause a computer device to execute any of the above-described method embodiments.

[0081] This invention provides a computer program product, which includes a computer program stored on a computer-readable storage medium. The computer program includes program instructions that, when executed on a computer, cause the computer to perform any of the above-described method embodiments.

[0082] The algorithms or displays provided herein are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems can also be used in conjunction with the teachings herein. The required structure for constructing such systems is apparent from the above description. Furthermore, the embodiments of the present invention are not directed to any particular programming language. It should be understood that the content of the invention described herein can be implemented using various programming languages, and the above description of specific languages ​​is for the purpose of disclosing the best mode of implementation of the invention.

[0083] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0084] Similarly, it should be understood that, in order to streamline the invention and aid in understanding one or more of the various aspects of the invention, features of the embodiments of the invention are sometimes grouped together in a single embodiment, figure, or description thereof in the above description of exemplary embodiments of the invention. However, this disclosure should not be construed as reflecting an intention that the claimed invention requires more features than are expressly recited in each claim.

[0085] Those skilled in the art will understand that modules in the computer device of the embodiments can be adaptively modified and placed in one or more computer devices different from that embodiment. Modules, units, or components in the embodiments can be combined into a single module, unit, or component, and can be divided into multiple sub-modules, sub-units, or sub-components. Except where at least some of such features and / or processes or units are mutually exclusive, any combination can be used to combine all features disclosed in this specification (including the accompanying claims, abstract, and drawings) and all processes or units of any method or computer device so disclosed. Unless expressly stated otherwise, each feature disclosed in this specification (including the accompanying claims, abstract, and drawings) may be replaced by an alternative feature that serves the same, equivalent, or similar purpose.

[0086] It should be noted that the above embodiments are illustrative of the invention and not restrictive, and that those skilled in the art can devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses should not be construed as limiting the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The invention can be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third, etc., does not indicate any order. These words can be interpreted as names. The steps in the above embodiments, unless otherwise specified, should not be construed as limiting the order of execution.

Claims

1. A cross-cloud network settlement method based on dynamic optimization of cost and performance, characterized in that, The method includes: Within a preset time window, collect the performance parameters and billing parameters corresponding to the current traffic path and multiple candidate traffic paths respectively. Based on the predetermined billing rules, convert the billing parameters into equivalent traffic cost. The billing rules include 95th percentile billing, tiered pricing, monthly overage fees, and cross-domain surcharges. The equivalent traffic cost is the cost of transmitting one unit of traffic within the current time window. Identify the service type of the data stream to be transmitted in the current traffic path, and obtain the weight strategy corresponding to the service type from the preset strategy library; The performance parameters and the equivalent traffic cost are normalized, and the comprehensive score of each traffic path is determined based on the normalized performance parameters, equivalent traffic cost and the corresponding weighting strategy. Based on the comprehensive score of the current traffic path, the comprehensive scores of multiple candidate traffic paths, the continuous usage time of the current traffic path, and the number of path switching times of the data stream to be transmitted within the time window, it is determined whether to switch the path of the data stream to be transmitted. This includes analyzing whether there is a candidate traffic path with a comprehensive score higher than the comprehensive score of the current traffic path, whether the continuous usage time of the current traffic path is greater than or equal to a predetermined minimum dwell time, and whether the number of path switching times of the data stream to be transmitted within the time window is less than a predetermined number of switching times. If the analysis results are all yes, then it is determined to switch the path of the data stream to be transmitted. If so, a path switching instruction is sent to the network execution module, which then switches the current traffic path of the data stream to be transmitted to the candidate traffic path with the highest comprehensive score based on the path switching instruction. During the transmission period of the data stream to be transmitted, a service quality compliance rate is determined based on the performance parameters, and a continuously changing quality coefficient is generated based on the compliance rate. The higher the compliance rate, the larger the quality coefficient. A traffic settlement bill is generated based on the equivalent traffic cost of the current traffic path, the equivalent traffic cost of the candidate traffic path with the highest comprehensive score, the path switching time, and the quality coefficient. The path switching time is used to determine the transmission time of traffic on the corresponding path. By recording the path switching time, the total traffic is divided into time periods of different traffic paths, and the equivalent traffic cost of each path is applied for calculation.

2. The method according to claim 1, characterized in that, The strategy library includes a dynamic optimization engine, which dynamically optimizes the weights in the weight strategy for each of the business types according to a predetermined optimization algorithm. The optimization algorithm includes a value iteration-based optimization algorithm, a strategy gradient algorithm, and a multi-armed gambling machine algorithm.

3. The method according to claim 1, characterized in that, The process of generating a traffic bill based on the equivalent traffic cost of the current traffic path, the equivalent traffic cost of the candidate traffic path with the highest comprehensive score, the path switching time, and the quality coefficient includes: Based on the equivalent traffic cost of the current traffic path, the equivalent traffic cost of the candidate traffic path with the highest comprehensive score, the path switching time, and the quality coefficient, a sharding bill with multiple predetermined dimensions is generated. The multiple predetermined dimensions include region, cloud vendor, operator, and tenant. The traffic settlement bill is generated based on the segmented bill.

4. The method according to claim 1, characterized in that, The method further includes: Write the traffic settlement bill, the performance parameters, the billing parameters, the equivalent traffic cost, and the weighting strategy into a blockchain or an immutable storage system; If a review request is received within the preset dispute period, the traffic settlement bill, performance parameters, billing parameters, equivalent traffic cost, and weighting strategy are obtained from the blockchain or the storage system. The review is performed based on the performance parameters, billing parameters, equivalent traffic cost, and weighting strategy, and a traffic settlement bill to be reviewed is generated. The obtained traffic settlement bill is then compared with the traffic settlement bill to be reviewed.

5. The method according to claim 1, characterized in that, After collecting the performance parameters and billing parameters corresponding to the current traffic path and multiple candidate traffic paths within a preset time window, the method further includes: The performance parameters are input into a pre-trained prediction model for prediction, and the future performance trends of the current traffic path and multiple candidate traffic paths output by the prediction model are obtained. If the future performance trend of the current traffic path is deterioration, then the overall score of the current traffic path will be reduced according to the deterioration state. If the future performance trend of any candidate traffic path is deteriorating, then after determining the path to switch the data stream to be transmitted, the candidate traffic path with a good future performance trend is selected as the target path for switching.

6. The method according to claim 1, characterized in that, The service types include at least real-time interaction, live streaming, and large file transfer. The service types for identifying the data stream to be transmitted based on the current traffic path include: The service type of the data stream to be transmitted can be identified based on the packet size, packet interval, traffic rate, and protocol type of the current traffic path; or based on a pre-trained multi-feature fusion classification model.

7. A cross-cloud network settlement system based on dynamic cost and performance optimization, characterized in that, The system includes: The data acquisition and cost calculation module is used to collect the performance parameters and billing parameters corresponding to the current traffic path and multiple candidate traffic paths within a preset time window. Based on the predetermined billing rules, the billing parameters are uniformly converted into equivalent traffic costs. The billing rules include 95th percentile billing, tiered pricing, monthly overage fees, and cross-domain surcharges. The equivalent traffic cost is the cost of transmitting one unit of traffic within the current time window. The identification module is used to identify the service type of the data stream to be transmitted in the current traffic path; A strategy library is used to obtain the weight strategy corresponding to the business type from a preset strategy library; A dynamic scheduling engine is used to normalize the performance parameters and the equivalent traffic cost. Based on the normalized performance parameters, equivalent traffic cost, and corresponding weighting strategies, a comprehensive score for each traffic path is determined. Based on the comprehensive score of the current traffic path, the comprehensive scores of multiple candidate traffic paths, the duration of use of the current traffic path, and the number of path switching times of the data stream to be transmitted within the time window, a decision is made on whether to switch the path of the data stream to be transmitted. This includes analyzing whether there is a candidate traffic path with a higher comprehensive score than the current traffic path, whether the duration of use of the current traffic path is greater than or equal to a predetermined minimum dwell time, and whether the number of path switching times of the data stream to be transmitted within the time window is less than a predetermined number of switching times. If all the analysis results are yes, then the path of the data stream to be transmitted is switched; otherwise, a path switching command is issued to the network execution module. The network execution module is used to switch the current traffic path of the data stream to be transmitted to the candidate traffic path with the highest comprehensive score based on the path switching instruction. The billing and settlement module is used to determine the service quality compliance rate based on the performance parameters within the transmission period of the data stream to be transmitted, and generate a continuously changing quality coefficient based on the compliance rate, wherein the higher the compliance rate, the larger the quality coefficient; and generate a traffic settlement bill based on the equivalent traffic cost of the current traffic path, the equivalent traffic cost of the candidate traffic path with the highest comprehensive score, the path switching time, and the quality coefficient, wherein the path switching time is used to determine the transmission time of traffic on the corresponding path, and by recording the path switching time, the total traffic is divided into time periods of different traffic paths, and the equivalent traffic cost of each path is applied for calculation.

8. A computer device, characterized in that, include: The processor, memory, communication interface, and communication bus are provided, wherein the processor, memory, and communication interface communicate with each other via the communication bus. The memory is used to store at least one executable instruction that causes the processor to perform the method as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The storage medium stores at least one executable instruction, which, when executed on a computer device, causes the computer device to perform the method as described in any one of claims 1-6.

Images