Control method and electronic equipment

Abstract

The present disclosure provides a control method and an electronic device, the method comprising: based on a first prediction model, determining an expected network transmission rate demand requested by one or more processes running on the device within an upcoming time window T; determining a predicted available network bandwidth at one or more time points within the time window T based on a second prediction model; determining a priority score of network resource scheduling of one or more processes based on a plurality of preset network resource scheduling parameters; comparing, at least in part, the expected network transmission rate demand with the predicted available network bandwidth, and generating a network resource allocation policy based on the comparison result and the priority score of the network resource scheduling of the one or more processes; and scheduling access of one or more processes to the network resources according to the network resource allocation strategy. According to the control method disclosed by the invention, the efficiency and reliability of network transmission can be remarkably improved through accurate prediction and an intelligent scheduling strategy.

Description

Technical Field

[0001] This disclosure relates to the technical field of network resource control, and more particularly to a control method and electronic device. Background Technology

[0002] With the rapid development of information technology, network transmission technology is becoming increasingly important in daily life and various applications. Most current network transmission technologies employ a best-effort transmission mode, aiming to transmit data as much as possible. However, this mode performs poorly in certain scenarios with high latency requirements, failing to effectively guarantee data transmission latency. For example, in applications such as real-time video conferencing, online gaming, and telemedicine, low latency is a key factor in user experience.

[0003] Furthermore, existing technologies have limited capabilities in handling network latency and lack effective mechanisms to monitor and adjust transmission strategies in real time to adapt to different network environments and application requirements. This limitation makes it difficult for existing network transmission technologies to meet user expectations in demanding application scenarios, and a new solution is urgently needed to improve the reliability and real-time performance of data transmission. Summary of the Invention

[0004] To address the problems in the related technologies, embodiments of this disclosure provide a control method and an electronic device.

[0005] One aspect of this disclosure provides a control method, comprising: determining, based on a first prediction model, an expected network transmission rate demand requested by one or more processes running on a device within an upcoming time window T; determining, based on a second prediction model, a predicted available network bandwidth at one or more time points within the time window T; determining a priority score for network resource scheduling of one or more processes based on a plurality of preset network resource scheduling parameters; at least partially comparing the expected network transmission rate demand with the predicted available network bandwidth, and generating a network resource allocation strategy based on the comparison result and the priority scores for network resource scheduling of one or more processes; and scheduling access to network resources by one or more processes according to the network resource allocation strategy.

[0006] According to embodiments of this disclosure, at least partially comparing the expected network transmission rate demand with the predicted available network bandwidth includes: generating a total expected transmission rate demand at each time point within a time window T based on the expected network transmission rate of each of one or more processes; and for at least one specific time point within the time window T, calculating the difference between the total expected transmission rate demand and the predicted available network bandwidth at that specific time point as a supply-demand gap.

[0007] According to embodiments of this disclosure, the plurality of preset network resource scheduling parameters include at least two of the following parameters: a preset priority category for the process specified by the user or the system; a congestion sensitivity coefficient output by a first prediction model to indicate the process's sensitivity to network congestion; an identifier indicating whether the process is in a critical data transmission window; and the network transmission rate currently allocated to the process.

[0008] According to embodiments of this disclosure, a network resource allocation strategy is generated based on comparison results and priority scores of network resource scheduling for one or more processes, including: dividing one or more processes into multiple groups based on priority scores of network resource scheduling for one or more processes, wherein the multiple groups include at least a guaranteed group with the highest priority score and a control group with the lowest priority score; and in response to determining that there is a supply-demand gap indicating insufficient bandwidth, prioritizing the allocation of available network bandwidth to processes in the guaranteed group.

[0009] According to an embodiment of this disclosure, the control method further includes, in response to determining that there is a supply-demand gap indicating insufficient bandwidth, selecting one or more first processes in the control group as first candidate processes for network rate compression; and reducing the current first network rate of the first candidate processes by a first compression ratio to reduce the supply-demand gap.

[0010] According to an embodiment of this disclosure, the control method further includes, in response to reducing the current first network rate of the first candidate process by a first compression ratio, if the supply-demand gap still exists, selecting one or more first processes in the watch group as second candidate processes for network rate compression; reducing the current second network rate of the second candidate processes by a second compression ratio to further reduce the supply-demand gap, wherein the second compression ratio is less than the first compression ratio, and wherein the watch group includes processes with medium priority scores.

[0011] According to embodiments of this disclosure, multiple time points include a first time point and a second time point, the second time point being later than the first time point; the method further includes: in response to determining that there is a supply-demand gap due to insufficient bandwidth at the first time point, reducing the first network rate allocation of one or more processes in the control group by a first compression ratio to reduce the supply-demand gap; in response to determining that there is sufficient bandwidth and no supply-demand gap at the second time point, restoring the first network rate allocation of one or more processes in the control group, but limiting the total rate of all processes in the control group to below a predetermined proportion of the predicted available network bandwidth.

[0012] According to embodiments of this disclosure, the first prediction model is a first AI model, which predicts the expected network rate demand of one or more processes based on learning the historical and real-time network usage characteristics; the second prediction model is a second AI model, which predicts the future availability of network bandwidth based on learning the current and historical network latency characteristics.

[0013] According to embodiments of this disclosure, determining the expected network transmission rate demand requested by one or more processes running on the device within the upcoming time window T, based on a first prediction model, includes: processing the historical and real-time network usage characteristics of one or more processes using a first AI model to obtain the expected network transmission rate demand requested by one or more processes running on the device within the upcoming time window T; the historical and real-time network usage characteristics include at least any two or more of the following: priority identifier, historical network rate, historical connection duration, size of historical transmitted data packets, identifier of the currently accessed port, and size of the currently transmitted data packet; determining the predicted available network bandwidth at one or more time points within the time window T, based on a second prediction model, includes: processing the current and historical network latency characteristics of one or more processes using a second AI model to obtain the predicted available network bandwidth at one or more time points within the time window T; the current and historical network latency characteristics include at least any two or more of the following: available bandwidth, bandwidth fluctuation range, network type, historical bandwidth usage data, historical bandwidth usage scenarios, current process latency, current process congestion, and current process bandwidth usage.

[0014] Another aspect of this disclosure provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the method described above.

[0015] According to the technical solution of this disclosure, through analysis based on the first and second prediction models, the network transmission requirements and available bandwidth of each process can be accurately predicted, thereby achieving reasonable scheduling and allocation of network resources, maximizing the utilization rate of network resources. By dynamically adjusting the network resource allocation strategy within the time window, changes in network status can be responded to in a timely manner, ensuring that high-priority processes obtain the required bandwidth, thereby effectively reducing data transmission latency and improving the performance of real-time applications. Based on multiple preset network resource scheduling parameters, priority scores are generated, which can flexibly adjust the allocation of network resources according to the actual needs and importance of different processes, optimize the overall network scheduling strategy, ensure the priority execution of critical tasks, and adapt to different network environments and load conditions through real-time monitoring and prediction of network status, thereby maintaining stable network transmission performance in various complex scenarios, improving user experience. Through scientific scheduling of network resources, congestion caused by uneven allocation of network resources can be effectively avoided, reducing packet loss rate and improving the reliability of network transmission. Attached Figure Description

[0016] Other features, objects, and advantages of this disclosure will become more apparent from the following detailed description of non-limiting embodiments, taken in conjunction with the accompanying drawings. In the drawings: Figure 1 A flowchart illustrating a control method according to an embodiment of the present disclosure is shown schematically. Figure 2 A flowchart illustrating a control method according to another embodiment of the present disclosure is shown schematically; Figure 3 A flowchart illustrating a control method according to another embodiment of the present disclosure is shown schematically; Figure 4 A flowchart illustrating a control method according to another embodiment of the present disclosure is shown schematically; Figure 5 This illustration schematically shows an application diagram of the control method according to an embodiment of the present disclosure; Figure 6 A schematic block diagram of an electronic device according to an embodiment of the present disclosure is shown. Detailed Implementation

[0017] Hereinafter, exemplary embodiments of the present disclosure will be described in detail with reference to the accompanying drawings to enable those skilled in the art to implement them more readily. Furthermore, for clarity, portions unrelated to the description of exemplary embodiments have been omitted from the drawings. In this disclosure, it should be understood that terms such as "comprising" or "having" are intended to indicate the presence of features, numbers, steps, behaviors, components, portions, or combinations thereof disclosed in this specification, and are not intended to exclude the possibility of the presence or addition of one or more other features, numbers, steps, behaviors, components, portions, or combinations thereof. It should also be noted that, without conflict, embodiments and features in the embodiments of this disclosure can be combined with each other. The present disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0018] Figure 1 A flowchart illustrating a control method according to an embodiment of the present disclosure is shown schematically.

[0019] like Figure 1 As shown, it includes steps S110-S150.

[0020] In step S110, based on the first prediction model, the expected network transmission rate demand requested by one or more processes running on the device within the upcoming time window T is determined.

[0021] For example, the first prediction model is a first AI model, which predicts the expected network rate requirements of one or more processes based on learning from their historical and real-time network usage characteristics. In this embodiment, different processes correspond to different applications.

[0022] In this embodiment, determining the expected network transmission rate demand requested by one or more processes running on the device within the upcoming time window T, based on the first prediction model, includes: processing the historical and real-time network usage characteristics of one or more processes using the first AI model to obtain the expected network transmission rate demand requested by one or more processes running on the device within the upcoming time window T; the historical and real-time network usage characteristics include at least any two or more of the following: priority identifier, historical network rate, historical connection duration, size of historical transmitted data packets, identifier of the currently accessed port, and size of the currently transmitted data packet.

[0023] For example, a priority identifier can be a priority label assigned to each process, indicating its relative importance in network resource scheduling. Processes with higher priority will receive more bandwidth during network resource allocation to ensure their performance requirements are met. Historical network rate can record the network transmission rate of each process over a period of time. This feature helps identify the process's network usage patterns and understand its performance under different load conditions. Historical connection duration can record the duration of continuous activity of a process in past connections. This feature can reflect the activity level of the process and help the model determine network demand in future time windows. Historical packet size can analyze the size of packets transmitted by a process in historical connections. This feature helps understand the bandwidth demand characteristics of a process, thus affecting the prediction of future network rates. The currently accessed port identifier can identify the network port currently used by a process. This information helps the model understand the specific network activity of the process and compare it with historical data. The size of the currently transmitted packet can monitor the size of the packets currently being transmitted by a process in real time. This feature can reflect the process's immediate needs in the current network environment and provide the latest information for prediction.

[0024] Based on the aforementioned features, the first AI model can be trained. For example, historical and real-time network usage features of each process on the device can be collected, and the data can be cleaned and standardized to ensure the quality and consistency of the model input. The aforementioned features can be extracted from the collected data to construct feature vectors for model training and prediction. The first AI model can be trained using historical data. By learning the network usage patterns of processes, the model establishes the relationship between features and network rate requirements. Within the upcoming time window T, the trained model is used to predict the expected network transmission rate requirements of one or more processes. The model will output the network rate requirements of each process based on the combination of real-time and historical features. Finally, the first prediction model will output the expected network transmission rate requirements of each process within the time window T, providing a basis for subsequent network resource scheduling.

[0025] The first prediction model mentioned above not only relies on rich historical and real-time network usage characteristics, but also, through an intelligent learning and prediction process, can provide accurate predictions of the network requirements of multiple processes on the device. The effectiveness and accuracy of this model directly affect subsequent network resource scheduling strategies, thereby improving overall network performance and user experience.

[0026] In step S120, based on the second prediction model, the predicted available network bandwidth at one or more time points within the time window T is determined.

[0027] For example, the second prediction model is a second AI model, which predicts the availability of future network bandwidth based on learning from current and historical network latency characteristics.

[0028] In this embodiment, determining the predicted available network bandwidth at one or more time points within a time window T, based on the second prediction model, includes: processing the current and historical network latency characteristics of one or more processes using the second AI model to obtain the predicted available network bandwidth at one or more time points within the time window T; the current and historical network latency characteristics include at least any two or more of the following: available bandwidth, bandwidth fluctuation amplitude, network type, historical bandwidth usage data, historical bandwidth usage scenarios, current process latency, current process congestion, and current process bandwidth usage. Different processes correspond to different applications.

[0029] For example, available bandwidth refers to the amount of bandwidth available for data transmission in the current network environment. Real-time monitoring of available bandwidth can reflect the actual network usage and provide basic data for prediction. Bandwidth fluctuation can describe changes in network bandwidth, reflecting the stability of the network at different points in time. Analyzing bandwidth fluctuation helps to determine the reliability of the network environment and affects the prediction of future bandwidth availability. Network type includes the type of network connection, such as WiFi, Ethernet, 4G / 5G, etc. Different network types have different bandwidth characteristics and latency performance. Historical bandwidth usage data can record bandwidth usage over a period of time, including bandwidth usage during peak and off-peak periods. This data can help the model identify patterns and trends in bandwidth usage. Historical bandwidth usage scenarios can analyze bandwidth usage in different scenarios, such as video streaming, file downloading, online gaming, etc. Different application scenarios have significantly different bandwidth requirements, and historical data can provide a reference for future predictions. Current process latency can monitor the network latency of the current process in real time, including the round-trip time of data packets. This feature can reflect the current health status of the network and has a direct impact on bandwidth availability prediction. Current process congestion can assess whether the current process is affected by network congestion, including data packet loss, retransmission, etc. Congestion analysis helps determine the actual available bandwidth of the network; the current process bandwidth usage can be monitored in real time to see how much bandwidth the current process is using. This information can directly affect the prediction of future bandwidth availability because the process's bandwidth usage is closely related to the available bandwidth.

[0030] Based on the aforementioned features, a second AI model can be trained. For example, current and historical network latency features of various processes on the device can be collected, and the data can be cleaned and standardized to ensure the quality and consistency of the model input. The aforementioned features can be extracted from the collected data to construct feature vectors for model training and prediction. The second AI model can be trained using historical data. By learning the relationship between network latency features and bandwidth availability, the model establishes a connection between features and future bandwidth predictions. Within a time window T, the trained model is used to predict the available network bandwidth at one or more time points. The model will output the available bandwidth at a specified time point based on the combination of real-time features and historical features. Finally, the second prediction model will output the predicted available network bandwidth at each time point within the time window T, providing a basis for subsequent network resource scheduling strategies.

[0031] The second prediction model mentioned above not only relies on rich current and historical network latency characteristics, but also provides accurate predictions of future network bandwidth availability through an intelligent learning and prediction process. The effectiveness and accuracy of this model directly affect subsequent network resource scheduling strategies, thereby improving overall network performance and user experience, and ensuring network quality in demanding application scenarios.

[0032] In step S130, the priority score of network resource scheduling for one or more processes is determined based on multiple preset network resource scheduling parameters.

[0033] For example, the multiple preset network resource scheduling parameters include at least two of the following parameters: a preset priority category for the process specified by the user or the system; a congestion sensitivity coefficient output by the first prediction model to indicate the process's sensitivity to network congestion; an identifier indicating whether the process is in a critical data transmission window; and the network transmission rate currently allocated to the process.

[0034] Specifically, the preset priority category of a process can be specified by the user or the system, reflecting the importance level of the process. Processes in higher priority categories will be assigned higher priority scores in network resource scheduling. For example, applications that are sensitive to latency, such as real-time communication and online games, may be assigned high priority, while non-real-time applications, such as background data synchronization, may be assigned low priority.

[0035] The congestion sensitivity coefficient can be a value output by the first prediction model, used to indicate the sensitivity of a process to network congestion. Processes with high sensitivity will be more affected by network congestion, and therefore should be given higher priority scores during resource scheduling. This coefficient can be calculated by analyzing the historical network performance of a process, reflecting its performance requirements under congestion conditions.

[0036] A critical data transmission window identifier indicates whether a process is in a critical data transmission window. For example, during video conferencing or real-time data transmission, a process may be in a critical window, in which case priority should be given to allocating network resources to it.

[0037] The currently allocated network transmission rate represents the actual network transmission rate that a process is currently receiving. This parameter reflects the process's current network usage and thus affects its priority score. If the process's current transmission rate is lower than its requirement, it may need to increase its priority to meet its network needs.

[0038] To effectively consider the impact of the above parameters, different weights can be pre-set for each parameter. The weights can be set based on user needs, system strategies, or historical data analysis. The following is an example of weight allocation in this embodiment: The default priority category of a process has a weight of 40%. This parameter is the main factor affecting scheduling priority, ensuring that high-priority processes receive priority support in resource allocation.

[0039] The congestion sensitivity coefficient has a weight of 30%. This parameter reflects the sensitivity of a process to changes in network state and can effectively identify processes that need to be prioritized in the event of network congestion.

[0040] The critical data transmission window identifier has a weight of 20%. This parameter ensures that the process can obtain the necessary network resources at critical moments (such as real-time data transmission), thereby guaranteeing the reliability and timeliness of data transmission.

[0041] The currently allocated network transmission rate has a weight of 10%. Although this parameter has a low weight, it still provides feedback on the process's current network usage and helps adjust the priority score.

[0042] By weighting the values ​​of each parameter with their corresponding weights, the network resource scheduling priority score for each process can be obtained. The specific calculation formula is as follows:

[0043] in: It is the value of the process's preset priority category. It is the value of the congestion sensitivity coefficient. This is the value of the critical data transmission window identifier (usually 0 or 1). It is the value of the currently allocated network transmission rate. These are the corresponding weights.

[0044] Through the detailed description of the above parameters and their weights, it can be seen that the priority score calculated based on multiple preset network resource scheduling parameters can effectively reflect the network resource needs and importance of a process. This mechanism ensures that, under the condition of limited network resources, processes that are sensitive to latency and of high importance can be given priority support, thereby optimizing overall network performance and user experience.

[0045] In step S140, the expected network transmission rate demand is compared with the predicted available network bandwidth at least in part, and a network resource allocation strategy is generated based on the comparison results and the priority scores of network resource scheduling of one or more processes.

[0046] For example, at least in part, the difference between the expected network transmission rate demand and the predicted available network bandwidth can be calculated and compared. Based on this difference and the priority scores of network resource scheduling for one or more processes, one or more processes can be classified. Based on the classification results, targeted network resource allocation strategies can be generated.

[0047] By partially comparing the expected network transmission rate demand with the predicted available network bandwidth and combining it with the priority scores of processes to generate a network resource allocation strategy, intelligent scheduling of network resources can be achieved. This not only improves the utilization efficiency of network resources, but also ensures support for high-priority and latency-sensitive processes, thereby optimizing overall network performance and user experience.

[0048] In step S150, one or more processes are scheduled to access network resources according to the network resource allocation strategy.

[0049] For example, based on the network resource allocation strategy, processes can be prioritized. High-priority processes should be given priority in scheduling to ensure they receive sufficient bandwidth when network resources are limited. Scheduling is then performed according to the allocated bandwidth for each process based on priority. If the demand exceeds the available bandwidth, dynamic bandwidth adjustment is required to ensure high-priority processes receive the necessary bandwidth. Network bandwidth is allocated to each process according to the scheduling order. For instance, if a high-priority process requires 500kbps and the available bandwidth is 1000kbps, that process may receive 100% of its demand, while other low-priority processes receive the remaining bandwidth according to the strategy. During resource allocation, network status and actual process usage should be monitored in real time. If a process's actual demand changes, bandwidth allocation should be dynamically adjusted to adapt to the new network conditions. If a process fails to use the allocated bandwidth as expected, or if network conditions change (such as bandwidth fluctuations or network congestion), the system should be able to dynamically adjust the resource allocation strategy based on real-time feedback and reallocate bandwidth to optimize network resource utilization.

[0050] Through the above scheduling process, the access of processes to network resources can be effectively managed according to the network resource allocation strategy. This process not only ensures that the network needs of high-priority processes are met, but also flexibly adjusts resource allocation when the network environment changes, thereby improving overall network performance and user experience. Intelligent network resource scheduling will help provide stable and efficient network services in various application scenarios.

[0051] By analyzing based on the first and second prediction models, the network transmission requirements and available bandwidth of each process can be accurately predicted, thereby achieving reasonable scheduling and allocation of network resources, maximizing network resource utilization. By dynamically adjusting the network resource allocation strategy within a time window, it can respond promptly to changes in network status, ensuring that high-priority processes obtain the necessary bandwidth, effectively reducing data transmission latency and improving the performance of real-time applications. Based on multiple preset network resource scheduling parameters, priority scores are generated, allowing for flexible adjustment of network resource allocation according to the actual needs and importance of different processes, optimizing the overall network scheduling strategy, ensuring the priority execution of critical tasks. Through real-time monitoring and prediction of network status, it can adapt to different network environments and load conditions, maintaining stable network transmission performance in various complex scenarios, improving user experience. Scientific scheduling of network resources can effectively avoid congestion caused by uneven network resource allocation, reduce packet loss rate, and improve network transmission reliability. This control method is not only suitable for latency-sensitive applications such as real-time communication and online games, but can also be widely applied in fields such as telemedicine and intelligent manufacturing, meeting the network transmission needs of different scenarios.

[0052] Figure 2 A flowchart illustrating a control method according to another embodiment of the present disclosure is shown schematically.

[0053] like Figure 2 As shown, based on the foregoing embodiments, comparing the expected network transmission rate requirement with the predicted available network bandwidth may include steps S210 and S220.

[0054] In step S210, the total expected transmission rate requirement at each time point within the time window T is generated based on the expected network transmission rate of each of one or more processes.

[0055] For example, a time window T can refer to a specific time period during which the network transmission requirements of various processes need to be predicted. The time window T can vary depending on the application scenario, for example, it can be a few seconds, a few minutes, or a longer period. The process of generating the total expected transmission rate requirement at each time point within the time window T can divide the time window T into multiple time points. These time points can be evenly spaced (e.g., per second, per minute) or dynamically adjusted according to actual needs (e.g., shortening the time interval under high load). The expected rate for each process is obtained. For each process running within the time window T, its expected network transmission rate requirement at each time point is obtained through a first prediction model. This requirement is usually derived from a combination of factors such as the process's historical usage patterns, current state, and preset priority. For each time point, the sum of the expected network transmission rate requirements of all processes is calculated, and the total expected transmission rate requirement at each time point is recorded to form a time series data. This data will be used for subsequent network resource scheduling and allocation.

[0056] Through the above steps, the total expected transmission rate demand at each time point within the time window T is generated based on the expected network transmission rate of one or more processes. This can comprehensively reflect the demand for network resources. This process provides important data support for subsequent resource allocation and scheduling, and helps to optimize network performance and improve user experience.

[0057] In step S220, for at least one specific time point within the time window T, the difference between the total expected transmission rate demand and the predicted available network bandwidth at the specific time point is calculated as the supply-demand gap.

[0058] For example, for time point t, the total expected transmission rate demand S(t) is compared with the predicted available network bandwidth B(t), and the difference between the total expected transmission rate demand at time t and the predicted available network bandwidth at that specific time point is calculated as the supply-demand gap. :

[0059] if This indicates that demand exceeds supply, and the network may face congestion risks, requiring network resource reallocation; if This indicates that the available bandwidth exceeds the demand, and network resources can be effectively utilized, meaning that there is sufficient bandwidth and no need to intervene in the allocation of network resources.

[0060] By calculating the difference between the total expected transmission rate demand and the predicted available network bandwidth as the supply-demand gap (G(t)), the supply and demand relationship of network resources at a specific point in time can be fully reflected. This process provides an important basis for the effective scheduling, monitoring and optimization of network resources, and ensures the stability and reliability of the network under high load conditions.

[0061] Figure 3 A flowchart illustrating a control method according to another embodiment of the present disclosure is shown schematically.

[0062] like Figure 3 As shown, based on the foregoing embodiments, generating a network resource allocation strategy based on the comparison results and the priority scores of the network resource scheduling of the one or more processes may include steps S310 and S320.

[0063] In step S310, based on the priority scores of network resource scheduling for one or more processes, the one or more processes are divided into multiple groups, and the multiple groups include at least the guarantee group with the highest priority score and the control group with the lowest priority score.

[0064] For example, all processes are sorted by priority score, arranged in descending order. This sorting ensures that high-priority processes are given priority in subsequent groupings. Based on the sorting results, the processes are divided into multiple groups, with the specific grouping rules as follows: The top 20% of processes are assigned to the guarantee group. These processes have the highest priority scores and are typically applications with extremely high network performance requirements. For example, video conferencing requires low latency and high bandwidth to ensure audio and video quality, and online office work requires a stable network connection to support real-time collaboration.

[0065] Watch Group: Processes in the middle 60% are classified into the watch group. These processes have moderate priority scores and are usually applications that have certain requirements for network performance but not as high as those in the guaranteed group. For example, social software needs a certain amount of bandwidth to support real-time message transmission, but its sensitivity to latency is relatively low. Online music needs a stable network connection to ensure sound quality, but its latency requirements are not as strict as those for video conferencing.

[0066] Control Group: Processes ranked in the bottom 20% are assigned to the control group. These processes have the lowest priority scores and are typically applications with low network resource requirements. For example, background updates can be performed when network load is low and real-time performance is not required, while broadcast loading has low bandwidth requirements and can be performed when network resources are sufficient.

[0067] By grouping processes into multiple groups based on priority scores, more granular network resource management can be achieved. This process not only ensures that the network needs of high-priority processes are met first, but also rationally regulates the resource usage of low-priority processes when resources are limited, thereby optimizing overall network performance and user experience.

[0068] In step S320, in response to determining that there is a supply-demand gap indicating insufficient bandwidth, available network bandwidth is preferentially allocated to processes in the guarantee group.

[0069] For example, when G(t) is greater than 0, it indicates insufficient bandwidth, and a bandwidth allocation response mechanism will be initiated. For instance, all processes in the protection group are identified from the divided process groups. Based on the available bandwidth (B(t)), the available bandwidth is first allocated to the processes in the protection group. After bandwidth allocation, network conditions and the actual performance of processes will continue to be monitored to ensure that the processes in the protection group can run smoothly. If it is found that the processes in the protection group still cannot meet their network requirements, the resource allocation strategy can be further adjusted. For example, the bandwidth allocation to the observation group and the control group can be temporarily reduced to free up more bandwidth for the processes in the protection group. The strategy of prioritizing the allocation of available network bandwidth to processes in the protection group can be widely applied in the following scenarios: ensuring stable network connections for participants during video conferences within enterprises to avoid delays and lag; ensuring the quality of real-time interaction between teachers and students on online education platforms to ensure the smooth transmission of teaching content; and ensuring high-quality audio and video communication between doctors and patients in telemedicine services to improve the effectiveness of medical services.

[0070] By responding to identified supply and demand gaps and prioritizing the allocation of available network bandwidth to processes within the support group, the efficiency of network resource utilization can be effectively improved, ensuring that applications with high network performance requirements receive the necessary support. This strategy not only optimizes the allocation of network resources but also enhances the user experience and ensures the stable operation of critical applications.

[0071] According to an embodiment of this disclosure, the control method further includes, in response to determining that there is a supply-demand gap indicating insufficient bandwidth, selecting one or more first processes in the control group as first candidate processes for network rate compression; and reducing the current first network rate of the first candidate processes by a first compression ratio to reduce the supply-demand gap.

[0072] For example, a control group is a group of processes divided according to their priority scores. It includes processes with lower network performance requirements, typically applications with low bandwidth demands. Processes in the control group can be reassigned when network resources are strained to free up bandwidth for higher-priority processes. For instance, one or more first-priority processes might be identified from the control group as primary candidates for network rate compression. Selection criteria could include the process's current bandwidth usage, priority score, and impact on network performance. The suitability of each process in the control group is evaluated, selecting those that are sensitive to bandwidth compression and can tolerate a certain reduction in network speed. For example, background update processes are typically insensitive to latency and are therefore more suitable for bandwidth compression.

[0073] In this example, the first compression ratio can be calculated based on the supply-demand gap ratio and the process control weight, such as first compression ratio = supply-demand gap ratio × process control weight.

[0074] The supply-demand gap ratio refers to the proportion of the supply-demand gap relative to total demand. The formula for calculation is:

[0075] This ratio reflects the degree of current network demand insufficiency; a higher value indicates a more severe bandwidth shortage. The process control weight is a value set based on process characteristics and importance, used to measure the process's adjustability during bandwidth compression. The weight can be set according to the process's priority, historical performance, and network usage patterns. For example, processes with higher weights may be more affected during compression.

[0076] Based on the calculated first compression ratio, the current network speed of the first candidate process will be adjusted. Specifically, the current network speed can be multiplied by (1 - the first compression ratio) to reduce the speed. For example, if the current network speed is 500 kbps and the first compression ratio is 20%, the new network speed will be:

[0077] After implementing network speed compression, the performance and network status of the process should be monitored in real time to ensure that the compression does not seriously affect its functionality. If it is found that the compressed network speed still cannot meet the supply and demand balance, further compression of the network speed of other processes or adjustment of the compression ratio can be considered.

[0078] By responding to the identified supply-demand gap, one or more first processes in the control group are selected as candidate processes for network rate compression. The first compression ratio is calculated based on the supply-demand gap ratio and the process control weight. This process can effectively optimize the allocation of network resources. It ensures that the network needs of high-priority processes are met, while reasonably controlling the resource usage of low-priority processes, thereby improving overall network performance and user experience.

[0079] According to an embodiment of this disclosure, the control method further includes, in response to reducing the current first network rate of the first candidate process by a first compression ratio, if the supply-demand gap still exists, selecting one or more first processes in the watch group as second candidate processes for network rate compression; reducing the current second network rate of the second candidate processes by a second compression ratio to further reduce the supply-demand gap, wherein the second compression ratio is less than the first compression ratio, and wherein the watch group includes processes with medium priority scores.

[0080] For example, in the first phase, network rate compression is implemented on the first candidate process in the control group, reducing its current network rate by a first compression ratio. If the supply-demand gap still exists after compression in this phase, further measures will be taken. For instance, one or more first processes are identified from the observation group as second candidate processes for network rate compression. The selection criteria for these processes may include current bandwidth usage, sensitivity to network performance, and priority scores. The second compression ratio should be lower than the first compression ratio to ensure a relatively small impact on the processes in the observation group. Based on the calculated second compression ratio, the current network rate of the second candidate processes will be adjusted. After implementing network rate compression, the performance and network status of the second candidate processes should be monitored in real time to ensure that compression does not seriously affect their functionality. If it is found that the compressed network rate still cannot meet the supply-demand balance, further consideration can be given to compressing the network rate of other processes or adjusting the compression ratio.

[0081] Following the initial network rate compression, if a supply-demand gap persists, one or more of the first processes in the observation group are designated as second candidate processes for network rate compression. Their network rates are further reduced using a second compression ratio, effectively optimizing network resource allocation. This process ensures that the network demands of high-priority processes are met while rationally controlling resource usage for medium-priority processes, thereby improving overall network performance and user experience.

[0082] Figure 4 A flowchart illustrating a control method according to another embodiment of the present disclosure is shown schematically.

[0083] like Figure 4 As shown, based on the foregoing embodiments, the plurality of time points may include a first time point and a second time point, wherein the second time point is later than the first time point; the method may also include steps S410-S430.

[0084] In step S410, in response to determining that there is a bandwidth shortage at a first time point.

[0085] In step S420, the first network rate allocation of one or more processes in the control group is reduced by a first compression ratio to reduce the supply-demand gap.

[0086] In step S430, in response to determining that there is sufficient bandwidth and no supply-demand gap at the second time point, the first network rate allocation of one or more processes in the control group is restored, but the total rate of all processes in the control group is limited to a predetermined proportion of the predicted available network bandwidth.

[0087] This method effectively manages network resources by dynamically monitoring and adjusting at multiple time points in response to bandwidth shortages and supply-demand gaps. First, it reduces the supply-demand gap by lowering the process rates in the control group. Second, when bandwidth is restored, the process rates are restored appropriately while ensuring the total rate does not exceed a predetermined proportion of available bandwidth. This method not only optimizes network resource utilization but also improves user experience and ensures the stable operation of critical applications.

[0088] For example, at the first time point: a bandwidth shortage (G(t)>0) is detected, requiring measures to optimize network resource allocation. At the second time point: later than the first time point, sufficient bandwidth is detected (G(t)<0), allowing the restoration of network rate allocation for processes in the control group.

[0089] For example, at the first time point, network conditions are monitored in real time to identify a bandwidth shortage. One or more processes are selected from the control group, and the network rate allocation for the selected processes is reduced according to the calculated first compression ratio. At the second time point, network conditions are monitored again. If sufficient bandwidth is detected at the second time point (i.e.,...) If the network rate allocation for processes in the control group is restored, the network rate of one or more processes in the control group can be restored to its original allocation value. This restoration process should ensure that the needs of all processes are met as much as possible when bandwidth is sufficient. Although the rate allocation is restored, the total rate of all processes in the control group should still be limited to a predetermined percentage of the predicted available network bandwidth (e.g., no more than 30% of the available network bandwidth). The purpose of setting this percentage is to avoid network congestion or insufficient bandwidth again.

[0090] If the sum of the recovery rates of processes in the control group exceeds the total rate limit, the rates of each process should be reduced proportionally to ensure that the total rate does not exceed the limit. For example, if the total rate after recovery is 600 kbps, and the total rate limit is 500 kbps, then the rate needs to be compressed to 500 kbps proportionally.

[0091] In some embodiments, the control method can be executed according to a preset period, such as every five minutes. First, the historical and real-time network usage characteristics of one or more processes and the current and historical network latency characteristics of one or more processes can be collected by the perception layer. Then, the historical and real-time network usage characteristics of one or more processes are processed by the first AI model of the modeling layer to obtain the expected network transmission rate demand requested by one or more processes running on the device within the upcoming time window T. Then, the current and historical network latency characteristics of one or more processes are processed by the second AI model to obtain the predicted available network bandwidth at one or more time points within the time window T. The difference between the total expected transmission rate demand at a specific time point within the time window T and the predicted available network bandwidth at the specific time point is calculated by the decision layer as the supply and demand gap. Based on the supply and demand gap and the priority score of the network resource scheduling of one or more processes, a network resource allocation strategy is generated and the network resource allocation strategy is sent to the execution layer.

[0092] refer to Figure 5 When the network resource scheduler in the execution layer receives a network request, it can schedule one or more processes to access network resources according to the network resource allocation strategy. In this way, through accurate prediction and intelligent scheduling strategies, the efficiency and reliability of network transmission can be significantly improved.

[0093] Figure 6 A schematic block diagram of an electronic device according to an embodiment of the present disclosure is shown.

[0094] like Figure 6 As shown, the electronic device 6 of this embodiment includes a processor 601, a memory 602, and a computer program 603 stored in the memory 602 and executable on the processor 601. When the processor 601 executes the computer program 603, it implements the steps in the various method embodiments described above. Alternatively, when the processor 601 executes the computer program 603, it implements the functions of each module in the various device embodiments described above.

[0095] Electronic device 6 can be a desktop computer, laptop, handheld computer, cloud server, or other electronic device. Electronic device 6 may include, but is not limited to, processor 601 and memory 602. Those skilled in the art will understand that... Figure 6 This is merely an example of electronic device 6 and does not constitute a limitation on electronic device 6. It may include more or fewer components than shown, or different components.

[0096] The processor 601 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0097] The memory 602 can be an internal storage unit of the electronic device 6, such as a hard disk or RAM of the electronic device 6. The memory 602 can also be an external storage device of the electronic device 6, such as a plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card, Flash Card, etc., equipped on the electronic device 6. The memory 602 can also include both internal and external storage units of the electronic device 6. The memory 602 is used to store computer programs and other programs and data required by the electronic device.

[0098] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0099] The present invention provides a computer program product, which stores a computer program, and when the computer program is executed by a processor, it implements the steps in the above method embodiments.

[0100] If the integrated module is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program may include computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium may include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium may be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0101] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0102] The units or modules described in the embodiments of this disclosure can be implemented in software or programmable hardware. The described units or modules can also be located in a processor, and the names of these units or modules do not necessarily constitute a limitation on the unit or module itself.

[0103] In another aspect, this disclosure also provides a computer-readable storage medium, which may be a computer-readable storage medium included in the electronic device or computer system described above; or it may be a standalone computer-readable storage medium not assembled into a device. The computer-readable storage medium stores one or more programs, which are used by one or more processors to execute the methods of the embodiments of this disclosure.

[0104] The above description is merely a preferred embodiment of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the inventive concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features disclosed in this disclosure that have similar functions.

Claims

1. A control method, comprising: Based on the first prediction model, determine the expected network transmission rate demand requested by one or more processes running on the device within the upcoming time window T. Based on the second prediction model, the predicted available network bandwidth at one or more time points within the time window T is determined; Based on multiple preset network resource scheduling parameters, the priority score for network resource scheduling of the one or more processes is determined; The expected network transmission rate requirement is compared at least partially with the predicted available network bandwidth, and a network resource allocation strategy is generated based on the comparison results and the priority scores of the network resource scheduling of the one or more processes. According to the network resource allocation strategy, the access of one or more processes to network resources is scheduled.

2. The control method according to claim 1, at least partially comparing the expected network transmission rate demand with the predicted available network bandwidth, includes: The total expected transmission rate requirement is generated at each time point within the time window T based on the expected network transmission rate of each of the one or more processes. For at least one specific time point within the time window T, the difference between the total expected transmission rate demand and the predicted available network bandwidth at that specific time point is calculated as the supply-demand gap.

3. The control method according to claim 2, wherein the plurality of preset network resource scheduling parameters include at least two of the following parameters: Process priority categories are preset by the user or the system. The congestion sensitivity coefficient, output by the first prediction model, is used to indicate the process's sensitivity to network congestion. An indicator that shows whether the process is in a critical data transfer window; The network transfer rate currently allocated to the process.

4. The control method according to claim 2, generating a network resource allocation strategy based on the comparison result and the priority scores of the network resource scheduling of the one or more processes, comprising: Based on the priority scores of the network resource scheduling of the one or more processes, the one or more processes are divided into multiple groups, and the multiple groups include at least the guarantee group with the highest priority score and the control group with the lowest priority score. In response to the determination of a supply-demand gap indicating insufficient bandwidth, available network bandwidth is preferentially allocated to processes in the guarantee group.

5. The control method according to claim 4 further includes: In response to determining that there is a supply-demand gap indicating insufficient bandwidth, one or more first processes in the control group are selected as first candidate processes for network rate compression; The supply-demand gap is reduced by lowering the current first network rate of the first candidate process by a first compression ratio.

6. The control method according to claim 5 further includes: In response to reducing the current first network rate of the first candidate process by a first compression ratio, if the supply-demand gap still exists, one or more first processes in the watch group are used as second candidate processes for network rate compression. The supply-demand gap is further reduced by lowering the current second network rate of the second candidate process by a second compression ratio, wherein the second compression ratio is less than the first compression ratio, and wherein the watch group contains processes with medium priority scores.

7. The control method according to claim 4, wherein the plurality of time points includes a first time point and a second time point, the second time point being later than the first time point; the method further includes: In response to the determination that there is a bandwidth shortage at the first point in time; The supply-demand gap is reduced by decreasing the first network rate allocation of one or more processes in the control group through a first compression ratio. In response to determining that there is sufficient bandwidth and no supply-demand gap at a second time point, the first network rate allocation for one or more processes in the control group is restored, but the total rate of all processes in the control group is limited to a predetermined proportion of the predicted available network bandwidth.

8. The control method according to claim 1, wherein the first prediction model is a first AI model, and the first AI model predicts the expected network rate demand based on learning the historical and real-time network usage characteristics of the one or more processes; The second prediction model is a second AI model, which predicts future network bandwidth availability based on learning current and historical network latency characteristics.

9. The control method according to claim 8, wherein determining the expected network transmission rate demand requested by one or more processes running on the device within the upcoming time window T, based on the first prediction model, includes: The first AI model processes the historical and real-time network usage characteristics of the one or more processes to obtain the expected network transmission rate requirements requested by the one or more processes running on the device within the upcoming time window T; the historical and real-time network usage characteristics include at least any two or more of the following: priority identifier, historical network rate, historical connection duration, size of historical transmitted data packets, identifier of the currently accessed port, and size of the currently transmitted data packets; Based on the second prediction model, determining the predicted available network bandwidth at one or more time points within the time window T includes: The second AI model processes the current and historical network latency characteristics of the one or more processes to obtain the predicted available network bandwidth at one or more time points within the time window T; the current and historical network latency characteristics include at least any two or more of the following: available bandwidth, bandwidth fluctuation range, network type, historical bandwidth usage data, historical bandwidth usage scenarios, current process latency, current process congestion, and current process bandwidth usage.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the method according to any one of claims 1 to 9.

Images