Call detail data collection method, device, medium and program product

Abstract

This invention provides a method, device, medium, and program product for collecting call detail record (CDR) data. The method includes: using a constructed intelligent decision-making model for collection strategies to dynamically select the optimal CDR data collection mode for each PBX within a predetermined decision period t; wherein the CDR data collection modes include: a periodic batch collection mode and a real-time streaming collection mode; the objective function of the intelligent decision-making model for collection strategies is used to minimize the total cost, which includes the cost of periodic batch collection and the cost of real-time streaming collection; the intelligent decision-making model for collection strategies predicts the CDR data generation rate in the periodic batch collection cost using a pre-trained LSTM network prediction model; utilizing the above technical solution, by integrating the LSTM prediction and multi-factor cost model decision layer, a dynamically adaptive CDR data collection strategy is achieved, significantly improving resource utilization efficiency and data real-time performance.

Description

Technical Field

[0001] This invention relates to the fields of communication network data processing and cloud computing technology, and in particular to a method, device, medium and program product for collecting call detail data. Background Technology

[0002] Large enterprises or communications service providers typically deploy hundreds or thousands of PBX devices globally. These devices continuously generate Call Detail Record (CDR) data for every call. A unified billing and analytics platform needs to collect CDR data from all these PBXs.

[0003] Currently, existing methods for acquiring CDR data mainly face the following problems: The limitations of static data collection mode mean that platforms typically pre-set a fixed collection mode for all PBXs, such as collecting data every 5 minutes or enabling real-time push for all. This "one-size-fits-all" strategy leads to uneven resource allocation: for highly active PBXs, periodic collection may result in excessive data latency, affecting the real-time performance of billing; for inactive PBXs, real-time push will cause long-term idleness and waste of network and computing resources.

[0004] In summary, existing technologies lack the ability to dynamically adjust strategies based on the real-time status of the PBX and business needs. Summary of the Invention

[0005] The embodiments of the present invention provide a method, device, medium and program product for collecting call detail record data, so as to dynamically select the optimal CDR data collection mode for each PBX within the decision cycle.

[0006] To achieve the above objectives, on the one hand, a method for collecting call detail data is provided for collecting call detail data from multiple PBXs, including: using a constructed intelligent decision model of the collection strategy to dynamically select the optimal call detail data collection mode for each PBX within a predetermined decision period t; wherein, the call detail data collection mode includes: periodic batch collection mode and real-time streaming collection mode. The decision variables of the intelligent decision-making model for the data acquisition strategy are: , =0 means Select the periodic batch data collection mode within the decision cycle t. =1 means Select the real-time streaming acquisition mode within the decision cycle t; The objective function of the intelligent decision-making model for the data acquisition strategy is:

[0007] The objective function is used to minimize the total cost, where N is the total number of PBXs; express The real-time streaming acquisition cost consumed by collecting call detail data using the real-time streaming acquisition mode within the decision cycle t; express The cost of periodic batch collection of call detail data using the periodic batch collection mode within the decision cycle t; Indicates the first One PBX;

[0008]

[0009] in, This indicates the time from the data collection platform to the decision-making period t. The average round-trip latency of the network; express The keyness coefficient of the data, its value is based on The criticality of the data is pre-defined; For predictions made using a pre-trained LSTM network prediction model The predicted rate of generation of detailed call data within the decision period t; for The number of call detail records that have been generated but not yet successfully collected within the decision period t; , , and These are predetermined cost weighting coefficients.

[0010] Preferably, in the acquisition method, wherein, The inputs to the LSTM network prediction model include: The generation rate sequence of historical call detail data over L consecutive time points prior to decision period t. L is a predetermined number; and, is the time characteristic corresponding to the historical call detail data generation rate sequence. The output of the LSTM network prediction model includes: The predicted generation rate of call detail data at one or more time points within the decision period t.

[0011] Preferably, in the acquisition method, the time feature includes one or more of the following: The hour of a day takes the value of an integer between 0 and 23. The day of the week, which takes the value of an integer between 0 and 6; Whether it is a holiday or not, the value can be 0 or 1, where 0 means it is not a holiday and 1 means it is a holiday.

[0012] Preferably, the data acquisition method further includes: The task is to receive requests and export all call detail records for the selected PBX cluster within a selected time period. The predictive sharding algorithm based on spatiotemporal correlation aims to minimize the total execution time of the task and decomposes the task into multiple sub-tasks that are executed in parallel. Among them, with the goal of minimizing the sum of squared network delays between PBXs within the group, the selected PBX cluster is divided into K1 groups using the K-Means clustering algorithm. Each sub-task is to obtain all call detail records of a PBX group in a time slice.

[0013] Preferably, the acquisition method further includes the step of decomposing the task into multiple sub-tasks to be executed in parallel, and performing the following processing on each of the K1 groups: Aggregate the historical total call detail data generation rate of each PBX in the group over a predetermined period of time, and form the aggregated rate time series of the group; For the aggregation rate time series of this group, a change point detection algorithm is used to find M change points within a predetermined time period, and the predetermined time period is divided into M+1 intervals using the M change points; Based on the preset data volume threshold and the estimated data volume of each interval, the M+1 intervals are merged or subdivided to form multiple time slices, so that the difference between the estimated data volumes of different time slices in the multiple time slices is less than the preset data volume threshold.

[0014] Preferably, in the acquisition method, wherein, The estimated data volume for each interval = the average of the aggregated call detail record generation rates corresponding to the two endpoints of the interval in the aggregated rate time series × the time length of the interval.

[0015] Preferably, the data acquisition method further includes: Each subtask is assigned to an idle collector for parallel execution; each collector establishes a connection with the corresponding target PBX group and pulls data according to the specified time slice.

[0016] On the other hand, an electronic device is also provided, including a memory and a processor, the memory storing at least one program, the at least one program being executed by the processor to implement the steps of any of the acquisition methods described above.

[0017] In another aspect, a computer-readable storage medium is also provided, wherein at least one program is stored therein, the at least one program being executed by a processor to implement the steps of any of the acquisition methods described above.

[0018] In another aspect, a computer program product is also provided, comprising a computer program, characterized in that, when the computer program is executed by a processor, it implements the steps of any of the acquisition methods described above.

[0019] The above technical solution has the following technical effects: This invention utilizes a constructed intelligent decision-making model for acquisition strategies to dynamically select the optimal CDR data acquisition mode for each PBX within a predetermined decision period t. This intelligent decision-making model constructs an objective function using a real-time streaming acquisition cost function and a periodic batch acquisition cost function. Furthermore, it predicts the CDR data generation rate in the periodic batch acquisition cost function using a pre-trained LSTM network prediction model. By integrating the LSTM prediction and multi-factor cost model into the decision layer, a dynamically adaptive CDR data acquisition strategy is achieved, significantly improving resource utilization efficiency and data real-time performance.

[0020] In a further embodiment, after receiving the task of exporting all CDR data of the selected PBX cluster within a selected time period, i.e., receiving the task of batch exporting CDR data, a predictive sharding algorithm based on spatiotemporal correlation is implemented to realize a batch task divide-and-conquer optimization model. This batch task divide-and-conquer optimization model combines non-uniform time sharding of network affinity clustering and change point detection to realize intelligent parallelization of batch export tasks. It fully considers the spatiotemporal characteristics of data, solves the problem of uneven load caused by simply sharding by PBX or by time, greatly shortens the response time of large data volume queries, and can perform efficient batch processing. Attached Figure Description

[0021] Figure 1 This is a schematic diagram illustrating the principle topology of a method for collecting call detail data according to an embodiment of the present invention. Detailed Implementation

[0022] To further illustrate the various embodiments, the present invention provides accompanying drawings. These drawings are part of the disclosure of the present invention, primarily used to illustrate the embodiments and to explain the operating principles of the embodiments in conjunction with the relevant descriptions in the specification. With reference to these drawings, those skilled in the art should be able to understand other possible implementations and the advantages of the present invention. Components in the drawings are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0023] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments.

[0024] Example 1: An embodiment of the present invention provides a method for collecting call detail data, which includes: using a constructed intelligent decision-making model for the collection strategy to dynamically select the optimal call detail data collection mode for each PBX within a predetermined decision period t; wherein, the call detail data collection mode includes: a periodic batch collection mode and a real-time streaming collection mode. The decision variables of the intelligent decision-making model for data collection strategy are: , =0 means Select the periodic batch data collection mode within the decision cycle t. =1 means Select the real-time streaming acquisition mode within the decision cycle t; The objective function of the intelligent decision-making model for data collection strategy is:

[0025] The objective function is used to minimize the total cost, where N is the total number of PBXs; express The real-time streaming acquisition cost consumed by collecting call detail data using the real-time streaming acquisition mode within the decision cycle t; express The cost of periodic batch collection of call detail data using the periodic batch collection mode within the decision cycle t; This is the cost function for real-time streaming acquisition; this function quantifies the cost of... The system resources consumed and the business risks associated with using real-time streaming acquisition mode within window t; This is a cost function for periodic batch data collection; it quantifies the costs incurred due to data latency and backlog when using batch collection mode.

[0026]

[0027]

[0028] in: This is a network latency metric used to represent the latency from the data acquisition platform to the destination network within a decision period t. The average round-trip latency of the network; in one specific implementation, its unit is milliseconds, which is measured by continuously sending ICMP or specific TCP / UDP probe packets. The larger the value, the higher the cost of ensuring the real-time performance of streaming transmission, so it is positively correlated with cost.

[0029] express The keyness coefficient of the data, its value is based on The criticality of the data is pre-defined; in one specific implementation, it is a predefined constant based on business rules, with a value range of [0, 1]. An exemplary assignment logic is: the PBX used for postpaid billing is set to 0.7, the PBX used for prepaid billing is set to 0.9, which means that extremely high real-time performance is required, and the PBX used only for internal office analysis is set to 0.3. The higher the value, the more important the data, and the more inclined to use a real-time streaming acquisition mode. The item is positively correlated with cost.

[0030] For predictions made using a pre-trained LSTM network prediction model The predicted rate of call detail record (CDR) data generation is the value within the decision period t. In one specific implementation, the unit of the CDR generation rate prediction is calls per second. This value characterizes... The activity level; and The item reflects the inactivity cost: the slower the generation rate, the longer it takes to collect a certain amount of data, and the higher the data latency may be, thus the higher the cost. This represents the amount of backlogged data, expressed in rows. This value directly reflects the severity of the data collection delay; the more data backlogged, the higher the cost.

[0031] for The number of call detail records that have been generated but not yet successfully collected within the decision period t; , , and These are predetermined cost weighting coefficients.

[0032] In one specific implementation, , A pre-defined constant used to balance the proportion of network conditions and data criticality in the total cost, typically satisfying... For example, if more attention is paid to network conditions, it can be set... =0.7, =0.3. and This is a pre-defined constant used to balance the effects of generation rate and backlog. For example, if data freshness is of greater concern, the value of δ can be set to be greater than γ; typically, it satisfies... .

[0033] Example 2: This embodiment of the invention is implemented based on Embodiment 1. The inputs to the LSTM network prediction model include: The generation rate sequence of historical call detail data over L consecutive time points prior to decision period t. L is a predetermined number; and, represents the time characteristics corresponding to the historical call detail data generation rate sequence; the output of the LSTM network prediction model includes: Predicted generation rate of call detail data at one or more time points within the decision period t. The LSTM network prediction model of this invention will be described in detail below.

[0034] Model Selection and Principles: Long Short-Term Memory (LSTM) network is used as the core prediction engine. LSTM is a special type of recurrent neural network (RNN). Through its internal gating mechanisms, such as input gate, forget gate, and output gate, it can effectively learn and capture long-term dependencies, complex nonlinear trends, and periodic patterns such as daily / weekly / monthly in time series data. This is crucial for accurately predicting the CDR generation rate, which has obvious business patterns.

[0035] Input feature engineering: Main feature sequence: Historical CDR generation rate sequence over the past L consecutive time points In one specific implementation, L=168, representing the past week in hours; here, "past" refers to the period before decision period t. Time features: These serve as auxiliary inputs and include, but are not limited to: the hour of the day (0-23, representing any integer from 0 to 23, indicating the time of day the data was collected); the day of the week (0-6, representing any integer from 0 to 6, indicating the day of the week the data was collected); and whether it is a holiday or public holiday (0 or 1, where 0 indicates a non-holiday and 1 indicates a holiday, indicating whether the data collection date falls on a holiday). These features help the model learn cyclical patterns.

[0036] Output and Model Update: The output of the LSTM network prediction model is the predicted CDR generation rate at one or more future time points. LSTM network prediction models are trained offline using historical data and can be fine-tuned online periodically, such as daily, with the latest data to adapt to changes in data distribution.

[0037] The CDR generation rate sequence prior to time t can be used to predict the CDR generation rate for period t, including the CDR generation rate at one or more time points within period t.

[0038] Example 3: The inventors of this application, while researching CDR data acquisition methods, discovered that existing acquisition methods, besides having the limitations of static acquisition modes, also suffer from performance bottlenecks in exporting massive amounts of historical data, lacking effective means for intelligent parallel decomposition and scheduling of batch tasks. Specifically, when users, such as finance or analysis departments, need to export CDR data for all PBXs in a certain region over a long period, such as one month, for reconciliation or analysis, the traditional serial query method is extremely time-consuming. This is because the task is simply decomposed into serial sub-tasks of "by PBX * by day," without considering the differences in data volume between different PBXs, network conditions, and the inherent time patterns of data generation, resulting in a significant "weakest link" effect and low overall execution efficiency. Existing technologies lack effective means for intelligent parallel decomposition and scheduling of batch tasks.

[0039] Based on this, the acquisition method of this embodiment of the invention also provides a CDR data acquisition scheme for efficiently parallelizing the decomposition of batch tasks.

[0040] The data acquisition method of this embodiment of the invention further includes the following steps: The task is to receive and export all call detail records for a selected PBX cluster within a selected time period. The selected PBX cluster includes multiple PBXs; in one specific implementation, the selected PBX cluster is PBX cluster S; the selected time period is... The corresponding task is: ; The predictive sharding algorithm based on spatiotemporal correlation aims to minimize the total execution time of the task and decomposes the task into multiple sub-tasks that are executed in parallel. Among them, the K-Means clustering algorithm is used to divide the selected PBX cluster into K1 groups with the goal of minimizing the sum of squared network delays between PBXs within the group. Each sub-task is to obtain all call detail records of a PBX group in a time slice.

[0041] The total execution time of the minimized task described above is the latest completion time among all subtasks.

[0042] In one specific implementation, a batch task divide-and-conquer optimization model is constructed to efficiently and parallelize the decomposition of batch tasks. The following is a detailed explanation of the batch task divide-and-conquer optimization model.

[0043] Optimization objective: Minimize the total execution time, i.e., the latest completion time among all subtasks. .

[0044] Decomposition Algorithm: Predictive Fragmentation Based on Spatiotemporal Correlation Spatial partitioning (K-Means clustering based on network latency): Metric: Network latency between PBXs As a distance metric.

[0045] Algorithm Execution: The K-Means clustering algorithm is used to divide the PBX cluster S into K1 groups, where K1 is a predetermined number. The goal of the algorithm is to minimize the sum of squared network delays between PBXs within a group, i.e., to minimize... This ensures that PBX networks within the same group have similar conditions, allowing for efficient processing by the same collector instance and reducing network hopping overhead. Time slicing, based on PELT change point detection for non-uniform segmentation: Algorithm execution: For each PBX group obtained after clustering, aggregate its historical total CDR generation rate, such as the historical total CDR generation rate within a predetermined time period, to form an aggregated rate time series. Using change point detection algorithms such as PELT (Pruned Exact Linear Time) within a time period... Find M variables within the inner circle The PELT algorithm efficiently finds the change point that minimizes the piecewise linear fitting cost function, thus dividing the entire time period into M+1 intervals, ensuring a relatively stable CDR generation rate within each interval. Piecewise equalization: Calculates the cost function for each interval. Estimated data volume: ( ), This represents the estimated data volume for the m-th interval, which is also the estimated CDR data volume. and They represent The corresponding time point and The values ​​at each point are then determined. Then, based on a preset data volume threshold, these intervals are merged or further subdivided to ultimately form a series of time slices. 'j' represents any time slice. The goal is to make the estimated data volume for each time slice as balanced as possible, avoiding the "weakest link" effect in parallel tasks. Subtask definition and scheduling: A subtask Defined as: acquiring all CDR data of a specific PBX group in a specific time slice, where k is the kth subtask.

[0046] General Task Decomposed into Sub-tasks .

[0047] In the implementation, a pool of collector threads is maintained to enable the parallel execution of subtasks. The scheduler assigns these subtasks to idle collectors for parallel execution. Each collector independently establishes a connection with the target PBX group and pulls data according to a specified time slice.

[0048] The data acquisition method in this embodiment of the invention achieves intelligent parallelization of batch export tasks by combining network affinity clustering and non-uniform time partitioning based on change point detection. This method fully considers the spatiotemporal characteristics of data, solves the load unevenness problem caused by simply partitioning by PBX or by time, and greatly shortens the response time for large-scale data queries.

[0049] Example 4: This embodiment of the present invention is a combination of the above embodiments one to three. Figure 1 This is a schematic diagram of the principle topology of this embodiment of the invention. Figure 1 In its specific implementation, the intelligent acquisition platform uses the acquisition method of this invention to acquire CDR data from data sources such as globally distributed PBX clusters. The intelligent acquisition platform, or CDR data acquisition platform, includes: an intelligent scheduling engine, used to dynamically select the optimal acquisition mode using a constructed intelligent decision model for the acquisition strategy; this intelligent scheduling engine uses an LSTM network prediction model to predict the CDR generation rate; and a task divide-and-conquer optimizer, used to decompose tasks into multiple parallel subtasks based on a predictive sharding algorithm based on spatiotemporal correlation, to achieve efficient batch processing when there are a large number of pull requests at the data consumption end.

[0050] In the embodiments of this invention, the LSTM deep learning model is used for accurate prediction of CDR generation patterns, a statistical method such as PELT for change point detection is used to identify the inherent rhythm of data generation, and a combination optimization such as K-Means clustering is used for network resource optimization. This achieves deep integration of multidisciplinary technologies and provides a new and efficient solution to the bottleneck of traditional data acquisition.

[0051] Example 5: The present invention also provides an electronic device comprising a processor, a memory, a bus, and a computer program stored in the memory and executable on the processor. The processor includes one or more processing cores. The memory is connected to the processor via the bus and is used to store program instructions. When the processor executes the computer program, it implements the steps in the method embodiments described in Embodiment 1 of the present invention. Preferably, the above-described electronic device is a server for data acquisition or scheduling.

[0052] Furthermore, as an executable solution, the electronic device can be a computer unit, which can be a desktop computer, laptop, handheld computer, cloud server, or other computing device. The computer unit may include, but is not limited to, a processor and memory. Those skilled in the art will understand that the above-described structure of the computer unit is merely an example and does not constitute a limitation on the computer unit. It may include more or fewer components, or combine certain components, or use different components. For example, the computer unit may also include input / output devices, network access devices, buses, etc., and this embodiment of the invention does not limit this.

[0053] Furthermore, as an executable solution, the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc. The processor is the control center of the computer unit, connecting various parts of the entire computer unit via various interfaces and lines.

[0054] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the computer unit by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a function; the data storage area may store data created based on the use of the mobile phone, etc. In addition, the memory may include high-speed random access memory and may also include non-volatile memory, such as hard disk, RAM, plug-in hard disk, smart media card (SMC), secure digital card (SD card), flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0055] Example 6: The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the method described in the embodiments of the present invention.

[0056] If the modules / units integrated in the computer unit are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), and software distribution media, etc. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction.

[0057] Example 7: The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps described above.

[0058] Although the invention has been specifically shown and described in conjunction with preferred embodiments, those skilled in the art should understand that various changes in form and detail may be made to the invention without departing from the spirit and scope of the invention as defined in the appended claims, all of which shall be within the scope of protection of the invention.

Claims

1. A method for collecting call detail data, used to collect call detail data from multiple PBXs, characterized in that, include: The constructed intelligent decision-making model for the collection strategy dynamically selects the optimal call detail data collection mode for each PBX within a predetermined decision period t; wherein, the call detail data collection modes include: periodic batch collection mode and real-time streaming collection mode. The decision variables of the intelligent decision-making model for the data acquisition strategy are: , =0 means Select the periodic batch data collection mode within the decision cycle t. =1 means Select the real-time streaming acquisition mode within the decision cycle t; The objective function of the intelligent decision-making model for the data acquisition strategy is: The objective function is used to minimize the total cost, where N is the total number of PBXs; express The real-time streaming acquisition cost consumed by collecting call detail data using the real-time streaming acquisition mode within the decision cycle t; express The cost of periodic batch collection of call detail data using the periodic batch collection mode within the decision cycle t; Indicates the first One PBX; in, This indicates the time from the data collection platform to the decision-making period t. The average round-trip latency of the network; express The keyness coefficient of the data, its value is based on The criticality of the data is pre-defined; For predictions made using a pre-trained LSTM network prediction model The predicted rate of generation of detailed call data within the decision period t; for The number of call detail records that have been generated but not yet successfully collected within the decision period t; , , and These are predetermined cost weighting coefficients.

2. The data acquisition method according to claim 1, characterized in that, The inputs to the LSTM network prediction model include: The generation rate sequence of historical call detail data over L consecutive time points prior to decision period t. L is a predetermined number; and, is the time characteristic corresponding to the historical call detail data generation rate sequence. The output of the LSTM network prediction model includes: The predicted generation rate of call detail data at one or more time points within the decision period t.

3. The data acquisition method according to claim 2, characterized in that, The time feature includes one or more of the following: The hour of a day takes the value of an integer between 0 and 23. The day of the week, which takes the value of an integer between 0 and 6; Whether it is a holiday or not, the value can be 0 or 1, where 0 means it is not a holiday and 1 means it is a holiday.

4. The data acquisition method according to claim 1, characterized in that, Also includes: The task is to receive requests and export all call detail data for the selected PBX cluster within a selected time period. The predictive sharding algorithm based on spatiotemporal correlation aims to minimize the total execution time of the task and decomposes the task into multiple sub-tasks that are executed in parallel. Among them, with the goal of minimizing the sum of squared network delays between PBXs within the group, the selected PBX cluster is divided into K1 groups using the K-Means clustering algorithm. Each sub-task is to obtain all call detail records of a PBX group in a time slice.

5. The data acquisition method according to claim 4, characterized in that, The step of decomposing the task into multiple subtasks to be executed in parallel also includes processing each of the K1 groups as follows: Aggregate the historical total call detail data generation rate of each PBX in the group over a predetermined period of time, and form the aggregated rate time series of the group; For the aggregation rate time series of this group, a change point detection algorithm is used to find M change points within a predetermined time period, and the predetermined time period is divided into M+1 intervals using the M change points; Based on the preset data volume threshold and the estimated data volume of each interval, the M+1 intervals are merged or subdivided to form multiple time slices, so that the difference between the estimated data volumes of different time slices in the multiple time slices is less than the preset data volume threshold.

6. The data acquisition method according to claim 5, characterized in that, The estimated data volume for each interval = the average of the aggregated call detail record generation rates corresponding to the two endpoints of the interval in the aggregated rate time series × the time length of the interval.

7. The data acquisition method according to claim 4, characterized in that, Also includes: Each subtask is assigned to an idle collector for parallel execution; each collector establishes a connection with the corresponding target PBX group and pulls data according to the specified time slice.

8. An electronic device, characterized in that, It includes a memory and a processor, the memory storing at least one program, the at least one program being executed by the processor to implement the steps of the acquisition method as described in any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that, The storage medium stores at least one program, which is executed by a processor to implement the steps of the acquisition method as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the acquisition method as described in any one of claims 1 to 7.

Images