Convergence node deployment method for data convergence peak demand

Abstract

According to the aggregation node deployment method for the data aggregation peak demand, the deployment positions and the number of the aggregation nodes are determined through the predicted peak demand and the time information, and normal transmission and processing of data are guaranteed when the data transmission amount is increased. Meanwhile, a plurality of nodes are divided into a plurality of node clusters, and transmission paths in the node clusters are optimized according to data traffic characteristics, so that the transmission efficiency can be improved on the basis of meeting transmission requirements. Therefore, the problem that the data aggregation efficiency and precision are reduced due to the fact that the aggregation node cannot meet the peak demand under static deployment is solved.

Description

Technical Field

[0001] This disclosure relates to the field of computer network technology, and in particular to a method for deploying aggregation nodes to meet peak data aggregation demands. Background Technology

[0002] Data aggregation refers to the collection and integration of data distributed across different data sources to form a complete dataset for subsequent processing and analysis. Traditional data aggregation methods typically employ either a centralized or distributed approach. Distributed data aggregation involves deploying multiple aggregation nodes across a network. Each node is responsible for collecting and aggregating data from surrounding data sources, then sending the aggregated data to a data center for further processing. Distributed data aggregation effectively reduces network traffic and improves the efficiency and accuracy of data aggregation.

[0003] However, in practical applications, the data aggregation needs of data sources are often uneven. Sometimes data aggregation is required frequently, while at other times it is relatively infrequent. Although distributed data aggregation methods can effectively reduce network traffic, statically deployed aggregation nodes may not be able to meet peak demand when the data aggregation needs of data sources are uneven, leading to reduced efficiency and accuracy of data aggregation. Summary of the Invention

[0004] This disclosure provides a method for deploying aggregation nodes to meet peak data aggregation demands. The method determines the deployment location and number of aggregation nodes by predicting peak times and peak demand, and optimizes the node cluster based on data traffic characteristics, thereby improving data transmission efficiency.

[0005] In one aspect, this embodiment provides a method for deploying aggregation nodes to meet peak data aggregation demands. The method includes: determining the deployment location and number of at least one aggregation node based on the predicted peak time and predicted peak demand of at least one data source; dividing the nodes into multiple node clusters, each node cluster including at least one aggregation node and a data source node; and determining the optimal data transmission path based on the data traffic characteristics within each node cluster, so as to configure computing resources for the aggregation node.

[0006] In embodiments of this disclosure, the method further includes: acquiring multiple historical data from at least one data source; determining the historical peak time and historical peak demand for each data source based on the multiple historical data; and determining the predicted peak time and predicted peak demand for each data source based on the historical peak time and historical peak demand.

[0007] In the embodiments of this disclosure, the data source includes at least one or more of the following: cloud service data source in the operator field, communication value-added service data source, capability open platform data source, and support service data source.

[0008] In the embodiments of this disclosure, determining the historical peak time and historical peak demand of each data source based on multiple historical data includes: determining a preset time range and time granularity; aggregating multiple historical data according to the time granularity within the preset time range; and determining the historical peak time and historical peak demand of each data source within the preset time range based on the aggregated historical data.

[0009] In the embodiments of this disclosure, determining the historical peak time and historical peak demand within a preset time range for each data source based on aggregated historical data includes: determining the data aggregation demand within multiple time windows, each time window having the same time length, and the time windows being determined by dividing a preset time range; comparing the data aggregation demand within the current time window with the data aggregation demand within the previous time window; and determining the historical peak time and historical peak demand within a preset time range for each data source based on the comparison results.

[0010] In the embodiments of this disclosure, determining the historical peak time and historical peak demand within a preset time range for each data source includes: in response to the data aggregation demand in the current time window being greater than the data aggregation demand in the previous time window, determining the current time window as a peak time period, the historical peak time being determined by the sum of the times corresponding to consecutive peak time periods, and the historical peak demand being determined by the sum of the data aggregation demand corresponding to consecutive peak time periods.

[0011] In embodiments of this disclosure, determining the deployment location and number of at least one aggregation node based on the predicted peak time and predicted peak demand of at least one data source includes: determining multiple aggregation time periods; determining the sum of the predicted data volume generated by multiple data sources within each aggregation time period, wherein the predicted data volume is determined at least based on historical data predictions; and determining the deployment location range and number of at least one aggregation node based on the largest sum of the multiple data volumes.

[0012] In embodiments of this disclosure, determining the deployment location and number of at least one aggregation node based on the largest sum of multiple data volumes includes: determining the deployment location of the aggregation node based on the geographical location information and network condition information of the data source; and determining the number of aggregation nodes based on the data processing capability of the aggregation node with the largest sum of data volumes.

[0013] In embodiments of this disclosure, the method further includes: evaluating the aggregation node based on at least one or more of bandwidth information, latency information, packet loss rate information, and throughput information.

[0014] In embodiments of this disclosure, dividing multiple node clusters includes: determining a network topology based on data sources and aggregation nodes, the network topology including at least one aggregation node and multiple data source nodes; and dividing multiple aggregation nodes and data source nodes into multiple node clusters based on the geographical location information of the data sources, predicted peak time, predicted peak demand, and deployment location of the aggregation nodes, each node cluster including at least one aggregation node and multiple data source nodes.

[0015] In the embodiments of this disclosure, multiple aggregation nodes and data source nodes are divided into multiple node clusters, including: dividing multiple aggregation nodes and data source nodes into multiple initial node clusters based on geographical location information, wherein each initial node cluster includes at least one aggregation node and multiple data source nodes; and further dividing the multiple initial node clusters using the data generation and data consumption of the aggregation nodes and data source nodes to obtain multiple node clusters.

[0016] In embodiments of this disclosure, multiple aggregation nodes and data source nodes are divided into multiple initial node clusters based on geographic location information, including: determining the number of clusters and the initial cluster center; calculating the distance from each aggregation node and data source node to the initial cluster center; and assigning each aggregation node and data source node to a different initial node cluster based on the distance to the initial cluster center.

[0017] In embodiments of this disclosure, an optimal data transmission path is determined based on the data traffic characteristics within each node cluster in order to configure computing resources for the aggregation node. This includes: constructing a minimum spanning tree for each node cluster based on the data traffic characteristics; and determining an optimal data transmission path for each node cluster based on the minimum spanning tree, wherein the optimal data transmission path includes at least one aggregation node.

[0018] In embodiments of this disclosure, the method further includes: allocating computing and storage resources to the aggregation nodes included in the optimal data transmission path based on the optimal data transmission path.

[0019] In embodiments of this disclosure, the method further includes: acquiring real-time monitoring data information, which includes at least latency rate, throughput, packet loss rate, and peak demand changes of the data source.

[0020] In embodiments of this disclosure, the method further includes: adjusting the deployment location and number of aggregation nodes based on real-time monitoring data.

[0021] On the other hand, embodiments of this disclosure provide a network device, including: a communication interface configured to perform wireless communication with a terminal device; a memory for storing computer-readable instructions; and a processor for executing the computer-readable instructions, causing the network device to perform the aggregation node deployment method.

[0022] In another aspect, embodiments of this disclosure provide a non-transitory computer-readable storage medium for storing computer-readable instructions that, when executed by a processor, cause the processor to perform the aforementioned aggregation node deployment method.

[0023] In another aspect, embodiments of this disclosure provide a computer program product, including a computer program that, when executed by a processor, implements the above-described aggregation node deployment method. Attached Figure Description

[0024] The above and other objects, features, and advantages of this disclosure will become more apparent from the more detailed description of the embodiments thereof in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this disclosure and form part of the specification. They are used together with the embodiments of this disclosure to explain the disclosure and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0025] Figure 1 The schematic diagram illustrates an environmental application according to an embodiment of the present disclosure.

[0026] Figure 2 The flowchart illustrates a method for deploying aggregation nodes to meet peak data aggregation demands according to an embodiment of the present disclosure.

[0027] Figure 3 The flowchart illustrates another method for deploying aggregation nodes to meet peak data aggregation demands according to an embodiment of the present disclosure.

[0028] Figure 4 A flowchart illustrating the determination of the deployment location and number of at least one aggregation node according to an embodiment of the present disclosure is provided.

[0029] Figure 5 The flowchart illustrating the division of multiple node clusters according to an embodiment of the present disclosure is shown in the illustration.

[0030] Figure 6 The flowchart illustrating the division of multiple node clusters according to an embodiment of the present disclosure is shown in the illustration.

[0031] Figure 7 A block diagram of a network device according to an embodiment of the present disclosure is shown schematically.

[0032] Figure 8 A block diagram illustrating a non-transitory computer-readable storage medium according to an embodiment of the present disclosure is shown.

[0033] Figure 9 A block diagram illustrating a computer program product according to an embodiment of the present disclosure is shown schematically. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of this disclosure more apparent, exemplary embodiments according to this disclosure will now be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this disclosure, and not all embodiments of this disclosure. It should be understood that this disclosure is not limited to the exemplary embodiments described herein.

[0035] Centralized data aggregation methods face challenges such as network bandwidth limitations, data transmission latency, and data security. While distributed data aggregation methods can effectively reduce network traffic, statically deployed aggregation nodes may fail to meet peak demand when data aggregation needs from data sources are uneven, leading to reduced efficiency and accuracy in data aggregation.

[0036] Based on this, embodiments of the present disclosure provide a method for deploying aggregation nodes to meet peak data aggregation demands, thereby improving the efficiency and accuracy of data aggregation, reducing network transmission volume and latency, and enhancing the real-time performance and accuracy of data processing.

[0037] Next, this article will combine Figures 1-6 This disclosure provides a detailed description of a method for deploying aggregation nodes to meet peak data aggregation demands, according to an embodiment of the present disclosure.

[0038] Figure 1 The schematic diagram illustrates an environmental application according to an embodiment of the present disclosure.

[0039] like Figure 1 As shown, it includes a computing device 101, multiple data source nodes 102, and multiple aggregation nodes 103. The computing device 101 and the data source nodes 102 and aggregation nodes 103 can perform real-time data transmission.

[0040] The computing device 101 can acquire multiple historical data points from the data source node 102 and perform predictive analysis on the historical data to predict data peak information in the future time period. Based on the data peak information and the geographical location information of the data source node 102, it dynamically deploys the aggregation node 103 and optimizes the data transmission path, thereby improving the efficiency and accuracy of data aggregation, reducing network transmission volume and latency, and improving the real-time performance and accuracy of data processing.

[0041] In some embodiments, computing device 101 may be a network device with computing capabilities, such as a server. Computing device 101 may also be a network device that includes a master device and distributed devices, such as a master device responsible for global decision-making and edge devices sharing local computing.

[0042] Data source node 102 can be a network device that generates data, such as a mobile device, tablet, laptop, workstation, or other device capable of generating network data. Aggregation node 103 can be a network device capable of receiving data generated by data source node 102, processing and forwarding the data, such as an edge server, rack server, cloud server, etc.

[0043] Figure 2 The flowchart illustrates a method for deploying aggregation nodes to meet peak data aggregation demands according to an embodiment of the present disclosure.

[0044] like Figure 2 As shown, an embodiment of this disclosure provides a method for deploying aggregation nodes to meet peak data aggregation demands, comprising steps S201, S202, and S203: S201. Based on the predicted peak time and predicted peak demand from at least one data source, determine the deployment location and number of at least one aggregation node.

[0045] In the embodiments of this disclosure, a data source can refer to the place where the data is generated or the entity that generates the data, encompassing both physical location / system and the explicit source of data generation. For example, the place where data is generated can include: a carrier's cloud service server (a physical device storing user storage space and CPU usage data), communication network equipment (hardware recording user call logs and SMS records), and a capability open platform system (a software system storing user data traffic consumption records). The entity that generates the data can include: a user's cloud service usage behavior (generating storage space and CPU usage data) and a user's communication behavior (generating call and SMS record data). It is understood that regardless of whether it is categorized by business scenario, storage medium, or generating entity, the essence of a data source is the source of data generation. As long as a certain entity, device, system, or scenario continuously generates data, it is a data source.

[0046] In embodiments of this disclosure, at least one data source can be predicted, its predicted peak time and predicted peak demand can be estimated, thereby dynamically estimating the deployment location and number of aggregation nodes from a prediction perspective.

[0047] It is evident that predicting peak time and peak demand is crucial for determining the deployment location and number of aggregation nodes. This disclosure provides a method for determining peak time and peak demand, such as... Figure 3 As shown.

[0048] Figure 3 The flowchart illustrates another method for deploying aggregation nodes to meet peak data aggregation demands according to an embodiment of the present disclosure.

[0049] like Figure 3 As shown, another method for deploying aggregation nodes to meet peak data aggregation demands, according to an embodiment of this disclosure, includes steps S301, S302, and S303: S301. Obtain multiple historical data from at least one data source.

[0050] By collecting and analyzing historical data from at least one data source within the telecom operator sector, data aggregation needs for different time periods can be identified. This collection of at least one data source can be automated, such as by accessing the operator's official website, using web crawlers, or data scraping.

[0051] In the embodiments of this disclosure, the data source includes at least one or more of the following: cloud service data source in the operator field, communication value-added service data source, capability open platform data source, and support service data source.

[0052] Cloud service data sources can include data from cloud storage, cloud computing, cloud databases, etc., such as user storage space, CPU utilization, and memory utilization. Communication value-added service data sources can include data from SMS, voice, and video communication services, such as user call logs, SMS records, and video call records. Capability open platform data sources can include data from internet billing, unified authentication, big data, etc., such as user data traffic consumption records, authentication login records, and user profiles. Support service data sources can include data from services such as HeLife, SuiEXing Wireless Local Area Network (WLAN), the Testing Center, and MengWang, such as user location information, WLAN access records, and user reviews.

[0053] S302. Based on multiple historical data, determine the historical peak time and historical peak demand for each data source.

[0054] In the embodiments of this disclosure, data can be processed based on collected data sources and historical data. For each data source, basic information such as its data type, data format, and data volume is determined. For historical data, its time range and time granularity are determined, and it is cleaned to remove outliers and missing values. Furthermore, multiple historical data are aggregated according to time granularity to identify peak values, thereby finding the historical peak time and historical peak demand for each data source within a specific time period.

[0055] In the embodiments of this disclosure, the determination of the historical peak time and historical peak demand of each data source in S302 can be determined by the following methods: determining a preset time range and time granularity; aggregating multiple historical data according to the time granularity within the preset time range; and determining the historical peak time and historical peak demand of each data source within the preset time range based on the aggregated historical data.

[0056] In the embodiments of this disclosure, a preset time range can be determined first, such as the past week or the past month, and then data for the corresponding time period can be extracted from historical data. It is also necessary to determine the time granularity. The time granularity can be the smallest statistical unit of historical data, such as 2 hours or 5 hours.

[0057] Furthermore, we can first determine which historical data each data source includes, then select data corresponding to a preset time range from these historical data, and then aggregate these selected historical data according to time granularity to obtain aggregated historical data for a single data source within the preset time range. Then, we can perform peak identification on the aggregated historical data to determine the historical peak time and historical peak demand for each data source within the preset time range.

[0058] In the embodiments of this disclosure, determining the historical peak time and historical peak demand within a preset time range for each data source based on aggregated historical data includes: determining the data aggregation demand within multiple time windows, each time window having the same time length, and the time windows being determined by dividing a preset time range; comparing the data aggregation demand within the current time window with the data aggregation demand within the previous time window; and determining the historical peak time and historical peak demand within a preset time range for each data source based on the comparison results.

[0059] First, a time window size can be selected and determined, such as 1 hour, 1 day, or 1 week. In some cases, the time window size can be equal to the time granularity to maximize the reliability of the analysis results and make subsequent node deployment more targeted. Further, the time axis corresponding to the preset time range is divided into multiple time windows, each with the same size as the selected time window, and the data aggregation requirement for each time window is determined. The data aggregation requirement is the sum of the data volume of multiple historical data from the same data source within the time period corresponding to that time window.

[0060] Furthermore, for each time window, its data aggregation demand can be compared with that of the previous time window. The comparison results determine the historical peak time and historical peak demand for each data source within a preset time range. For example, if the data aggregation demand for multiple consecutive time windows is greater than that of the previous time window, it can be determined that the data aggregation demand is gradually increasing within the time period comprised of these multiple time windows. When the upward trend ends, the historical peak demand for that data source can be determined. The upward trend ends in the first time window following these consecutive time windows, when its data aggregation demand is less than that of the previous time window.

[0061] In the embodiments of this disclosure, determining the historical peak time and historical peak demand within a preset time range for each data source includes: in response to the data aggregation demand in the current time window being greater than the data aggregation demand in the previous time window, determining the current time window as a peak time period, the historical peak time being determined by the sum of the times corresponding to consecutive peak time periods, and the historical peak demand being determined by the sum of the data aggregation demand corresponding to consecutive peak time periods.

[0062] If the data aggregation demand in the current time window is greater than that in the previous time window, then the current time window can be marked as a peak time period, and the above steps can be repeated until all time windows have been traversed.

[0063] Therefore, a historical peak demand can be determined by the sum of the aggregated demand corresponding to n consecutive peak time periods, that is:

[0064] Among them, P d D represents the historical peak demand. i Let represent the data aggregation demand within the i-th time window marked as the peak time period, and n represent the number of time windows marked as the peak time period.

[0065] Correspondingly, a historical peak demand corresponds to a historical peak time, which is determined by the sum of the times corresponding to n consecutive peak time periods. Within a preset time range, there may be multiple historical peak times, each corresponding to a historical peak demand. By performing statistical and visual analysis on historical data, it is possible to identify multiple potential periodic or non-periodic peak periods, thus providing guidance for the subsequent dynamic deployment of aggregation nodes. Furthermore, calculating the data aggregation demand for different time periods can help enterprises optimize resource allocation and improve data processing capabilities.

[0066] In some embodiments, the historical peak time can refer to a period of time during which the data aggregation demand is increasing. The historical peak demand can refer to the sum of the data aggregation demands corresponding to that period, or it can refer to the data aggregation demand corresponding to the highest point within that period.

[0067] S303. Based on historical peak times and historical peak demand, determine the predicted peak time and predicted peak demand for each data source.

[0068] By identifying multiple historical peak times and historical peak demand for each data source within a preset time range, the predicted peak time and predicted peak demand for each data source within a specific future time period can be predicted. For example, for a data source that experienced peak demand from 9 AM to 10 AM Monday through Friday over the past week, it can be predicted, based on the historical peak times and historical peak demand for that data source, that it is also likely to experience peak demand from 9 AM to 10 AM Monday through Friday over the next week.

[0069] In some embodiments, the data aggregation demand of each data source within a preset time range can be analyzed to predict the data aggregation demand of each data source in a specific future time period. For example, by analyzing the data aggregation demand of a data source in the past month, the data aggregation demand of that data source in the next month can be predicted. This can be done by predicting future peak times and peak demands based on historical peak times and historical peak demands, or by predicting off-peak times and off-peak demands.

[0070] Therefore, by using multiple historical peak times and historical peak demand for each data source within a preset time range, the predicted peak time and predicted peak demand for each data source can be predicted, thus providing accurate demand basis for the subsequent deployment of aggregation nodes.

[0071] After predicting the peak time and peak demand for each data source, aggregation nodes can be dynamically deployed based on these predictions. For example... Figure 4 As shown, Figure 4 A flowchart illustrating the determination of the deployment location and number of at least one aggregation node according to an embodiment of the present disclosure is provided.

[0072] like Figure 4 As shown, the above S201 includes S401, S402 and S403: S401. Determine multiple convergence time periods.

[0073] In the embodiments of this disclosure, the aggregation time period can refer to the smallest unit of aggregation data statistics. The aggregation time period can be 1 hour or 1 day. In some cases, the size of the aggregation time period can be the same as the size of the time window and the size of the time granularity.

[0074] S402. Determine the sum of the predicted data volume from multiple data sources within each aggregation time period. The predicted data volume shall be determined at least based on historical data.

[0075] In the embodiments of this disclosure, referring to S303 above, the data aggregation demand of each data source within a specific future time period can be determined first. Then, within the specific future time period, the data aggregation demand of each data source in different aggregation time periods can be determined. Then, for the same aggregation time period, the data aggregation demand of all data sources within that aggregation time period is aggregated, and the sum of the predicted data volume generated by all data sources within that aggregation time period is taken as the data aggregation demand for that aggregation time period. Thus, the data aggregation demand for multiple aggregation time periods can be obtained.

[0076] S403. Determine the deployment location range and number of at least one aggregation node based on the largest sum of multiple data volumes.

[0077] Furthermore, the maximum peak demand can be identified by utilizing the data aggregation demand across multiple aggregation time periods. Specifically, the maximum peak demand can be determined by summing the largest sum of multiple data volumes, and the number of aggregation nodes can be based on this sum. In some embodiments, the maximum peak demand can also be identified using a time window, such as the method described in S302 above.

[0078] It is understandable that the maximum peak demand is related to the predicted peak demand from at least multiple data sources. When the size of the aggregation period and the time window are the same, the larger the sum of the data volume in the aggregation period is when the data volume from more data sources in the aggregation period is the predicted peak demand.

[0079] In the embodiments of this disclosure, S403 includes: determining the deployment location of the aggregation node based on the geographical location information and network condition information of the data source; and determining the number of aggregation nodes based on the data processing capacity of the aggregation node with the largest sum of data volume.

[0080] Given a determined peak demand, the number of aggregation nodes can be determined based on this peak demand and the processing capacity of a single aggregation node. For example, based on historical data, the peak demand during peak periods is: Daily peak data volume is... D max =10TB, assuming the maximum processing capacity of a single aggregation node is...C node =2TB / hour (determined based on hardware configuration and test results), then the number of aggregation nodes can be determined using the following formula:

[0081] in, N The number of aggregation nodes, D max For the maximum peak demand, C node For the processing capacity of a single aggregation node, T The peak duration is represented by units matching the processing capacity of a single aggregation node. For example, if the peak lasts for 2 hours, then N = 10 / (2×2) =3 aggregation nodes.

[0082] In some embodiments, load balancing and fault tolerance can also be considered by adding redundant aggregation nodes, for example, adding 20% ​​redundant aggregation nodes. Then, the final deployment... N =3 × 1.2 ≈ 4 convergence nodes.

[0083] When the number of aggregation nodes is determined, the deployment location of each aggregation node can be determined based on the geographical location and network conditions of the data source. For example, the Euclidean distance between the data source and multiple candidate deployment locations can be calculated using GPS coordinates, and the candidate deployment location with the smallest distance can be selected as the aggregation node's deployment location. For instance, if the coordinates of data source A are (longitude 113.3, latitude 23.1) and the coordinates of candidate deployment point B are (longitude 113.5, latitude 23.0), then the Euclidean distance L ≈ 0.22 degrees. Furthermore, when there are multiple data sources, the Euclidean distances between each data source and each candidate deployment location can be calculated separately, thereby comprehensively determining the deployment location of the aggregation node.

[0084] Network condition information can refer to the use of network testing tools (Internet Performance Test, IPERF) to measure the bandwidth, latency, and packet loss rate between candidate deployment locations and data sources, and to select candidate deployment locations with high bandwidth (e.g., ≥100Mbps), low latency (e.g., ≤50ms), and low packet loss rate (e.g., ≤1%) as the deployment locations for aggregation nodes.

[0085] Furthermore, the Euclidean distance and network condition information between each candidate deployment location and multiple data sources can be calculated. Then, a weighted score is calculated by combining geographical distance and network indicators, for example: weight = 0.6 × distance score + 0.4 × network score. This allows the candidate deployment location with the highest total score, i.e., the one that is geographically close to the data source and has good network conditions, to be selected as the deployment location of the aggregation node.

[0086] In embodiments of this disclosure, such as Figure 4 The method also includes: evaluating the aggregation node based on at least one or more of the following: bandwidth information, latency information, packet loss rate information, and throughput information.

[0087] When determining the location and number of aggregation nodes, the feasibility and efficiency of data transmission can be evaluated through key performance indicators (bandwidth, latency, packet loss rate, throughput), thereby ensuring that data can be aggregated stably and quickly after node deployment.

[0088] Bandwidth information refers to the maximum transmission rate along a data transmission path. Network testing tools (such as iperf or Network Flow Information Export, NetFlow) are used to measure bandwidth usage during peak and off-peak times. Latency information refers to the time it takes for data to travel from the source to the destination. Packet loss rate information refers to the percentage of data lost during transmission, reflecting network reliability. A high packet loss rate can severely impact data transmission quality. Throughput information refers to the actual amount of data successfully transmitted over the network, reflecting actual network performance and influenced by both bandwidth and latency.

[0089] return Figure 2 S202. Divide into multiple node clusters, each node cluster including at least one aggregation node and a data source node.

[0090] After determining the deployment location and number of aggregation nodes, multiple aggregation nodes and multiple data source nodes can be divided into multiple node clusters based on geographical location information and data traffic characteristics. For example, aggregation nodes and data source nodes that are geographically close can be grouped into one node cluster.

[0091] like Figure 5 As shown, S202 includes S501 and S502. Figure 5 The flowchart illustrating the division of multiple node clusters according to an embodiment of the present disclosure is shown in the illustration.

[0092] S501. Based on the data source and the aggregation node, determine the network topology diagram, which includes at least one aggregation node and multiple data source nodes.

[0093] In the embodiments of this disclosure, a network topology map can be drawn based on the determined deployment location and number of aggregation nodes, as well as the geographical location information and data traffic characteristics of the data source. Data traffic characteristics may include network data information such as peak time periods, data flow direction, data type, frequency, size, node connection method, and bandwidth limitations.

[0094] The data types of data sources can include: classified by business attributes: cloud business data (such as user storage space, CPU utilization), communication data (such as call records, SMS content), capability open data (such as traffic consumption records, authentication logs), etc.; classified by data format: structured data (database tables, such as user ID + storage amount + time), semi-structured data (JSON / XML, such as cloud service status returned by application programming interface (API), and unstructured data (log files, audio and video clips, such as call recordings).

[0095] Data source frequency statistics can refer to the interval at which each data source generates data: high frequency (e.g., cloud server CPU usage, 1 record per second), medium frequency (e.g., user WLAN access records, 1 record every 5 minutes), and low frequency (e.g., user monthly bills, 1 record per day). Data source size measurement can refer to calculating the total amount of data per single record or per unit of time: small data (e.g., SMS records, 1KB per record), medium data (e.g., user profiles, 1MB per profile), and large data (e.g., cloud drive file transfers, 100GB per hour).

[0096] Furthermore, the connection relationships and limitations between data sources, potential aggregation nodes, and data centers can be clearly defined, i.e., the node connection methods. By obtaining the devices where all data sources are located (such as server IP, user terminal location), existing network devices (routers, switches, gateways), and data center addresses, it can be determined whether the connections between nodes are wired (fiber optic, network cable) or wireless (5G, WLAN), the physical links of the connections are recorded (such as server A being connected to switch B via fiber optic), and the maximum bandwidth of each link (such as the link bandwidth from switch B to the data center being 1Gbps) and the minimum guaranteed bandwidth (such as a dedicated cloud service data link guaranteeing 500Mbps) can be determined.

[0097] Then, based on the node location → connection relationship → bandwidth label, a network topology diagram is drawn, and the physical topology is transformed into a "logical mapping table" to clarify the network attributes (IP address, subnet) and link performance parameters (bandwidth, latency, packet loss rate) of each node.

[0098] In some embodiments, regions can be divided according to the coverage of the aggregation node (e.g., "Node 1 covers cloud service data sources in the eastern region, and Node 2 covers communication data sources in the western region") to ensure that the data source in each region is as close as possible to the responsible aggregation node. Then, the data flow patterns within the region are analyzed, including peak period analysis and data flow direction analysis. Peak period analysis: Based on historical data and forecast results, the peak periods for data sources in each region are statistically analyzed (e.g., cloud service data in the eastern region reaches its peak between 9:00 and 11:00 on weekdays). Data flow direction analysis: The path of data from "data source → aggregation node → data center" is recorded.

[0099] S502. Based on the geographic location information of the data source, the predicted peak time, the predicted peak demand, and the deployment location of the aggregation node, multiple aggregation nodes and data source nodes are divided into multiple node clusters, each node cluster including at least one aggregation node and multiple data source nodes.

[0100] Furthermore, based on the network topology diagram drawn above, and the network attributes, link performance parameters, peak periods, data flow directions, and other data information of each node analyzed based on the network topology diagram, multiple aggregation nodes and data source nodes can be divided into multiple node clusters.

[0101] Figure 6 The flowchart illustrating the division of multiple node clusters according to an embodiment of the present disclosure is shown in the illustration.

[0102] like Figure 6 As shown, S502 includes S601 and S602: S601. Based on geographic location information, multiple aggregation nodes and data source nodes are divided into multiple initial node clusters, each initial node cluster including at least one aggregation node and multiple data source nodes.

[0103] During node clustering, nodes are first grouped based on their geographical location information. Nodes with similar geographical locations, such as sink nodes and data source nodes, can be grouped into the same initial node cluster. Each initial node cluster includes at least one sink node and multiple data source nodes.

[0104] In the embodiments of this disclosure, S601 includes: determining the number of clusters and the initial cluster center; calculating the distance from each sink node and data source node to the initial cluster center; and dividing each sink node and data source node into different initial node clusters based on the distance to the initial cluster center.

[0105] First, collect the geographical location information of all nodes, including longitude and latitude. Then, determine the number of clusters (K value), i.e., how many groups to divide all nodes into. Next, randomly select K nodes as the initial cluster centers. For each node, calculate its distance to all K center points and assign it to the cluster represented by the nearest center point. The distance can be calculated using the Euclidean distance formula; the formula for calculating distance in geographic locations is:

[0106] Where L is the distance, x1 is the longitude of the center point, y1 is the latitude of the center point, x2 is the longitude of the node, and y2 is the latitude of the node.

[0107] Once all nodes have been assigned to clusters, the center point of each cluster is updated to the average of the positions of all nodes within the cluster. Then, the distances between nodes and all K center points are repeatedly calculated, and the cluster center points are updated, until the positions of the center points no longer change, thus completing the determination of the initial node clusters.

[0108] S602. Using the data generation and data consumption of the aggregation node and the data source node, the multiple initial node clusters are further divided to obtain multiple node clusters.

[0109] Data generation and data consumption are core characteristics used to describe the data interaction behavior of nodes, quantifying the data output and input attributes of nodes in the network. Data generation can refer to a node's output capacity as a data source, specifically the total amount of data a single node generates and sends to the network per unit time. Nodes with high generation (such as core business servers) require more uplink bandwidth and aggregation resources, and are tended to be grouped with nodes that can handle large amounts of data input during clustering. Data consumption can refer to a node's input requirements as a data receiver, specifically the total amount of data a single node receives and processes from the network per unit time. Nodes with high consumption (such as regional aggregation centers) require more downlink bandwidth and processing power, and are tended to be grouped with nodes that generate large amounts of data during clustering to reduce cross-group transmission costs.

[0110] In some embodiments, in addition to data generation and data consumption being used as the basis for optimizing the initial node cluster, the initial node cluster can also be comprehensively optimized based on indicators such as data type matching degree, interaction frequency, and peak synchronization.

[0111] When optimizing the initial node cluster, we can first collect the data generation and consumption characteristics of each node. Then, we set the search radius and the minimum number of neighbors. If the distance between two points is less than or equal to the search radius, these two points are considered neighbors. The minimum number of neighbors refers to the minimum number of neighbors required for a point to become a core point. If the number of neighbor points within the search radius of a point is greater than or equal to the minimum number of neighbors, then that point is considered a core point.

[0112] For example, the distance between two nodes can be determined using data generation and data consumption. For instance, "data generation per hour (GB)" and "data consumption per hour (GB)" can be used as two feature dimensions to describe nodes. The distance between node A (generating 10GB, consuming 5GB) and node B (generating 12GB, consuming 6GB) can be calculated (e.g., Euclidean distance ≈ 2.24). If the search radius is set to 3, then node A and node B are considered neighbors; if the search radius is set to 2, then node A and node B are not considered neighbors.

[0113] For each node, count the number of points within its search radius. If a node has at least the minimum number of neighboring points within its search radius, then that node is marked as a core node. For each core node, if it has not yet been assigned to any cluster, create a new cluster and add all points within its search radius to this cluster. This process is repeated recursively until all core nodes have been visited. Points that are not core nodes and do not belong to any cluster are marked as noise.

[0114] By clustering nodes according to data behavior, we can overcome the limitations of physical location and dynamically adapt to changes in data characteristics. Ultimately, this provides accurate grouping criteria for resource allocation and topology optimization, achieving the goal of "efficient collaboration within the same group and the lowest overall cost".

[0115] return Figure 2 S203. Based on the data traffic characteristics within each node cluster, determine the optimal data transmission path in order to configure computing resources for the aggregation node.

[0116] Data traffic characteristics can include key performance indicators of nodes within a cluster, such as latency, bandwidth, and packet loss rate. These characteristics are used to determine the connection cost weights between any two nodes within the cluster; these weights reflect transmission efficiency and reliability. Based on these connection cost weights, an optimal connection topology is constructed within the cluster, allowing for adjustments to the connection methods and data transmission paths of nodes within the cluster. This makes data flow within the cluster more efficient and provides a clear local basis for global optimization of the entire network.

[0117] According to embodiments of this disclosure, by determining the deployment location and number of aggregation nodes based on predicted peak demand and time information, it is beneficial to ensure normal data transmission and processing when data transmission volume increases. Furthermore, by dividing multiple nodes into multiple node clusters and optimizing the transmission paths within each cluster based on data traffic characteristics, transmission efficiency can be improved while meeting transmission requirements. This overcomes the problem of statically deployed aggregation nodes failing to meet peak demand, leading to reduced efficiency and accuracy in data aggregation.

[0118] In the embodiments of this disclosure, S203 includes: constructing a minimum spanning tree for each node cluster based on data traffic characteristics; and determining an optimal data transmission path for each node cluster based on the minimum spanning tree, wherein the optimal data transmission path includes at least one aggregation node.

[0119] In some embodiments, the connection cost (weight) between two nodes can be determined by both network latency and bandwidth. If the business prioritizes speed (e.g., real-time video transmission), a larger value for α (e.g., α=0.7, β=0.3) is used, allowing latency to have a greater impact on cost; if the business prioritizes transmission volume (e.g., large file backup), a larger value for β (e.g., α=0.3, β=0.7) is used, allowing bandwidth to have a greater impact on cost. The weight can be determined by the following formula:

[0120] Here, α and β are coefficients that adjust the effects of latency and bandwidth.

[0121] Furthermore, for any two nodes within the cluster (such as node A and node B, A and C, B and C, etc.), their delays and bandwidths are measured and substituted into the formula to calculate the weights. Then, a weighted graph is constructed within the cluster: all nodes within the cluster are used as vertices of the weighted graph, the physical or logical connection between any two nodes is used as an edge of the weighted graph, and the weight between any two nodes is calculated as the weight of the edge.

[0122] Then, a minimum spanning tree is generated using a weighted graph, for example, using Prim's algorithm (Prim), to find the lowest-cost connection method. The cluster's sink node is used as the root node, and each time, the edge with the smallest weight is added to the tree from multiple edges until all nodes are included (ensuring connectivity and minimizing the total weight). The minimum spanning tree determines the optimal data transmission path for each node cluster. It is the set of lowest-cost paths that allow all nodes to interconnect; it forms the basic transportation network for all data transmission within the cluster, ensuring the lowest overall transmission cost. By using the sink node as the root node, the path cost from all nodes to the sink node is minimized, and the logic is clearest, ultimately achieving the business goal of "efficiently aggregating data within the cluster to the core node."

[0123] In some embodiments, the generated minimum spanning tree can be evaluated to determine whether it effectively reduces data transmission costs within the cluster. The evaluation can be performed by comparing total weights and actual transmission efficiency. For example, the total weight of the minimum spanning tree can be calculated and compared with the total weights of random or fully connected connections before optimization; if the total weight of the minimum spanning tree is smaller, it indicates cost reduction. Alternatively, network testing tools (such as iperf) can be used to test the latency and bandwidth utilization of data transmission within the cluster after optimization to verify whether it is better than before. In some cases, if the evaluation finds the minimum spanning tree ineffective (e.g., a sudden increase in actual latency on a certain edge), the α and β coefficients are readjusted or the weights are recalculated, and the minimum spanning tree is generated again until the cost reduction target is met.

[0124] In some cases, for the same node cluster, there may be different candidate aggregation nodes and corresponding transmission paths, resulting in multiple solutions. Furthermore, the total cost of each solution (including node hardware cost, transmission latency, bandwidth usage, maintenance difficulty, etc.) can be compared to select the lowest-cost solution that meets the requirements. For example, consider users in the South China region as a node cluster. The peak demand for this cluster is between 8 PM and 10 PM, and the data flow is: User -> Aggregation Node -> Guangzhou Core Data Center. Solution A: Select a node in Guangzhou with strong computing power but limited outbound bandwidth as the aggregation node. Solution B: Select a node in Foshan with slightly weaker computing power but direct connection to the Guangzhou center and ample bandwidth as the aggregation node. Evaluation: During peak evening hours, the node in Solution A may become a bottleneck due to bandwidth constraints, leading to a surge in latency; while the node in Solution B has slightly weaker computing power, its network path is superior, potentially resulting in higher overall throughput. Decision: Through simulation or performance evaluation, we found that Solution B's overall performance (latency + throughput) during peak periods is better than Solution A; therefore, Solution B is the optimal choice.

[0125] By conducting detailed analysis and classification of data sources, and clustering nodes based on data source characteristics (geographical location, data generation, and consumption), the layout of aggregation nodes can be rationally planned. Combined with optimization of intra-cluster topology (reducing internal transmission distance and cost through the minimum spanning tree algorithm), data processing efficiency can be significantly improved, especially when dealing with large-scale datasets.

[0126] In embodiments of this disclosure, such as Figure 2 The method also includes allocating computing and storage resources to the aggregation nodes included in the optimal data transmission path based on the optimal data transmission path.

[0127] Computing and storage resources can be allocated to aggregation nodes based on data processing and data storage requirements. This allocation includes CPU allocation, memory allocation, and storage allocation. The number of CPU cores is calculated as follows: (Data processing time × Data volume) / (Processing capacity × Target processing time). Here, data processing time refers to the time required to process a single piece of data using one CPU core; data volume refers to the total amount of data the aggregation node needs to process; processing capacity refers to the basic processing performance coefficient of a single CPU core; and the target processing time refers to the maximum time desired to complete all data processing. Storage space (GB) is calculated as follows: Storage space (GB) = Data volume (GB) × Data retention period (days) / Data compression ratio. Here, data volume (GB) refers to the amount of raw data the aggregation node needs to store each day; data retention period refers to the number of days the data needs to be stored in the node; and data compression ratio refers to the percentage reduction in file size after compression.

[0128] By allocating sufficient computing and storage resources to aggregation nodes, CPU resources are prevented from causing data processing delays due to insufficient cores, nor from becoming idle and wasted due to excessive cores. Storage resources meet data retention requirements while avoiding overcapacity through compression optimization, thus achieving quantitative and controllable resource allocation.

[0129] In embodiments of this disclosure, such as Figure 2 The method also includes: acquiring real-time monitoring data, which includes at least latency, throughput, packet loss rate, and peak demand changes of the data source.

[0130] Network monitoring software can be used to collect metrics in real time to monitor network status, including latency (data transmission time), throughput (actual data transmission volume), and packet loss rate, to determine network congestion and link reliability. Monitoring data traffic allows for real-time tracking of data transmission volume in each cluster and node (e.g., whether peak values ​​have been reached or if there are sudden traffic spikes). Monitoring node performance allows for monitoring of aggregation node CPU utilization, memory usage, and remaining storage space (e.g., whether high load is causing processing delays), enabling timely detection of problems (e.g., a sudden drop in throughput or excessive latency at a node), providing a basis for subsequent adjustments.

[0131] Through geographic clustering and data traffic characteristic-based clustering, as well as fine-grained optimization of intra-cluster topology, the deployment of aggregation nodes is closer to data sources and data consumers, effectively reducing data transmission latency. Simultaneously, a dynamic adjustment mechanism can adjust the number and location of aggregation nodes based on real-time traffic and network conditions, further reducing potential congestion and latency. Furthermore, precise demand analysis and node clustering ensure that the computing and storage resources of aggregation nodes match data traffic demands. This dynamic adjustment strategy allows the network to flexibly adjust resource allocation based on actual traffic conditions, thereby optimizing the overall network throughput and ensuring efficient data processing even during peak traffic periods.

[0132] Furthermore, the embodiments of this disclosure can also encrypt all data passing through the aggregation node. By encrypting all data passing through the aggregation node and implementing a secure key management strategy, the security of data transmission can be effectively guaranteed. In addition, the redundant design and disaster recovery plan for the aggregation node can provide a higher level of data security for the network. Simultaneously, the implementation of redundancy design and disaster recovery plan, along with a dynamic adjustment mechanism based on real-time monitoring data, can jointly improve the reliability and availability of the network. Even in the face of node failures or other unforeseen circumstances, the system can recover quickly, minimizing the impact on data processing and transmission.

[0133] In embodiments of this disclosure, such as Figure 2 The method also includes adjusting the deployment location and number of aggregation nodes based on real-time monitoring data.

[0134] During real-time monitoring of network status, if abnormal metrics are detected, the deployment location and number of aggregation nodes can be dynamically adjusted. For example, if monitoring reveals that "the CPU utilization of the aggregation node in cluster A reaches 90% due to a sudden increase in traffic (insufficient performance)," then one aggregation node can be temporarily added to share the load (adjusting the number). Alternatively, if "the latency from the original aggregation node to the edge node in cluster B suddenly increases from 20ms to 100ms (poor location)," then the aggregation node can be switched to another node within the cluster with lower latency (adjusting the location).

[0135] In some embodiments, network performance can be evaluated periodically, and network topology and aggregation strategies can be adjusted based on the evaluation results. For example, the overall network performance (such as average latency, resource utilization, and data transmission success rate) can be reviewed on a daily / weekly / monthly basis, compared with optimization targets, and the trends of monitoring data and the effectiveness of aggregation strategies can be analyzed to achieve periodic optimization of network status.

[0136] Figure 7 A block diagram of a network device according to an embodiment of the present disclosure is schematically illustrated; like Figure 7As shown, the network device 700 of this embodiment includes a memory 701 and a processor 702.

[0137] Memory 701 is used to store computer-readable instructions. Processor 702 is used to execute the computer-readable instructions, causing the network device to perform the aggregation node deployment method.

[0138] Figure 8 A block diagram illustrating a non-transitory computer-readable storage medium according to an embodiment of the present disclosure is shown schematically. like Figure 8 As shown, a non-transitory computer-readable storage medium 800 of this disclosure embodiment is used to store computer-readable instructions 801, which, when executed by a processor, cause the processor to perform the aggregation node deployment method as described above.

[0139] Figure 9 A block diagram illustrating a computer program product according to an embodiment of the present disclosure is shown schematically. like Figure 9 As shown, a computer program product 900 according to an embodiment of this disclosure includes a computer program 901, which, when executed by a processor, implements the aggregation node deployment method as described above.

[0140] The above description, with reference to the accompanying drawings, illustrates a method for deploying aggregation nodes to meet peak data aggregation demands according to embodiments of the present disclosure. This method comprehensively utilizes various techniques, including data source analysis, network topology mapping, traffic pattern analysis, node clustering, and intra-cluster topology optimization. By scientifically analyzing and planning network resources, data processing efficiency can be effectively improved, latency reduced, network throughput increased, and data transmission security and network high availability ensured.

[0141] The basic principles of this disclosure have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this disclosure are merely examples and not limitations, and should not be considered as essential features of each embodiment of this disclosure. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the scope of this disclosure to the necessity of employing the aforementioned specific details for implementation.

[0142] The block diagrams of devices, apparatuses, devices, and systems disclosed herein are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context clearly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.

[0143] Additionally, as used herein, the "or" used in a list of items beginning with "at least one" indicates a separate list, such that a list of, for example, "at least one of A, B, or C" means A or B or C, or AB or AC or BC, or ABC (i.e., A and B and C). Furthermore, the word "exemplary" does not imply that the described example is preferred or better than other examples.

[0144] It should also be noted that in the systems and methods of this disclosure, the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered as equivalent solutions to this disclosure.

[0145] Various changes, substitutions, and modifications can be made to the technology described herein without departing from the teachings defined by the appended claims. Furthermore, the scope of the claims of this disclosure is not limited to the specific aspects of the processes, machines, manufactures, events, means, methods, and actions described above. Currently existing or later-developed processes, machines, manufactures, events, means, methods, or actions that perform substantially the same function or achieve substantially the same result as the corresponding aspects described above can be utilized. Therefore, the appended claims include such processes, machines, manufactures, events, means, methods, or actions within their scope.

[0146] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this disclosure. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of this disclosure. Therefore, this disclosure is not intended to be limited to the aspects shown herein, but rather to be carried out within the widest scope consistent with the principles and novel features disclosed herein.

[0147] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this disclosure to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations therein.

Claims

1. A method for deploying aggregation nodes to meet peak data aggregation demands, characterized in that, The method includes: Based on the predicted peak time and predicted peak demand from at least one data source, determine the deployment location and number of at least one aggregation node; The nodes are divided into multiple clusters, each cluster including at least one aggregation node and a data source node; and Based on the data traffic characteristics within each node cluster, the optimal data transmission path is determined in order to configure computing resources for the aggregation node.

2. The aggregation node deployment method according to claim 1, characterized in that, Also includes: Obtain multiple historical data entries from at least one of the aforementioned data sources; Based on the aforementioned multiple historical data, the historical peak time and historical peak demand for each of the data sources are determined; as well as Based on the historical peak time and the historical peak demand, the predicted peak time and the predicted peak demand for each of the data sources are determined.

3. The aggregation node deployment method according to claim 2, characterized in that, The data source includes at least one or more of the following: cloud service data source in the operator field, communication value-added service data source, capability open platform data source, and support service data source.

4. The aggregation node deployment method according to claim 3, characterized in that, The step of determining the historical peak time and historical peak demand for each data source based on the multiple historical data sets includes: Determine the preset time range and time granularity; Within the preset time range, aggregate the multiple historical data entries according to the time granularity; and Based on the aggregated historical data, determine the historical peak time and historical peak demand within the preset time range for each data source.

5. The aggregation node deployment method according to claim 4, characterized in that, The step of determining the historical peak time and historical peak demand within the preset time range for each data source based on the aggregated historical data includes: Determine the data aggregation demand within multiple time windows, each time window having the same time length, and the time windows are determined by dividing the preset time range; Compare the data aggregation demand in the current time window with the data aggregation demand in the previous time window; and Based on the comparison results, the historical peak time and historical peak demand within the preset time range for each of the data sources are determined.

6. The aggregation node deployment method according to claim 5, characterized in that, Determining the historical peak time and historical peak demand within the preset time range for each of the data sources includes: in response to the data aggregation demand within the current time window being greater than the data aggregation demand within the previous time window, determining the current time window as a peak time period, wherein the historical peak time is determined by the sum of the times corresponding to consecutive peak time periods, and the historical peak demand is determined by the sum of the data aggregation demand corresponding to consecutive peak time periods.

7. The aggregation node deployment method according to claim 4, characterized in that, The determination of the deployment location and number of at least one aggregation node based on the predicted peak time and predicted peak demand from at least one data source includes: Determine multiple convergence time periods; Determine the sum of the predicted data volume from multiple data sources within each said convergence time period, wherein the predicted data volume is determined at least based on the historical data predictions; and The deployment location range and number of at least one aggregation node are determined based on the largest sum of the sums of the multiple data volumes.

8. The aggregation node deployment method according to claim 7, characterized in that, The step of determining the deployment location and number of at least one aggregation node based on the largest sum of the sums of multiple data volumes includes: Based on the geographical location information and network conditions of the data source, the deployment location of the aggregation node is determined; and The number of aggregation nodes is determined based on the maximum sum of the data volume and the data processing capacity of the aggregation nodes.

9. The aggregation node deployment method according to claim 8, characterized in that, Also includes: The aggregation node is evaluated based on at least one or more of the following: bandwidth information, latency information, packet loss rate information, and throughput information.

10. The aggregation node deployment method according to claim 1, characterized in that, The division into multiple node clusters includes: Based on the data source and the aggregation node, a network topology is determined, the network topology including at least one aggregation node and multiple data source nodes; and Based on the geographic location information of the data source, the predicted peak time, the predicted peak demand, and the deployment location of the aggregation node, the multiple aggregation nodes and the data source nodes are divided into multiple node clusters, and each node cluster includes at least one aggregation node and multiple data source nodes.

11. The aggregation node deployment method according to claim 10, characterized in that, The process of dividing the multiple aggregation nodes and data source nodes into multiple node clusters includes: Based on the geographic location information, the multiple aggregation nodes and the multiple data source nodes are divided into multiple initial node clusters, each initial node cluster including at least one aggregation node and multiple data source nodes; and Using the data generation and consumption of the aggregation node and the data source node, the initial node clusters are further divided to obtain multiple node clusters.

12. The aggregation node deployment method according to claim 11, characterized in that, Based on the geographic location information, the multiple aggregation nodes and the data source nodes are divided into multiple initial node clusters, including: Determine the number of clusters and the initial cluster centers; Calculate the distance from each of the aggregation nodes and the data source nodes to the initial cluster center; and Based on the distance to the initial cluster center, each of the aggregation nodes and the data source nodes is assigned to a different initial node cluster.

13. The aggregation node deployment method according to claim 12, characterized in that, The step of determining the optimal data transmission path based on the data traffic characteristics within each node cluster, in order to configure computing resources for the aggregation node, includes: Based on data traffic characteristics, a minimum spanning tree is constructed for each node cluster; and Based on the minimum spanning tree, an optimal data transmission path is determined for each node cluster, and the optimal data transmission path includes at least one of the aggregation nodes.

14. The aggregation node deployment method according to claim 13, characterized in that, Also includes: Based on the optimal data transmission path, computing and storage resources are allocated to the aggregation nodes included in the optimal data transmission path.

15. The aggregation node deployment method according to claim 1, characterized in that, Also includes: Acquire real-time monitoring data, which includes at least latency rate, throughput, packet loss rate, and peak demand changes of the data source.

16. The aggregation node deployment method according to claim 15, characterized in that, Also includes: Based on the real-time monitoring data, the deployment location and number of aggregation nodes are adjusted.

17. A network device, characterized in that, include: Memory, used to store computer-readable instructions; as well as A processor for executing the computer-readable instructions, causing the network device to perform the aggregation node deployment method as described in any one of claims 1 to 16.

18. A non-transitory computer-readable storage medium for storing computer-readable instructions, characterized in that, When computer-readable instructions are executed by a processor, the processor performs the aggregation node deployment method as described in any one of claims 1 to 16.

19. A computer program product, comprising a computer program, characterized in that, When a computer program is executed by a processor, it implements the aggregation node deployment method as described in any one of claims 1 to 16.

Images