Network resource scheduling method, device and equipment

Abstract

The embodiment of the invention discloses a network resource scheduling method, device and equipment. According to the scheme, the method comprises the following steps: receiving local data transmitted by a client, and obtaining content distribution network state data transmitted by a content distribution network node; analyzing according to the local data and the content distribution network state data, and determining a network resource state corresponding to the client; and generating a corresponding network resource control strategy according to the network resource state, and feeding back the network resource control strategy to the client, so that the client executes the network resource control strategy.

Description

Technical Field

[0001] This specification relates to the field of Internet technology, and in particular to methods, apparatus and equipment for network resource scheduling. Background Technology

[0002] With the development of digitalization and mobile internet, the distribution of high-bandwidth content such as video and audio, and critical business scenarios, have become core components of application services. Users are placing high demands on smooth, high-quality experiences and high service availability.

[0003] In traditional solutions, before service deployment, operations personnel or R&D teams typically pre-configure scheduling policies in the cloud based on the business type (e.g., payment, live streaming, video-on-demand). Based on these cloud-preset rules, network resource scheduling is executed through static scheduling policies to statically deliver content. Furthermore, with surges in business traffic or operational activities, sudden increases in business requests can lead to bursts of traffic on the Content Delivery Network (CDN), potentially incurring additional CDN bandwidth costs. In such cases, traditional solutions usually employ global rate limiting, setting global bandwidth thresholds through static scheduling policies to ensure all users share the same set of traffic control rules.

[0004] Meanwhile, in scenarios such as node failure, link anomaly, or domain name blocking, cloud service failover in CDN or backend cloud systems relies on manual intervention or DNS re-resolution.

[0005] Therefore, a more efficient and personalized network resource scheduling solution is needed to cater to different users. Summary of the Invention

[0006] This specification provides one or more embodiments of a network resource scheduling method, apparatus, device, and storage medium to solve the following technical problem: the need for a more efficient and personalized network resource scheduling scheme for different users.

[0007] To solve the above-mentioned technical problems, one or more embodiments of this specification are implemented as follows: This specification provides a network resource scheduling method according to one or more embodiments, including: Receive local data transmitted by the client, and obtain content delivery network status data transmitted by content delivery network nodes; Based on the analysis of the local data and the content delivery network status data, the network resource status corresponding to the client is determined. Based on the network resource status, a corresponding network resource control policy is generated, and the network resource control policy is fed back to the client so that the client can execute the network resource control policy.

[0008] This specification provides a network resource scheduling device according to one or more embodiments, comprising: The data receiving module receives local data transmitted by the client and obtains content delivery network status data transmitted by content delivery network nodes. The status confirmation module analyzes the local data and the content delivery network status data to determine the network resource status corresponding to the client. The policy execution module generates a corresponding network resource control policy based on the network resource status and feeds the network resource control policy back to the client so that the client executes the network resource control policy.

[0009] This specification provides one or more embodiments of a network resource scheduling device, comprising: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to: Receive local data transmitted by the client, and obtain content delivery network status data transmitted by content delivery network nodes; Based on the analysis of the local data and the content delivery network status data, the network resource status corresponding to the client is determined. Based on the network resource status, a corresponding network resource control policy is generated, and the network resource control policy is fed back to the client so that the client can execute the network resource control policy.

[0010] This specification provides one or more embodiments of a non-volatile computer storage medium storing computer-executable instructions, wherein the computer-executable instructions are configured as follows: Receive local data transmitted by the client, and obtain content delivery network status data transmitted by content delivery network nodes; Based on the analysis of the local data and the content delivery network status data, the network resource status corresponding to the client is determined. Based on the network resource status, a corresponding network resource control policy is generated, and the network resource control policy is fed back to the client so that the client can execute the network resource control policy.

[0011] The above-described at least one technical solution adopted in one or more embodiments of this specification can achieve the following beneficial effects: 1. Achieve precise resource scheduling while ensuring both cost control and user experience: By integrating real-time CDN operation and maintenance data with multi-dimensional local client data, distribution strategies can be dynamically formulated. Under cost pressure, the system can implement refined cost reduction or degradation for low-priority users or scenarios, while ensuring a high-quality experience for high-value users, effectively balancing cost and experience goals.

[0012] 2. Ensure high service availability and business continuity: Relying on the cloud control system for second-level monitoring of nodes and links, a switchover command can be issued in real time once an anomaly is detected, and the client automatically switches to the backup node. Compared with traditional manual or DNS switching, this significantly shortens fault recovery time and improves business robustness and user retention.

[0013] 3. Achieve personalized content distribution for each user: By integrating multi-dimensional data such as user network, device, package, and preferences, the cloud control center can dynamically issue differentiated content strategies to provide users with a better experience in different scenarios (e.g., high definition on strong networks and smooth playback on weak networks), thereby improving user satisfaction while avoiding resource waste.

[0014] 4. Enhance system adaptability and risk resilience: Through a closed-loop mechanism of policy distribution and effect data feedback, the system can continuously learn and optimize. Combined with client-side event-driven mechanisms and periodic data reporting, it enables early identification and local control of abnormal traffic, preventing the spread of risks and improving the overall system resilience. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A flowchart illustrating a network resource scheduling method provided in one or more embodiments of this specification; Figure 2 A schematic diagram of the application architecture of a network resource scheduling method in an application scenario, provided for one or more embodiments of this specification; Figure 3 A flowchart of a network resource cost reduction process in an application scenario, provided for one or more embodiments of this specification; Figure 4 A schematic diagram of a CDN node switching process in an application scenario, provided for one or more embodiments of this specification; Figure 5A schematic diagram of a client-side personalized network resource allocation process in an application scenario, provided for one or more embodiments of this specification; Figure 6 A schematic diagram of the structure of a network resource scheduling device provided for one or more embodiments of this specification; Figure 7 This is a schematic diagram of the structure of a network resource scheduling device provided for one or more embodiments of this specification. Detailed Implementation

[0017] This specification provides network resource scheduling methods, apparatus, devices, and storage media through its embodiments.

[0018] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this application.

[0019] Figure 1 This diagram illustrates a network resource scheduling method provided in one or more embodiments of this specification. The method can be applied to various business domains, such as internet video live streaming, internet finance, e-commerce, instant messaging, gaming, and government services. The process can be executed by computing devices in the corresponding domain (e.g., servers or smart mobile terminals for live streaming services). Certain input parameters or intermediate results in the process can be manually adjusted to improve accuracy.

[0020] High-bandwidth content distribution, such as video and audio, and critical business scenarios (e.g., payment, live streaming, video-on-demand) have become core components of application services. Users are demanding smooth, high-quality experiences and high service availability, but traditional solutions for network resource scheduling still face numerous technical challenges in actual distribution processes. 1. The Conflict Between Cost Control and User Experience Assurance: Surges in business traffic or operational activities can lead to sudden increases in CDN traffic, potentially resulting in additional CDN bandwidth costs. Traditional solutions typically employ global rate limiting, setting global bandwidth thresholds through static scheduling strategies. This forces all users to share the same traffic control rules, leading to forced degradation of all users during peak periods and difficulty in differentiating between different types of services. Furthermore, traditional solutions distribute CDN status, user profiles, service quality, and experience data across different systems, lacking unified aggregation and intelligent analysis, further complicating the differentiation of traffic control rules for different users.

[0021] 2. Efficiency Bottleneck of Failover: In scenarios such as node failures, link anomalies, or domain name blocking, CDNs or backend cloud systems can easily lead to prolonged service unavailability. Traditional solutions rely on manual intervention or DNS re-resolution for cloud service failover, resulting in delays in detection and switching processes, poor business continuity, and failure detection delays typically exceeding one minute, with failover taking even more than five minutes—unacceptable in critical scenarios (such as payment scenarios). Simultaneously, clients struggle to detect and automatically switch to backup resources in a timely manner, leading to service interruptions.

[0022] 3. Resource allocation is disconnected from user environment, making it difficult to achieve precise personalized experiences: Different users have significantly different network environments (e.g., 4G, 5G, Wi-Fi), remaining data allowances, terminal device performance, and user preferences. If the system statically distributes content, it is difficult to personalize the user experience, which may lead to problems such as buffering and poor image quality, affecting service satisfaction. For example, based on statically distributed content, 5G users may be assigned 480P videos due to global bandwidth policies, or users in weak network environments may be forced to watch high-definition content, resulting in severe buffering.

[0023] 4. Delayed response and simplistic strategy: In traditional distribution methods, the cloud and the client lack timely and unified data collaboration and dynamic decision-making. They can only be passively adjusted or manually prepared. The strategy is configured manually on a regular basis or synchronized periodically. It is difficult to automatically optimize and schedule according to real-time scenarios. The technical response is delayed, resulting in both cost control and user experience being compromised. This leads to both resource waste and loss of user experience.

[0024] Based on this, the following was proposed: Figure 1 The network resource scheduling method shown is as follows: Figure 1 The process may include the following steps: S102: Receive local data transmitted by the client and obtain content delivery network status data transmitted by the content delivery network node.

[0025] Figure 2This diagram illustrates the application architecture of a network resource scheduling method for one or more embodiments of this specification. The cloud control center, also known as the control center, is typically located on a server corresponding to the client, or it may function as a separate server. Generally, the cloud control center is responsible for real-time collection and aggregation of metrics such as bandwidth, latency, and packet loss rate from the Internet Content Delivery Network (CDN). It also receives and aggregates multi-dimensional data reported by various clients, including user characteristics, network status, available bandwidth, playback buffering, and payment results. Through intelligent decision-making algorithms, it comprehensively analyzes this data and automatically generates network resource control strategies (e.g., distribution degradation, priority enhancement, and failover).

[0026] Specifically, local data includes two types: static data used to describe user characteristics (referred to as static user data) and dynamic data used to describe the business execution process (referred to as dynamic business data). The client terminal periodically or event-triggeredly collects local data and uploads it to the cloud control system.

[0027] The system receives static user data transmitted from the client, which describes user characteristics, and builds a user profile based on this data. Static user data may include device level, network type, and user geographic location information. Device level refers to the tiers based on terminal performance, including high-end, mid-range, and low-end devices. Network type refers to the type of network the user accesses, which may include 5G, cellular networks, and Wi-Fi. User geographic location information refers to the user's location; the accuracy of this information can be adjusted based on requirements.

[0028] The "static" in static user data does not mean it remains unchanged forever, but rather data that has a high probability of remaining unchanged within a short period or during a single transaction. However, in practice, users may switch network types or move their geographical location within a short period or during a single transaction. In such cases, the static user data will be updated based on event triggers.

[0029] Generally speaking, when a client first establishes a connection with the cloud control center based on business (for example, when a user registers through the client), a user profile can be created for that client and updated as subsequent events are triggered (including new business transactions, events during the business transaction process, etc.) or periodically.

[0030] The system receives dynamic business data from the client in real time, which describes the business execution process. Dynamic business data can include user payment success or failure, playback stuttering, page loading white screen, etc. Compared with static user data, it changes more frequently and has higher real-time requirements. However, it does not occur in every business execution process, so it needs to be transmitted and confirmed by the client in real time.

[0031] The client, also known as the terminal application, is installed on the user's terminal. During the user's business transactions, it can monitor its own network environment (e.g., network type such as Wi-Fi, 4G, 5G, etc., and whether it is in a weak network environment), data plan, terminal hardware capabilities, and content experience (e.g., buffering rate, payment success rate). Simultaneously, the client can automatically adjust download behavior, request resolution, content quality, etc., according to the network resource control policies issued by the cloud control center. It can also execute actions such as failover of the cloud control center and content distribution network nodes, and based on the state before adjustment, the local execution results after adjustment, and the latest experience data, it provides real-time feedback to the cloud control center, supporting intelligent system summarization and self-learning.

[0032] A Content Delivery Network (CDN) is a distributed network that caches content from an origin server using edge nodes deployed in various locations (referred to as CDN nodes in this article). These nodes can be located in the cloud, origin server data centers, etc., and are responsible for distributing the actual content (including video, audio, images, etc.). CDNs support multi-level resource storage and dynamic node routing, allowing users to access data from the nearest node, thereby improving access speed, reducing pressure on the origin server, and ensuring distribution reliability. Simultaneously, CDNs provide an operation and maintenance monitoring data interface, allowing the cloud control system to collect relevant CDN status data globally. This CDN status data can include bandwidth data, packet loss rate, latency, etc.

[0033] S104: Analyze the local data and the content delivery network status data to determine the network resource status corresponding to the client.

[0034] Network resource status is a quantitative assessment result based on a comprehensive analysis of client-side local data and content delivery network status data. It describes the overall quality and availability of the service path from the client to its connected CDN nodes and can be represented by multiple dimensions.

[0035] Specifically, content delivery network (CDN) status data can include bandwidth utilization, packet loss rate, latency, and error rate. Bandwidth utilization rate (Utilization) refers to the ratio of the actual amount of data transmitted per unit time to the theoretical maximum transmission capacity of a network channel, usually expressed as a percentage, reflecting the traffic saturation level of a link or node. Packet loss rate (Loss) refers to the ratio of the number of data packets lost or corrupted during data transmission that fail to reach the destination to the total number of packets sent, usually expressed as a percentage, reflecting the reliability or congestion level of the network link. Latency (Delay) refers to the one-way or round-trip time required for a data packet to travel from the sender to the receiver, usually measured in milliseconds, reflecting the speed of network response. Error rate (Error) refers to the proportion of bit errors or frame errors that occur during data transmission at the physical or data link layer. It directly measures the inherent reliability of the channel or transmission medium and is usually caused by hardware defects, electromagnetic interference, etc.

[0036] like Figure 2 As shown, during the data analysis process, local data and content delivery network (CDN) status data are uniformly converted into a standardized format through data cleaning and formatting. Simultaneously, data aggregation and statistical calculations are performed to summarize user profiles, business experience metrics, CDN node performance parameters, and other data from multiple dimensions. Anomaly detection is also conducted to identify any abnormal data, facilitating subsequent business evaluation and confirmation of the current network resource status.

[0037] During business evaluation, a network failure index can be determined based on multiple preset network status dimensions and content distribution network status data. A quantified, unified network failure index represents these multiple network status dimensions simultaneously, providing a faster and clearer reflection of the current network status. Multiple network status dimensions can be selected, and corresponding weights can be assigned to each dimension, resulting in a weighted sum to obtain the network failure index. For example, if the network status dimensions are packet loss rate, error rate, and latency, and weights are set, the network failure index = 0.4 * error rate + 0.3 * packet loss rate + 0.3 * latency, where latency is pre-normalized.

[0038] Based on the network failure index and dynamic service data, the network resource status corresponding to the client is obtained. The dynamic service data reflects the status of the user's current service, such as whether there are any issues like lag or blank screens. The network failure index reflects the overall network situation. Combining these two metrics allows for a comprehensive assessment of the client's network resource status from both the CDN and client perspectives.

[0039] S106: Generate a corresponding network resource control policy based on the network resource status, and feed the network resource control policy back to the client so that the client executes the network resource control policy.

[0040] For different network resource states, the cloud control center generates corresponding network resource control policies, which are then distributed to the client. For example... Figure 2 As shown, the multi-objective decision engine has built-in alarms and a rule base, used to trigger anomaly warnings and match preset policy templates, respectively. The alarms can generate alarm signals based on key indicators in the network resource status, while the rule base stores policy templates for different scenarios, such as bandwidth emergency control templates, fault rapid switching templates, and user experience optimization templates.

[0041] A policy distribution and command synchronization module is set up in the cloud control center. After the cloud control system generates control policies, it can quickly and securely generate a policy priority queue, indicating the network resource control policies and their respective priorities for each client. It also generates encrypted control commands and distributes the policies to each client through a dedicated command channel. It supports real-time policy synchronization and acknowledgments for multiple clients and multiple scenarios.

[0042] Upon receiving a control command, the client decrypts it to obtain the corresponding network resource control policy. During execution, it also provides feedback on static user data and dynamic business data.

[0043] Figure 3 This specification provides a flowchart of a network resource cost degradation process in one or more embodiments. The process involves determining that bandwidth utilization is higher than a first preset threshold based on content delivery network status data, identifying the presence of preset service anomalies based on dynamic business data, and determining that the network fault index is lower than a second preset threshold. A bandwidth utilization higher than the first preset threshold (e.g., set to 0.9) indicates that the bandwidth resources of the current CDN node are nearing or have reached saturation, potentially making it difficult to meet the normal request needs of all users. The presence of preset service anomalies in the dynamic business data (e.g., playback stuttering, payment timeouts) indicates that the user's current service experience has been negatively impacted. A network fault index lower than the second preset threshold (e.g., set to 0.8) suggests that the current network link itself (e.g., packet loss, latency) has not yet experienced a serious, systemic failure. Based on these factors, the service anomaly at this point is more likely due to bandwidth resource constraints.

[0044] Corresponding network resource control policies are generated to reduce network resource consumption by clients. Based on this network resource status, the network resource control policies generated by the cloud control center focus on cost control of network resources, ensuring basic business operations for most clients by reducing their network resource consumption. For example, in video playback services, video resolution is reduced, and data transmission rates are delayed in non-urgent data synchronization scenarios, thereby alleviating network resource pressure.

[0045] After executing the network resource control policy, the client returns an execution result acknowledgment (ACK), reporting the specific results of this execution (such as whether the execution was successful or failed, the video playback bitrate after the downgrade, the target node to be switched, etc.), which is then monitored by the cloud control center.

[0046] Furthermore, in some special business scenarios, if this cost downgrade operation is performed during business execution, it may easily lead to abnormal execution of the business.

[0047] Based on this, when generating the corresponding network resource control policy, the service type corresponding to the dynamic service data is determined. Service types may include video playback services, live streaming services, payment services, security verification services, etc.

[0048] If the business type is a preset type, the current business stage is determined based on the preceding business actions already executed in the dynamic business data. Preset types may include payment services, security verification services, etc. This type of business is more sensitive to latency but has low bandwidth requirements. If a rate reduction or throttling policy is triggered at this moment due to high bandwidth utilization, it may increase additional processing latency or cause retransmissions, which may easily cause the entire critical transaction to fail (e.g., payment timeout, login failure), resulting in a degraded user experience.

[0049] Based on this, the current business stage is determined by the preceding business actions, thereby determining whether the user is currently in a critical operation stage. For example, in security verification, the preceding business action could be that the user completes image CAPTCHA recognition and submits a verification request, or that the user has already applied for a verification code. In this case, the current business stage is the verification information upload stage, which is a preset stage. In payment, the preceding business action could be that the user enters the payment page. In this case, the current business stage is the payment request submission stage, which is also a preset stage.

[0050] If the business phase belongs to a preset phase, a corresponding network resource control policy is generated. This policy maintains the client's network resource consumption during the execution of the business phase and reduces it after the phase is completed. If both the current business type and phase belong to a preset type, cost degradation is not immediately implemented. Instead, the current network resource consumption is maintained until the current phase is completed. This ensures the smooth execution of critical business processes and prevents core operations from failing due to resource policy adjustments. For example, when a user is detected in the payment request submission phase, even if the CDN node bandwidth utilization is high, the cloud control center will maintain the current network resource configuration to prioritize the transmission and processing of payment data packets. After payment is completed, the policy will be dynamically adjusted to reduce resource consumption based on real-time network conditions.

[0051] Figure 4 This specification provides a schematic diagram of a CDN node switching process in one or more embodiments. If the network failure index is determined to be higher than a second preset threshold, a backup content distribution network node is determined based on the client's current primary content distribution network node. A network failure index higher than the second preset threshold indicates that the current network link itself may have experienced a serious, systemic failure; therefore, simply reducing network resource costs is insufficient to solve the client's problems.

[0052] At this point, the backup content delivery network (CDN) node corresponding to the client is determined. The primary CDN node refers to the one currently being used by the client and primarily responsible for content distribution, while the backup CDN node is the one that can replace the primary node and provide services to the client when the primary node fails or performs poorly.

[0053] In its daily operations, the cloud control center maintains a list of CDN node health statuses and a mapping relationship between backup nodes. This list is dynamically updated based on real-time collected CDN status data (such as bandwidth, latency, packet loss rate, etc.). When the network failure index of the primary CDN node is detected to be higher than a second preset threshold, the cloud control center matches the client with a backup CDN node from the backup node mapping relationship that is geographically close, has better network conditions, and is currently under lower load.

[0054] Based on dynamic business data, corresponding business data is pre-cached to backup content distribution network nodes. Business data refers to the specific content data that the client is currently accessing or about to access, such as video files, audio clips, and web resource packages. To ensure seamless service continuity during failover, after determining the backup CDN node, the cloud control center immediately pre-caches the critical business data required by the client to the backup CDN node based on the client's dynamic business data (such as current playback progress, loaded resource identifiers, user request queues, etc.). For example, if a client is watching a video and is currently at the 15-minute mark, the cloud control center will instruct the backup node to prioritize caching the critical segments of the video after the 15-minute mark, thereby pre-warming up the backup resources.

[0055] A corresponding network resource control policy is generated to enable clients to switch to backup content distribution network nodes based on this policy. After the cloud control center sends the network resource control policy to the client, the client's Software Development Kit (SDK) executes the corresponding callback function (e.g., onSwitchStart) to notify the client to initiate the CDN node switching process. Once the client successfully establishes a new connection with the backup CDN node, it notifies the client of the established connection via the corresponding callback function (e.g., onSwitchDone), allowing normal interaction to resume and achieving a smooth connection with the backup CDN node.

[0056] Furthermore, in some large-scale events or breaking news scenarios, most users receive highly similar trending content, which may cause anomalies in the main content distribution network node. If a large number of clients escape from the main content distribution network node at the same time, it may instantly overwhelm the backup content distribution network node, resulting in congestion and transfer.

[0057] Based on this, if the number of clients that need to switch to the backup content distribution network node is higher than the preset number, it means that a large number of clients need to switch nodes in the current business activities. If the cloud control center allocates the switching strategy for the clients, a rough strategy calculation is likely to cause secondary congestion caused by the switching, while a precise strategy calculation requires a lot of computing power due to the large number of clients. Therefore, an inquiry command is generated.

[0058] The query command is sent to each client, enabling each client to perform local calculations, determine the corresponding switching benefit based on its local data, and then determine its own switching strategy based on the switching benefit. This query command transforms resource allocation from a centralized optimization problem into a distributed game problem. By utilizing the local computation of the clients, it avoids secondary congestion caused by herding behavior, resulting in a more robust overall system.

[0059] Switching benefits refer to the gains a client receives when switching from the current primary content delivery network node to a backup content delivery network node. The goal is to allow each client to selfishly calculate a value that maximizes its own interests (including experience and cost), and use this as the basis for generating switching strategies. It dynamically adjusts weights based on user profiles and local data, thereby achieving differentiated decision-making for the group.

[0060] The switching strategy includes whether to switch and the switching target. When multiple backup content distribution network nodes exist, after selecting whether to switch, the corresponding switching target must also be selected. The switching target can be determined based on factors such as geographical location. Additionally, a switching delay can be set based on requirements to determine the switching time.

[0061] The system receives the switching strategies returned by each client and combines them to obtain a set of switching strategies. This is achieved by adding the corresponding identifier of each client to the switching strategies of each client and then combining them.

[0062] Based on the number of switching strategies of each type, some switching strategies in the switching strategy set are adjusted. This adjustment is a fine-tuning process. For example, when there are multiple backup content distribution network nodes, if too many clients switch to one of the backup content distribution network nodes, the cloud control center can balance the number of clients that have switched to that backup content distribution network node and adjust them to switch to other backup content distribution network nodes.

[0063] Generate corresponding network resource control policies so that each client can execute its own switching policy based on the network resource control policies, such as determining whether to switch itself and the backup content distribution network node of the target to switch to.

[0064] Furthermore, when determining the switching benefits, the corresponding switching benefits are determined based on multiple preset benefit dimensions and local data. These benefit dimensions include the experience gain estimation dimension and the switching cost estimation dimension.

[0065] The experience gain estimation dimension represents the potential experience improvement brought about by switching to a backup content delivery network node. It can include multiple first sub-dimensions, each determined based on the network status dimension, and the weight of each first sub-dimension is determined by the business type corresponding to the dynamic business data. For example, the first sub-dimensions include latency gain, packet loss gain, and bandwidth gain, which are calculated using the latency, packet loss rate, and available bandwidth of the current primary content delivery network node and each backup content delivery network node, respectively. The relevant data of the primary content delivery network node can be obtained through interaction between the client and the current content delivery network node, while the relevant data of the backup content delivery network node can be obtained based on the interaction between the cloud control center and each backup content delivery network node, and with the issuance of query commands.

[0066] The estimated switching cost includes several second sub-dimensions, each determined based on static user data and business type. For example, the second sub-dimensions may include switching latency penalties, failure risk penalties, and resource consumption penalties. Switching latency penalties and failure risk penalties are determined based on the current business type and average switching time. For business types with high real-time requirements (e.g., real-time communication services), switching latency penalties are higher; while for newly launched business types, their stability is uncertain and may experience business anomalies during node switching, hence their failure risk penalties are higher. Resource consumption penalties represent the additional power and data consumed in establishing new connections, re-handshaking, and re-buffering. These are usually fixed values, but need to be increased for clients with low power or low data usage.

[0067] Based on this, the final switching benefit is obtained by combining the experience gain prediction dimension and the switching cost prediction dimension through weighted summation. When the switching benefit is higher than the preset value, a switching node is selected, and the backup content distribution network node with the highest switching benefit is selected as the target node for this switching.

[0068] Figure 5This specification provides a schematic diagram of a personalized network resource allocation process for a client in one or more embodiments, illustrating an application scenario. Based on user profiles and the service type corresponding to dynamic service data, the service level of the client is determined. The service level represents the priority or importance of a user during service usage; a higher service level requires more network resources. It is typically determined by combining static user data from multiple dimensions, such as historical consumption behavior, membership level, device tier, and network type, with the service type corresponding to dynamic service data. For example, paid members, users using high-end devices, or users in 5G network environments may be classified as having a higher service level; while ordinary free users, users using low-end devices, or users in weak network environments may be classified as having a general or lower service level. Furthermore, clients performing important services such as payment or medical consultation may have their service level increased, while clients performing ordinary services such as live video streaming may have their service level decreased. Determining the service level provides a basis for formulating subsequent differentiated resource scheduling strategies, ensuring that limited network resources prioritize the experience of high-level services.

[0069] Obtain the network failure index and bandwidth utilization rate for each content delivery network (CDN) node. The network failure index measures the link health of each CDN node, while the bandwidth utilization rate reflects the current traffic load level of the node. By comparing these two core indicators of different CDN nodes, the cloud control center selects CDN nodes with low network failure indices and reasonable bandwidth utilization rates as healthy CDN nodes.

[0070] Corresponding network resource control policies are generated to allocate appropriate content delivery network (CDN) nodes to clients based on their service levels, and to assign appropriate network resource usage patterns to clients based on network failure index and bandwidth utilization. When allocating CDN nodes, healthy CDN nodes are allocated to clients with higher service levels in priority order according to service level.

[0071] Furthermore, during the interaction between the client and CDN nodes, network resource usage modes are allocated based on the current content delivery network status data (including network failure index and bandwidth utilization). For example, a low network failure index and low bandwidth utilization indicate that the current CDN bandwidth is abundant, and the user's network type is 5G with sufficient traffic. In this case, a high-consumption network resource usage mode can be selected, such as choosing high-definition video for the user in live video streaming and automatically loading more images during business execution. Conversely, a high network failure index or high bandwidth utilization, in addition to switching CDN nodes for the user, a low-consumption network resource usage mode can be allocated, such as choosing standard-definition video for the user in live video streaming and reducing the automatic loading of low-value images during business execution.

[0072] Meanwhile, the client side can perform feedback loop, feeding back relevant static user data and dynamic business data during the business process to the cloud control center, thereby facilitating the cloud control center to update user profiles and network resource control strategy generation methods.

[0073] 1. Achieve precise resource scheduling while ensuring both cost control and user experience: By integrating real-time CDN operation and maintenance data with multi-dimensional local client data, distribution strategies can be dynamically formulated. Under cost pressure, the system can implement refined cost reduction or degradation for low-priority users or scenarios, while ensuring a high-quality experience for high-value users, effectively balancing cost and experience goals.

[0074] 2. Ensure high service availability and business continuity: Relying on the cloud control system for second-level monitoring of nodes and links, a switchover command can be issued in real time once an anomaly is detected, and the client automatically switches to the backup node. Compared with traditional manual or DNS switching, this significantly shortens fault recovery time and improves business robustness and user retention.

[0075] 3. Achieve personalized content distribution for each user: By integrating multi-dimensional data such as user network, device, package, and preferences, the cloud control center can dynamically issue differentiated content strategies to provide users with a better experience in different scenarios (e.g., high definition on strong networks and smooth playback on weak networks), thereby improving user satisfaction while avoiding resource waste.

[0076] 4. Enhance system adaptability and risk resilience: Through a closed-loop mechanism of policy distribution and effect data feedback, the system can continuously learn and optimize. Combined with client-side event-driven mechanisms and periodic data reporting, it enables early identification and local control of abnormal traffic, preventing the spread of risks and improving the overall system resilience.

[0077] In one or more embodiments of this specification, when determining content distribution network status data, in addition to directly interacting with CDN nodes to obtain the current content distribution network status data for each network status dimension, the content distribution network status data for future moments can also be estimated through prediction, thereby facilitating advance strategy formulation.

[0078] Based on multiple preset network state dimensions, state curves corresponding to each network state dimension are fitted using content delivery network state data. These network state dimensions include, but are not limited to, key indicators such as bandwidth, latency, packet loss rate, and jitter. The cloud control center first cleans and preprocesses the historically collected content delivery network state data for each network state dimension, removing outliers and noisy data to ensure data accuracy and usability. Then, for each network state dimension, time series analysis methods are used to fit the historical data to obtain the state curve of that dimension over time. For example, models such as Autoregressive Integrated Moving Average (ARIMA) and Long Short-Term Memory (LSTM) networks are used. By fitting the state curves, the historical change characteristics of each network state dimension can be intuitively observed, laying the foundation for subsequent future state predictions.

[0079] Based on state curves, predictions are made for the corresponding network state dimensions. If an ARIMA model is used, the trend, seasonal, and stochastic components of the state curve can be decomposed and predicted using model parameters (e.g., autoregressive order, differencing order, moving average order). If an LSTM network is used, a multi-layer neural network can learn long-term dependencies in historical state curves to output predicted values ​​for future moments. By predicting future network state data, the cloud control center can identify potential risks such as node overload and link congestion in advance, providing data support for developing proactive resource scheduling strategies (e.g., preheating backup nodes and adjusting resource allocation ratios), further improving the initiative and effectiveness of network resource scheduling.

[0080] Based on the same idea, one or more embodiments of this specification also provide apparatus and devices corresponding to the above methods, such as... Figure 6 , Figure 7 As shown.

[0081] Figure 6 This specification provides a schematic diagram of the structure of a network resource scheduling device according to one or more embodiments. The device includes: The data receiving module 602 receives local data transmitted by the client and obtains content delivery network status data transmitted by the content delivery network nodes. The status confirmation module 604 analyzes the local data and the content delivery network status data to determine the network resource status corresponding to the client. The policy execution module 606 generates a corresponding network resource control policy based on the network resource status and feeds the network resource control policy back to the client so that the client executes the network resource control policy.

[0082] Optionally, the data receiving module 602 receives static user data transmitted by the client to describe user characteristics, and establishes a user profile based on the static user data; It also receives dynamic business data from the client in real time, which describes the business execution process.

[0083] Optionally, the status confirmation module 604 determines the corresponding network fault index based on multiple preset network status dimensions and the content distribution network status data. The network resource status corresponding to the client is obtained based on the network fault index and the dynamic service data.

[0084] Optionally, the policy execution module 606 determines that the bandwidth utilization rate is higher than a first preset threshold based on the content delivery network status data, and determines that there is a preset service abnormal event based on the dynamic service data, and determines that the network fault index is lower than a second preset threshold. Generate a corresponding network resource control policy to reduce the network resource consumption of the client.

[0085] Optionally, the strategy execution module 606 determines the business type corresponding to the dynamic business data; If the business type belongs to a preset type, the current business stage is determined based on the preceding business actions already executed in the dynamic business data. If the service phase belongs to a preset phase, a corresponding network resource control policy is generated to maintain the network resource consumption of the client during the execution of the service phase and reduce the network resource consumption of the client after the service phase is completed.

[0086] Optionally, if the strategy execution module 606 determines that the network failure index is higher than the second preset threshold, it determines the corresponding backup content distribution network node based on the main content distribution network node currently corresponding to the client. Based on the dynamic business data, the corresponding business data is pre-cached to the backup content distribution network node; A corresponding network resource control policy is generated so that the client can switch the backup content distribution network node based on the network resource control policy.

[0087] Optionally, if the strategy execution module 606 determines that the number of clients that need to switch to the backup content distribution network node is higher than a preset number, it generates an inquiry command. The query command is sent to each client so that each client can perform local calculations, determine the corresponding switching benefits based on its own local data, and determine its own corresponding switching strategy based on the switching benefits. Receive the switching strategies returned by each client and combine them to obtain a set of switching strategies; Based on the number of switching strategies of each type, some switching strategies in the switching strategy set are adjusted to generate corresponding network resource control strategies, so that each client can execute its own corresponding switching strategy based on the network resource control strategies.

[0088] Optionally, the strategy execution module 606 determines the corresponding switching benefit based on its own local data, according to multiple preset benefit dimensions. The benefit dimensions include the experience gain estimation dimension and the switching cost estimation dimension; The experience gain prediction dimension includes multiple first sub-dimensions. Each first sub-dimension is determined according to each network state dimension, and the weight corresponding to each first sub-dimension is determined by the service type corresponding to the dynamic service data. The switching cost estimation dimension includes multiple second sub-dimensions, each of which is determined based on the static user data and the business type.

[0089] Optionally, the strategy execution module 606 determines the service level corresponding to the client based on the user profile and the service type corresponding to the dynamic service data; Obtain the network failure index and bandwidth utilization rate for each content delivery network node; Generate corresponding network resource control policies to allocate corresponding content delivery network nodes to the client according to the service level, and allocate corresponding network resource usage modes to the client based on the network failure index and the bandwidth utilization rate.

[0090] Optionally, the status confirmation module 604, based on multiple preset network status dimensions, fits the status curve corresponding to each network status dimension according to the content distribution network status data. Based on the state curve, the corresponding network state dimension is estimated.

[0091] Figure 7 This specification provides a schematic diagram of the structure of a network resource scheduling device according to one or more embodiments. The device includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to: Receive local data transmitted by the client, and obtain content delivery network status data transmitted by content delivery network nodes; Based on the analysis of the local data and the content delivery network status data, the network resource status corresponding to the client is determined. Based on the network resource status, a corresponding network resource control policy is generated, and the network resource control policy is fed back to the client so that the client can execute the network resource control policy.

[0092] Based on the same idea, one or more embodiments of this specification also provide a non-volatile computer storage medium corresponding to the above method, storing computer-executable instructions, wherein the computer-executable instructions are configured as follows: Receive local data transmitted by the client, and obtain content delivery network status data transmitted by content delivery network nodes; Based on the analysis of the local data and the content delivery network status data, the network resource status corresponding to the client is determined. Based on the network resource status, a corresponding network resource control policy is generated, and the network resource control policy is fed back to the client so that the client can execute the network resource control policy.

[0093] In the 1990s, improvements to a technology could be clearly distinguished as either hardware improvements (e.g., improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to the methodology). However, with technological advancements, many methodological improvements today can be considered direct improvements to the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved methodology into the hardware circuit. Therefore, it cannot be said that a methodological improvement cannot be implemented using hardware physical modules. For example, a Programmable Logic Device (PLD) (such as a Field Programmable Gate Array (FPGA)) is such an integrated circuit whose logic function is determined by the user programming the device. Designers can program and "integrate" a digital system onto a PLD themselves, without needing chip manufacturers to design and manufacture dedicated integrated circuit chips. Furthermore, nowadays, instead of manually manufacturing integrated circuit chips, this programming is mostly implemented using "logic compiler" software. Similar to the software compiler used in program development, the original code before compilation must be written in a specific programming language, called a Hardware Description Language (HDL). There are many HDLs, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, and RHDL (Ruby Hardware Description Language). Currently, VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog are the most commonly used. Those skilled in the art should also understand that by simply performing some logic programming on the method flow using one of these hardware description languages ​​and programming it into an integrated circuit, the hardware circuit implementing the logical method flow can be easily obtained.

[0094] The controller can be implemented in any suitable manner. For example, it can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicon Labs C8051F320. A memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art will also recognize that, in addition to implementing the controller in purely computer-readable program code form, the same functionality can be achieved by logically programming the method steps to make the controller take the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the means included therein for implementing various functions can also be considered as structures within the hardware component. Alternatively, the means for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0095] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0096] For ease of description, the above devices are described in terms of function, divided into various units. Of course, in implementing this specification, the functions of each unit can be implemented in one or more software and / or hardware components.

[0097] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, the embodiments of this specification can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the embodiments of this specification can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0098] This specification is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this specification. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0099] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0100] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0101] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0102] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0103] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0104] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0105] This specification can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This specification can also be practiced in distributed computing environments, where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0106] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments of apparatus, devices, and non-volatile computer storage media are basically similar to the method embodiments, so the descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0107] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0108] The above description is merely one or more embodiments of this specification and is not intended to limit this specification. Various modifications and variations can be made to the one or more embodiments of this specification by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of one or more embodiments of this specification should be included within the scope of the claims of this specification.

Claims

1. A network resource scheduling method, comprising: Receive local data transmitted by the client, and obtain content delivery network status data transmitted by content delivery network nodes; Based on the analysis of the local data and the content delivery network status data, the network resource status corresponding to the client is determined. Based on the network resource status, a corresponding network resource control policy is generated, and the network resource control policy is fed back to the client so that the client can execute the network resource control policy.

2. The method as described in claim 1, wherein receiving local data transmitted by the client specifically includes: Receive static user data transmitted from the client to describe user characteristics, and build a user profile based on the static user data; It also receives dynamic business data from the client in real time, which describes the business execution process.

3. The method as described in claim 2, wherein analysis is performed based on the local data and the content delivery network status data to determine the network resource status corresponding to the client, specifically including: Based on multiple preset network status dimensions, the corresponding network fault index is determined according to the content distribution network status data. The network resource status corresponding to the client is obtained based on the network fault index and the dynamic service data.

4. The method of claim 3, wherein a corresponding network resource control policy is generated based on the network resource status, specifically includes: Based on the content delivery network status data, it is determined that the bandwidth utilization rate is higher than a first preset threshold, and based on the dynamic service data, it is determined that there is a preset service abnormal event, and the network fault index is determined to be lower than a second preset threshold. Generate a corresponding network resource control policy to reduce the network resource consumption of the client.

5. The method as described in claim 4, generating a corresponding network resource control policy to reduce the network resource consumption of the client through the network resource control policy, specifically including: Determine the business type corresponding to the dynamic business data; If the business type belongs to a preset type, the current business stage is determined based on the preceding business actions already executed in the dynamic business data. If the service phase belongs to a preset phase, a corresponding network resource control policy is generated to maintain the network resource consumption of the client during the execution of the service phase and reduce the network resource consumption of the client after the service phase is completed.

6. The method as described in claim 3, wherein generating a corresponding network resource control policy based on the network resource status specifically includes: If it is determined that the network failure index is higher than the second preset threshold, then the corresponding backup content distribution network node is determined based on the main content distribution network node currently corresponding to the client. Based on the dynamic business data, the corresponding business data is pre-cached to the backup content distribution network node; A corresponding network resource control policy is generated so that the client can switch the backup content distribution network node based on the network resource control policy.

7. The method as described in claim 6, generating a corresponding network resource control policy to enable the client to switch the backup content distribution network node based on the network resource control policy, specifically includes: If it is determined that the number of clients that need to switch to the backup content distribution network node is higher than a preset number, an inquiry command is generated; The query command is sent to each client so that each client can perform local calculations, determine the corresponding switching benefits based on its own local data, and determine its own corresponding switching strategy based on the switching benefits. Receive the switching strategies returned by each client and combine them to obtain a set of switching strategies; Based on the number of switching strategies of each type, some switching strategies in the switching strategy set are adjusted to generate corresponding network resource control strategies, so that each client can execute its own corresponding switching strategy based on the network resource control strategies.

8. The method as described in claim 6, wherein determining the corresponding switching benefits based on its own local data specifically includes: Based on multiple preset revenue dimensions, the corresponding switching revenue is determined according to its own local data; The benefit dimensions include the experience gain estimation dimension and the switching cost estimation dimension; The experience gain prediction dimension includes multiple first sub-dimensions. Each first sub-dimension is determined according to each network state dimension, and the weight corresponding to each first sub-dimension is determined by the service type corresponding to the dynamic service data. The switching cost estimation dimension includes multiple second sub-dimensions, each of which is determined based on the static user data and the business type.

9. The method as described in claim 3, wherein generating a corresponding network resource control policy based on the network resource status specifically includes: Based on the user profile and the business type corresponding to the dynamic business data, the business level corresponding to the client is determined; Obtain the network failure index and bandwidth utilization rate for each content delivery network node; Generate corresponding network resource control policies to allocate corresponding content delivery network nodes to the client according to the service level, and allocate corresponding network resource usage modes to the client based on the network failure index and the bandwidth utilization rate.

10. The method as described in claim 1, wherein analysis is performed based on the local data and the content delivery network status data to determine the network resource status corresponding to the client, specifically including: Based on multiple preset network state dimensions, and according to the content distribution network state data, the state curves corresponding to each network state dimension are fitted to obtain the state curves. Based on the state curve, the corresponding network state dimension is estimated.

11. A network resource scheduling device, comprising: The data receiving module receives local data transmitted by the client and obtains content delivery network status data transmitted by content delivery network nodes. The status confirmation module analyzes the local data and the content delivery network status data to determine the network resource status corresponding to the client. The policy execution module generates a corresponding network resource control policy based on the network resource status and feeds the network resource control policy back to the client so that the client executes the network resource control policy.

12. The apparatus of claim 11, wherein the data receiving module receives static user data transmitted by the client for describing user characteristics, and establishes a user profile based on the static user data; It also receives dynamic business data from the client in real time, which describes the business execution process.

13. The apparatus of claim 12, wherein the status confirmation module determines the corresponding network fault index based on the content distribution network status data according to multiple preset network status dimensions; The network resource status corresponding to the client is obtained based on the network fault index and the dynamic service data.

14. The apparatus of claim 13, wherein the policy execution module determines, based on the content delivery network status data, that the bandwidth utilization rate is higher than a first preset threshold, and based on the dynamic service data, that a preset service abnormal event exists, and determines that the network fault index is lower than a second preset threshold. Generate a corresponding network resource control policy to reduce the network resource consumption of the client.

15. The apparatus of claim 14, wherein the strategy execution module determines the service type corresponding to the dynamic service data; If the business type belongs to a preset type, the current business stage is determined based on the preceding business actions already executed in the dynamic business data. If the service phase belongs to a preset phase, a corresponding network resource control policy is generated to maintain the network resource consumption of the client during the execution of the service phase and reduce the network resource consumption of the client after the service phase is completed.

16. The apparatus of claim 13, wherein if the strategy execution module determines that the network failure index is higher than a second preset threshold, it determines the corresponding backup content distribution network node based on the main content distribution network node currently corresponding to the client. Based on the dynamic business data, the corresponding business data is pre-cached to the backup content distribution network node; A corresponding network resource control policy is generated so that the client can switch the backup content distribution network node based on the network resource control policy.

17. The apparatus of claim 16, wherein the policy execution module generates an inquiry instruction if it determines that the number of clients that need to switch to the backup content distribution network node is higher than a preset number; The query command is sent to each client so that each client can perform local calculations, determine the corresponding switching benefits based on its own local data, and determine its own corresponding switching strategy based on the switching benefits. Receive the switching strategies returned by each client and combine them to obtain a set of switching strategies; Based on the number of switching strategies of each type, some switching strategies in the switching strategy set are adjusted to generate corresponding network resource control strategies, so that each client can execute its own corresponding switching strategy based on the network resource control strategies.

18. The apparatus of claim 16, wherein the strategy execution module determines the corresponding switching benefit based on its own local data, according to multiple preset benefit dimensions; in, The benefit dimensions include the experience gain estimation dimension and the switching cost estimation dimension; The experience gain prediction dimension includes multiple first sub-dimensions. Each first sub-dimension is determined according to each network state dimension, and the weight corresponding to each first sub-dimension is determined by the service type corresponding to the dynamic service data. The switching cost estimation dimension includes multiple second sub-dimensions, each of which is determined based on the static user data and the business type.

19. The apparatus of claim 13, wherein the strategy execution module determines the service level corresponding to the client based on the user profile and the service type corresponding to the dynamic service data; Obtain the network failure index and bandwidth utilization rate for each content delivery network node; Generate corresponding network resource control policies to allocate corresponding content delivery network nodes to the client according to the service level, and allocate corresponding network resource usage modes to the client based on the network failure index and the bandwidth utilization rate.

20. The apparatus of claim 11, wherein the state confirmation module, based on a preset plurality of network state dimensions, fits the state curve corresponding to each network state dimension according to the content distribution network state data; Based on the state curve, the corresponding network state dimension is estimated.

21. A network resource scheduling device, comprising: At least one processor; as well as, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enable the at least one processor to: Receive local data transmitted by the client, and obtain content delivery network status data transmitted by content delivery network nodes; Based on the analysis of the local data and the content delivery network status data, the network resource status corresponding to the client is determined. Based on the network resource status, a corresponding network resource control policy is generated, and the network resource control policy is fed back to the client so that the client can execute the network resource control policy.

Images