Network element performance prediction method and device, equipment, storage medium and program product
By constructing a multi-dimensional knowledge graph and a time series prediction model, the integrated network element objects are identified, which solves the problem of low operation and maintenance efficiency in existing communication networks, realizes the prediction and early warning of network element performance, and improves network operation and maintenance efficiency and overall performance.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-11
- Publication Date
- 2026-03-27
AI Technical Summary
The existing communication network has low operation and maintenance efficiency, poor overall network performance, and difficulty in effectively predicting and preventing network element failures, which affects network stability.
By constructing a multi-dimensional knowledge graph and identifying integrated network element objects, combined with a time series prediction model, network element performance can be predicted and early warnings can be provided, thereby improving network operation and maintenance efficiency.
It enables early prediction and warning of network element performance, improves network operation and maintenance efficiency, and enhances overall network performance and stability.
Smart Images
Figure CN121750501A_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the field of communication technology, and in particular to a method, apparatus, device, storage medium, and program product for predicting network element performance. Background Technology
[0002] With the continuous evolution and convergence of 2G / 4G / 5G networks, the core network is gradually transforming towards a virtualized and cloud-based architecture, placing higher demands on the automation and intelligence of network operation and maintenance by telecommunications operators. As the core hub carrying critical services such as user access, mobility management, and session control, the stability of the core network's network element performance directly affects the overall network service quality. Therefore, real-time monitoring and analysis of core network element performance data has become an important means for operators to ensure reliable network operation.
[0003] Current mainstream telecommunications operator network monitoring systems mainly rely on collected performance indicator data and make alarm judgments based on preset multi-level threshold rules. When a certain performance indicator (such as CPU utilization, registration success rate, signaling processing latency, etc.) exceeds or falls below the set threshold, the system automatically generates an alarm event of the corresponding level, and maintenance personnel carry out troubleshooting and handling accordingly.
[0004] Currently, network maintenance is carried out after a network element fails, resulting in low network maintenance efficiency and impacting overall network performance. Summary of the Invention
[0005] This disclosure provides a method, apparatus, device, storage medium, and program product for predicting network element performance, in order to at least solve the problems of low network operation and maintenance efficiency and poor overall network performance.
[0006] The technical solution disclosed herein is as follows: This disclosure provides a method for predicting network element performance, including: Based on the network element names of each network element in the core network, identify and aggregate the converged network elements that carry the functions of multiple generations of core networks, and construct a first-level data model centered on network element objects; Based on the logical topological relationship between network elements, resource pools, data centers and network regions, a second-level spatial resource data model is constructed. By associating the performance index information of each network element in the core network with PIM data index and VIM data index through the network element object, a third-level virtualization resource data model is constructed. A multi-dimensional knowledge graph is constructed based on the first-level data model, the second-level spatial resource data model, and the third-level virtualization resource data model. Based on the historical performance data in the multi-dimensional knowledge graph, a time series prediction model is used to perform univariate probability prediction to obtain network element performance prediction information for future periods.
[0007] Optionally, the performance indicator information includes resource data and performance data, and the method further includes: The performance data files reported by the northbound interface of the core network OMC equipment are parsed to obtain the performance index information of each network element of the core network. The resource data and performance data are stored separately according to the management object class.
[0008] Optionally, the network element performance prediction information is a network element performance prediction value, and the method further includes: Compare the predicted network element performance value of the target network element with the preset alarm threshold; If the predicted performance value of the network element is less than the preset alarm threshold, a performance alarm operation is performed.
[0009] Optionally, the step of performing a performance alarm operation when the predicted network element performance value is less than the preset alarm threshold includes: If the predicted performance value of the network element is less than the preset alarm threshold, the target network element will be marked as an early warning node in the multi-dimensional knowledge graph. Output early warning information for the target network element.
[0010] Optionally, based on the historical performance data in the multi-dimensional knowledge graph, a time series prediction model is used to perform univariate probability prediction to obtain network element performance prediction information for future periods, including: Obtain historical performance data from a multi-dimensional knowledge graph; Based on the historical performance data, construct time series data; Extract lag features from the time series data; The lag features and static covariates are used to form a label for a univariate time series at a given time step; By inputting the label of the univariate time series at the given time step into the time series prediction model, network element performance prediction information for future periods can be obtained.
[0011] Optionally, the converged network element includes at least one of the following: AMF-MME-SGSN, SMF-PGWC-SGWC, UPF-PGWU-SGWU, PCF-PCRF, and UDM-HSS.
[0012] This disclosure also provides a network element performance prediction device, including: The first construction module is used to identify and aggregate converged network elements that carry the functions of multiple generations of core networks based on the network element names of each network element in the core network, and to construct a first-level data model centered on network element objects. The second construction module is used to construct a second-level spatial resource data model based on the logical topological relationship between network elements, resource pools, data centers and network regions. The third construction module is used to associate the performance index information of each network element in the core network with the PIM data index and VIM data index through the network element object, and construct the third-level virtualization resource data model. The fourth construction module is used to construct a multi-dimensional knowledge graph based on the first-level data model, the second-level spatial resource data model, and the third-level virtualization resource data model. The prediction module is used to perform univariate probability prediction using a time series prediction model based on historical performance data in the multi-dimensional knowledge graph, so as to obtain network element performance prediction information for future periods.
[0013] This disclosure also provides an electronic device, including: processor; Memory used to store processor-executable instructions; The processor is configured to execute instructions to implement the steps in the above method.
[0014] This disclosure also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.
[0015] This disclosure also provides a computer program product, including a computer program / instructions, which, when executed by a processor, implement the steps of the method described above.
[0016] The technical solutions provided by the embodiments of this disclosure have at least the following beneficial effects: In some embodiments of this disclosure, based on the network element names of each network element in the core network, converged network elements carrying functions of multiple generations of core networks are identified and aggregated to construct a first-level data model centered on network element objects; based on the logical topological relationship between network elements, resource pools, data centers, and network regions, a second-level spatial resource data model is constructed; through network element objects, the performance index information of each network element in the core network is associated with PIM data indicators and VIM data indicators to construct a third-level virtualization resource data model; the spatial dimension and traditional physical network element relationships are expanded to add performance monitoring data dimensions; based on the first-level data model, the second-level spatial resource data model, and the third-level virtualization resource data model, a multi-dimensional knowledge graph is constructed to intuitively traverse network element objects and their associated objects, as well as upper and lower level indicator data; based on historical performance data in the multi-dimensional knowledge graph, a time series prediction model is used to perform univariate probability prediction to obtain network element performance prediction information for future periods; the model is used to predict network element performance in advance, improving network operation and maintenance efficiency and overall network performance.
[0017] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description
[0018] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure, and are not intended to unduly limit this disclosure.
[0019] Figure 1 A schematic diagram of a network element performance prediction process provided for an exemplary embodiment of this disclosure; Figure 2 A schematic diagram of a bottom-up topology flow of a knowledge graph provided for an exemplary embodiment of this disclosure; Figure 3 A schematic diagram of the structure of a network element performance prediction device provided as an exemplary embodiment of this disclosure; Figure 4 A schematic diagram of the structure of an electronic device provided for an exemplary embodiment of this disclosure. Detailed Implementation
[0020] To enable those skilled in the art to better understand the technical solutions of this disclosure, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings.
[0021] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure.
[0022] It should be noted that the user information involved in this disclosure includes, but is not limited to, user device information and user personal information; the collection, storage, use, processing, transmission, provision and disclosure of user information in this disclosure all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.
[0023] To address the aforementioned technical issues, in some embodiments of this disclosure, based on the network element names of each core network element, converged network elements carrying functions of multiple generations of core networks are identified and aggregated to construct a first-level data model centered on network element objects; based on the logical topological relationship between network elements, resource pools, data centers, and network regions, a second-level spatial resource data model is constructed; through network element objects, the performance index information of each network element in the core network is associated with PIM data indicators and VIM data indicators to construct a third-level virtualization resource data model; the spatial dimension and traditional physical network element relationships are expanded to increase the dimension of performance monitoring data; based on the first-level data model, the second-level spatial resource data model, and the third-level virtualization resource data model, a multi-dimensional knowledge graph is constructed to intuitively traverse network element objects and their associated objects, as well as upper and lower level indicator data; based on historical performance data in the multi-dimensional knowledge graph, a time series prediction model is used to perform univariate probability prediction to obtain network element performance prediction information for future periods; the model is used to predict network element performance in advance, improving network operation and maintenance efficiency and overall network performance.
[0024] The technical solutions provided by the embodiments of this disclosure are described in detail below with reference to the accompanying drawings.
[0025] Figure 1 This is a schematic diagram illustrating a network element performance prediction process, provided as an exemplary embodiment of this disclosure. For example... Figure 1 As shown, the method includes: S101: Based on the network element names of each network element in the core network, identify and aggregate the converged network elements that carry the functions of multiple generations of core networks, and construct a first-level data model centered on network element objects; S102: Construct a second-level spatial resource data model based on the logical topological relationship between network elements, resource pools, data centers, and network regions; S103: By associating the performance index information of each network element in the core network with PIM data index and VIM data index through network element objects, a third-level virtualization resource data model is constructed. S104: Construct a multi-dimensional knowledge graph based on the first-level data model, the second-level spatial resource data model, and the third-level virtualization resource data model; S105: Based on historical performance data in a multi-dimensional knowledge graph, a time series prediction model is used to perform univariate probability prediction to obtain network element performance prediction information for future periods.
[0026] In this embodiment, the subject executing the above method is a terminal device or a server.
[0027] The terminal device includes, but is not limited to, mobile stations (MS), mobile terminals, mobile phones, handsets, and portable equipment. This terminal device can communicate with one or more core networks via a radio access network (RAN). For example, the terminal device can be a mobile phone (or "cellular" phone), a computer with wireless communication capabilities, a computer with wireless transceiver capabilities, a virtual reality (VR) terminal device, an AR terminal device, a wireless terminal in industrial control, a wireless terminal in self-driving, a wireless terminal in remote medical care, a wireless terminal in a smart grid, a wireless terminal in transportation safety, a wireless terminal in a smart city, a wireless terminal in a smart home, etc. The operating systems installed on the terminal device include, but are not limited to, iOS, Android, Windows, Linux, and Mac OS. In different networks, terminals may be called by different names, such as: user equipment, mobile station, user unit, station, cellular phone, personal digital assistant, wireless modem, wireless communication device, handheld device, laptop, cordless phone, wireless local loop station, television, etc. For ease of description, this embodiment will simply refer to it as terminal device.
[0028] In this embodiment, the implementation form of the server is not limited. For example, the server can be a conventional server, a cloud server, a cloud host, a virtual center, or other server devices. The server mainly consists of a processor, hard disk, memory, system bus, and other common computer architecture types.
[0029] In this embodiment, based on the network element names of each core network element, converged network elements carrying functions of multiple generations of core networks are identified and aggregated to construct a first-level data model centered on network element objects. A second-level spatial resource data model is constructed based on the logical topology relationships between network elements, resource pools, data centers, and network regions. A third-level virtualization resource data model is constructed by associating the performance indicators of each core network element with PIM and VIM data indicators through network element objects. Spatial dimensions and traditional physical network element relationships are expanded to add performance monitoring data dimensions. A multi-dimensional knowledge graph is constructed based on the first-level data model, the second-level spatial resource data model, and the third-level virtualization resource data model, allowing for intuitive traversal of network element objects and their associated objects, as well as upper and lower level indicator data. Based on historical performance data in the multi-dimensional knowledge graph, a time series prediction model is used for univariate probability prediction to obtain network element performance prediction information for future periods. The model is used to predict network element performance in advance, improving network operation and maintenance efficiency and overall network performance.
[0030] In one optional embodiment, this disclosure is applicable to large-scale core network performance data monitoring scenarios. Based on the traditional three-layer topology relationship of VNF, PIM, and VIM, it expands the spatial dimension and the traditional physical network element relationship to enrich the dimensions of performance monitoring data; it introduces the concept of knowledge graph, and based on the traditional IMS, EPC, and 5GC topology, it intuitively traverses network element objects and their associated objects, as well as upper and lower level indicator data, through a top-down structured knowledge graph; it introduces the Lag-LLaMA model, and utilizes its decoder-only Transformer architecture advantage to achieve efficient time-series prediction of network element performance indicators.
[0031] To better explain this disclosure, the following explanations are provided for the technical terms used in this disclosure.
[0032] The core network is a key component of a communication network, responsible for managing functions such as data and voice transmission, user registration, authentication, billing, and roaming. With the evolution of communication technologies, the core network has gradually evolved from the circuit-switched and packet-switched architecture of the 2G era to the new cloud-native architecture and service-oriented design of the 5G era. It has introduced new technologies such as NFV (Network Functions Virtualization) and SDN (Software Defined Networking) to support more efficient service-oriented and virtualized architectures.
[0033] NFV virtualization requires the use of large data centers to build the core network. By centrally investing in hardware resources such as servers and deploying virtualized network functions in batches, it can not only improve the quality of service and security of the network, but also effectively reduce costs, improve operation and maintenance efficiency, and provide users with more stable and reliable services.
[0034] To ensure stable network operation and a good user experience, network monitoring and assurance are essential. Traditional network monitoring often focuses on the smallest granular unit, such as network elements, monitoring performance data metrics and thresholds. Multiple alarm thresholds are set for key indicators; once data degradation occurs, alarm information is generated for the corresponding network element. In scenarios where core network equipment was deployed independently, this approach effectively located problematic devices and assisted in timely troubleshooting. However, with the centralization of 5G network deployments and the virtualization of equipment, the causes of data degradation have gradually shifted from network elements to the equipment itself, making faults highly interconnected and impactful. Therefore, it is necessary to explore a more efficient and intuitive method for network fault monitoring, taking into account this current state of network infrastructure.
[0035] This disclosure takes the core network virtualization data architecture as its starting point. From the perspective of OMC northbound interface management, according to the 3GPP protocol, the resource data of each type of network element follows the IOC (Information Object Class) structure. Through the hierarchical management object classes of network elements, their corresponding performance indicators can be analyzed level by level. In the virtualization data architecture, its virtual layer adopts PIM-VIM-VNFM hierarchical management. VNFM (Virtualized Network Functions Manager) is responsible for the management of OMC network elements. Once a problem occurs in any of the above three levels of nodes, its impact will be reflected in the OMC network element performance indicators or monitoring indicators (such as CPU utilization). Therefore, most network element performance management systems also manage PIM, VIM, and VNFM layer devices, and even establish a three-level topology, which is a common solution in the industry.
[0036] Based on this virtualized network element construction structure, analysis of data from 2G / 4G / 5G core network virtualized network elements reveals instances of co-construction of network elements by manufacturers. For example, a 5G core network access and mobility management (AMF) network element object may simultaneously carry the functions and data metrics of a 4G mobility management entity (MME) and a 2G serving GPRS support node (SGSN) network element object—a phenomenon known as network element convergence. Therefore, we consider leveraging this characteristic to connect different northbound resource space object classes, PIM, VIM, VNFM, regions, resource pools, and network element topologies using the same network element object as a unique connecting key. This will construct a core network performance data model, serving as the underlying data for the knowledge graph.
[0037] By using a knowledge graph data model, alarm network elements and their corresponding data indicators and alarm times can be automatically labeled, thereby forming a bottom-up aggregation and correlation analysis. The knowledge graph can be used to find the network element topology to which the alarm belongs, the same-time indicators of other network element objects under related topologies, and the upper-level object entities (such as VIM, PIM, etc.) and even the overall data situation of the same data center, so as to determine the impact range of network element alarms and execute operation and maintenance plans.
[0038] Based on this data model, long-term continuous network element performance data is accumulated to form data samples. Combined with time-series prediction algorithms, training samples are accumulated by analyzing the characteristics of abnormal periods and poor-quality data in the data samples. Using knowledge graph data models and time-series prediction algorithms, network element performance alarm prediction is realized, and data fluctuations of different dimensions such as poor-quality network elements, resource pools, and data centers are predicted and presented in the knowledge graph.
[0039] Therefore, this disclosure, taking into account the characteristics of the data model, introduces the Lag-LLaMA model for prediction training. The Lag-LLaMA model is a basic model for univariate probabilistic time series prediction. By using lagged features as covariates, it can be pre-trained on diverse time series data and demonstrates strong zero-shot generalization ability on unseen datasets.
[0040] In some embodiments of this disclosure, based on the network element names of each network element in the core network, converged network elements carrying functions of multiple generations of core networks are identified and aggregated to construct a first-level data model centered on network element objects; based on the logical topological relationship between network elements, resource pools, data centers, and network regions, a second-level spatial resource data model is constructed; through network element objects, the performance index information of each network element in the core network is associated with PIM data indicators and VIM data indicators to construct a third-level virtualization resource data model; based on the first-level data model, the second-level spatial resource data model, and the third-level virtualization resource data model, a multi-dimensional knowledge graph is constructed.
[0041] In the above embodiments, the performance indicator information includes resource data and performance data. The performance data file reported by the northbound interface of the core network OMC device is parsed to obtain the performance indicator information of each network element in the core network. The resource data and performance data are stored separately according to the management object class. Specifically, core network elements report network element resource and performance data through the northbound interface of the core network OMC device. Network management and maintenance personnel collect and parse the northbound interface performance data file to conduct performance analysis of network element operation services. Specifically, the performance data file reported by the northbound interface is parsed to extract relevant performance indicator information. The parsed data is then classified and stored as resource data and performance data respectively, according to the corresponding management object class.
[0042] The parsed data is stored in a relational database to build a long-term, stable, and continuous data model, ensuring data persistence and consistency. The data model includes key fields such as network element objects, performance metrics, and timestamps, supporting fast data querying and correlation analysis.
[0043] As 2G / 4G / 5G core network elements gradually shift to convergence and backward compatibility modes, 2G / 4G network services are also gradually being carried by virtualized network elements instead of traditional equipment. Based on the core network convergence deployment, information on various converged network elements can be compiled. Converged network elements include at least one of the following: AMF-MME-SGSN, SMF-PGWC-SGWC, UPF-PGWU-SGWU, PCF-PCRF, and UDM-HSS.
[0044] By using network element names (such as userlabel or OMCid), data from different network elements (such as AMF, MME, etc.) can be chained and merged. For a merged network element, performance index data of different core network elements it carries can be queried to obtain the service data carried by that network element. For example, an AMF network element object, such as the network element name APP-AMF001XX-04AXX011, can also carry corresponding MME services in a 4G network and corresponding SGSN services in a 2G network. Therefore, this proposal will use such merged network element objects for data aggregation, construct a multi-dimensional network element performance index model using the same network element name, and label it as the first-level data model. This is to meet the need for simultaneous analysis of data from this network element under various core network architectures.
[0045] In the construction of network cloud virtualization, virtualized network elements have a logical relationship of "network element - resource pool - data center - network region". Specific implementation includes: For a given network element, its performance data for the resource pool to which it belongs, as well as the performance data of other network elements within that resource pool, can be queried based on their logical relationships. For example, for AMF001XX-04AXX011, the performance data for its resource pool can be queried.
[0046] For a given resource pool, its logical relationships can be used to query the data center to which it belongs, as well as the performance data of other resource pools and related network elements under that data center. This is of great significance for analyzing regional network quality fluctuations.
[0047] Data centers can be assigned to network regions, helping monitoring personnel respond quickly and specify solutions when large-scale network failures occur in a province or region. The spatial resource data model from network elements to regions is labeled as the second-level spatial resource data model.
[0048] After network element virtualization, PIM (Physical Infrastructure Manager) and VIM (Virtualized Infrastructure Manager) data are involved. This data includes various performance metrics for servers, disk arrays, distributed block storage clusters, switches, firewalls, routers, load balancers, web application firewalls, anti-distributed blocking service devices, intrusion detection / prevention systems, network device ports, hosts, and virtual machines. The data includes the following metrics, and a list of key network cloud metrics is shown in Table 1 below:
[0049] Table 1 The above indicators also have network element object and resource pool dimensions. Therefore, we can associate core network performance indicators with virtualized PIM and VIM data indicators through network element objects to construct a third-level virtualized resource data model.
[0050] The first-level data model, the second-level spatial resource data model, and the third-level virtualization resource data model are associated through network element objects to construct a multi-dimensional network element data model. This model covers multiple dimensions such as network element objects, PIM / VIM, resource pools, data centers, and network cloud regions, and a knowledge graph model is constructed. Compared with the previous PIM-VIM-VNF association method, this model has a more intuitive data analysis and resource topology structure, enabling data analysis applications such as rapid location of poor-quality data, historical data feature extraction, and indicator fluctuation trend analysis.
[0051] In network element performance monitoring, different thresholds are typically set for key indicators. When an indicator falls below a certain threshold, an alarm of the corresponding level is triggered. Upon receiving the alarm, monitoring personnel can dispatch tasks and handle the system to repair the fault and clear the alarm. Currently, this process has been largely automated, but the overall process is still in the post-event repair stage. How to effectively carry out pre-event early warning and improve the intelligence level of network monitoring is a major focus for monitoring and maintenance personnel.
[0052] In some embodiments of this disclosure, based on historical performance data in a multi-dimensional knowledge graph, a time series prediction model is used to perform univariate probabilistic prediction to obtain network element performance prediction information for future periods. One possible approach is to obtain historical performance data from the multi-dimensional knowledge graph; construct time series data based on the historical performance data; extract lag features from the time series data; combine the lag features and static covariates to form a label for a univariate time series at a given time step; and input the label of the univariate time series at the given time step into the time series prediction model to obtain network element performance prediction information for future periods.
[0053] Based on the constructed multi-dimensional network element performance data model, machine learning algorithms are introduced to perform feature extraction analysis and quality prediction on historical data. The Lag-LLaMA model is selected as the time series prediction algorithm. It is a fundamental model for univariate probabilistic time series prediction. By using lag features as covariates, it can be pre-trained on diverse time series data, effectively aligning with the data scenario of analyzing network element performance data using knowledge graphs in this proposal. The Lag-LLaMA model architecture is based on a decoder-only Transformer architecture. The model input is the label of a univariate time series at a given time step, which consists of lag time steps and static covariates.
[0054] For the multi-dimensional network data model mentioned above, performance metrics are continuously statistically analyzed at a 15-minute data granularity. For each time point, a state vector is constructed:
[0055] in, Indicates the first The vector formed by concatenating all the indicators of each network element. This represents the number of network elements.
[0056] Lag-LLaMA's tokenization scheme involves constructing lagged features from previous values of a time series. These features are built upon a specified set of lag indices, including quarterly, monthly, weekly, daily, hourly, and second-level frequencies. Its decoder-only Transformer architecture maps the input sequence to hidden dimensions through a linear projection layer and outputs the parameters of the predicted distribution through a distribution head. The final output layer of Lag-LLaMA, the distribution head, is responsible for outputting the parameters of the probability distribution; currently, it uses Student's t-distribution to construct the uncertainty interval.
[0057] Lag-LLaMA is pre-trained on large-scale time series data and uses a series of training strategies, such as stratified sampling and time series augmentation techniques, to improve the model's generalization ability. It also provides a training interface to facilitate the generation of highly customized time series prediction models.
[0058] The Lag-LLaMA model supports zero-shot learning, a relatively new concept. Its basic idea is to learn shared representations across multiple domains or tasks. This allows the model to identify and generalize to new categories or tasks without explicit training data. Specifically, this is typically achieved by using shared embedding layers that map input data from different domains or tasks to a common vector space, preserving the similarity between inputs. However, to ensure the accuracy of the data model, this disclosure uses data from the AMF-MME-SGSN model over the past year as samples for model training, improving prediction performance through long-term, consistent real-world data.
[0059] Large Language Models (LLMs) originate from time-series RNNs / LSTMs. This disclosure does not directly input time-series data into LLMs because these two types of data are different. The underlying time-series model aims to take time-series data as input, encode it accordingly, and capture time dependencies. Lag-LLaMA utilizes the lag characteristics of past values in time series to capture time dependencies. Prepare continuous data for a specific metric of AMF-MME-SGSN over a period of time, such as the number of successful registrations, and process it. Any time-series data should contain three basic elements: start date, target data, and data frequency.
[0060] By leveraging the key elements of the data, a prediction function is further constructed. This prediction function is then integrated with the network element performance data model to generate network element performance index values for the next time period.
[0061] In some embodiments of this disclosure, the predicted performance value of the target network element is compared with a preset alarm threshold; if the predicted performance value is less than the preset alarm threshold, a performance alarm operation is performed. If the predicted performance value is less than the preset alarm threshold, the target network element is marked as an early warning node in a multi-dimensional knowledge graph; an early warning message for the target network element is output. It should be noted that this disclosure does not limit the preset alarm threshold and can be adjusted according to the network element type and actual situation.
[0062] After generating the predicted performance data value of the network element object for the next time period, it is compared with the indicator alarm threshold. If it is lower than the threshold, an alarm is triggered, and it is marked in the knowledge graph for early warning analysis applications.
[0063] Based on the above scheme, the data foundation and early warning prediction data for the knowledge graph were constructed. In constructing the knowledge graph display scheme, a bottom-up approach was applied to summarize and organize network element relationships to form underlying concepts. Simultaneously, the relationships between network elements, interfaces, and service signaling processes in core network domains such as 5GC, EPC, and IMS were systematically analyzed and gradually abstracted upwards to form upper-layer graph concepts.
[0064] The network element graph is constructed according to the logic of "network element-interface-network element" and "network element-network element object-indicator-indicator value," building a network element metadata graph relationship database. Network element interactions are designed based on the following interface relationships between network elements. The list of interfaces between core network elements is shown in Table 2 below:
[0065] Table 2 Figure 2 This is a schematic diagram of a bottom-up topology flow for a knowledge graph, provided as an exemplary embodiment of this disclosure. For example... Figure 2As shown, the display of poor quality issues is divided into two dimensions: the problematic network elements and their indicator values for the current period are displayed according to their parent node (i.e., network element type A); the potential poor quality indicators of network element objects predicted by the Lag-LLaMA model are output to the graph database in the manner of "network element-network element object-indicator-indicator value" and displayed interactively according to the expected occurrence time of the poor quality issues.
[0066] This disclosure utilizes a knowledge graph-based network performance correlation analysis algorithm and apparatus to achieve multi-dimensional, intelligent, and real-time network performance data analysis and alarm prediction. Through a unified data model and visualized knowledge graph, it improves the efficiency and accuracy of network performance management, exhibits good scalability and intelligence, and enhances the optimization and management of communication networks.
[0067] Figure 3 This is a schematic diagram of the structure of a network element performance prediction device 30 provided for an exemplary embodiment of this disclosure. Figure 3 As shown, the network element performance prediction device 30 includes: a first construction module 31, a second construction module 32, a third construction module 33, a fourth construction module 34, and a prediction module 35.
[0068] The first construction module 31 is used to identify and aggregate converged network elements that carry the functions of multiple generations of core networks based on the network element names of each network element in the core network, and to construct a first-level data model centered on the network element object. The second construction module 32 is used to construct a second-level spatial resource data model based on the logical topology relationship between network elements, resource pools, data centers and network regions. The third construction module 33 is used to associate the performance index information of each network element in the core network with the PIM data index and VIM data index through the network element object, and construct the third-level virtualization resource data model. The fourth construction module 34 is used to construct a multi-dimensional knowledge graph based on the first-level data model, the second-level spatial resource data model, and the third-level virtualization resource data model. The prediction module 35 is used to perform univariate probability prediction based on historical performance data in a multi-dimensional knowledge graph and a time series prediction model to obtain network element performance prediction information for future periods.
[0069] Optionally, the performance metrics information includes: resource data and performance data. The first building module 31 can also be used for: The performance data files reported by the northbound interface of the core network OMC equipment are parsed to obtain the performance index information of each network element of the core network. Resource data and performance data are stored separately according to the managed object class.
[0070] Optionally, the network element performance prediction information is the predicted value of network element performance. The prediction module 35 can also be used for: Compare the predicted network element performance value of the target network element with the preset alarm threshold; If the predicted performance value of a network element is less than the preset alarm threshold, a performance alarm operation will be executed.
[0071] Optionally, when the prediction module 35 performs a performance alarm operation if the predicted value of the network element performance is less than a preset alarm threshold, it is used to: If the predicted performance value of a network element is less than the preset alarm threshold, the target network element will be marked as an early warning node in the multi-dimensional knowledge graph. Output early warning information for the target network element.
[0072] Optionally, when the prediction module 35 obtains network element performance prediction information for future periods by using a time series prediction model to perform univariate probability prediction based on historical performance data in a multi-dimensional knowledge graph, it is used for: Obtain historical performance data from a multi-dimensional knowledge graph; Construct time-series data based on historical performance data; Extracting lag features from time series data; The lag features and static covariates are used to create a label for a univariate time series at a given time step; By inputting the labels of a univariate time series at a given time step into a time series prediction model, network element performance prediction information for future periods can be obtained.
[0073] Optionally, the converged network element includes at least one of the following: AMF-MME-SGSN, SMF-PGWC-SGWC, UPF-PGWU-SGWU, PCF-PCRF, and UDM-HSS.
[0074] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.
[0075] Figure 4 This is a schematic diagram of the structure of an electronic device provided as an exemplary embodiment of the present disclosure. For example... Figure 4 As shown, the electronic device includes a memory 41 and a processor 42. Additionally, the electronic device also includes a power supply component 43 and a communication component 44.
[0076] Memory 41 is used to store computer programs and can be configured to store various other data to support operation on the electronic device. Examples of this data include instructions for any application or method used to operate on the electronic device.
[0077] The memory 41 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.
[0078] Communication component 44 is used for data transmission with other devices.
[0079] The processor 42 executes computer instructions stored in the memory 41 to: identify and aggregate converged network elements carrying multi-generation core network functions based on the network element names of each core network element, and construct a first-level data model centered on network element objects; construct a second-level spatial resource data model based on the logical topology relationship between network elements, resource pools, data centers, and network regions; construct a third-level virtualization resource data model by associating the performance index information of each network element in the core network with PIM data indicators and VIM data indicators through network element objects; construct a multi-dimensional knowledge graph based on the first-level data model, the second-level spatial resource data model, and the third-level virtualization resource data model; and obtain network element performance prediction information for future periods by using a time series prediction model to perform univariate probability prediction based on historical performance data in the multi-dimensional knowledge graph.
[0080] Accordingly, embodiments of this disclosure also provide a computer-readable storage medium storing a computer program. When the computer-readable storage medium stores a computer program, and the computer program is executed by one or more processors, it causes one or more processors to perform... Figure 1 Each step in the method embodiment.
[0081] Accordingly, this disclosure also provides a computer program product, which includes a computer program / instructions that are executed by a processor. Figure 1 Each step in the method embodiment.
[0082] The above Figure 4The communication component is configured to facilitate wired or wireless communication between the device containing the communication component and other devices. The device containing the communication component can access wireless networks based on communication standards, such as WiFi, 2G, 3G, 4G / LTE, 5G, or combinations thereof. In one exemplary embodiment, the communication component receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, the communication component also includes a Near Field Communication (NFC) module to facilitate short-range communication. For example, the NFC module may be implemented based on Radio Frequency Identification (RFID), Infrared Data Association (IrDA) technology, Ultra-Wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.
[0083] The above Figure 4 The power supply component provides power to the various components of the device in which it resides. The power supply component may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the device in which it resides.
[0084] The aforementioned electronic devices also include a display screen and audio components.
[0085] The display includes a screen, which may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen can be implemented as a touchscreen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors can sense not only the boundaries of touch or swipe actions, but also the duration and pressure associated with the touch or swipe operation.
[0086] An audio component may be configured to output and / or input audio signals. For example, the audio component includes a microphone (MIC) configured to receive external audio signals when the device containing the audio component is in an operating mode, such as call mode, recording mode, or voice recognition mode. The received audio signals may be further stored in memory or transmitted via a communication component. In some embodiments, the audio component also includes a speaker for outputting audio signals.
[0087] Those skilled in the art will understand that embodiments of this disclosure can be provided as methods, systems, or computer program products. Therefore, this disclosure can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this disclosure can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0088] This disclosure is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0089] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0090] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0091] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.
[0092] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.
[0093] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.
[0094] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.
[0095] The above are merely specific embodiments of this disclosure, enabling those skilled in the art to understand or implement this disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this disclosure. Therefore, this disclosure is not to be limited to these embodiments, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for predicting network element performance, characterized in that, include: Based on the network element names of each network element in the core network, identify and aggregate the converged network elements that carry the functions of multiple generations of core networks, and construct a first-level data model centered on network element objects; Based on the logical topological relationship between network elements, resource pools, data centers and network regions, a second-level spatial resource data model is constructed. By associating the performance index information of each network element in the core network with PIM data index and VIM data index through the network element object, a third-level virtualization resource data model is constructed. A multi-dimensional knowledge graph is constructed based on the first-level data model, the second-level spatial resource data model, and the third-level virtualization resource data model. Based on the historical performance data in the multi-dimensional knowledge graph, a time series prediction model is used to perform univariate probability prediction to obtain network element performance prediction information for future periods.
2. The method according to claim 1, characterized in that, The performance metric information includes resource data and performance data, and the method further includes: The performance data files reported by the northbound interface of the core network OMC equipment are parsed to obtain the performance index information of each network element of the core network. The resource data and performance data are stored separately according to the management object class.
3. The method according to claim 1, characterized in that, The network element performance prediction information is the network element performance prediction value, and the method further includes: Compare the predicted network element performance value of the target network element with the preset alarm threshold; If the predicted performance value of the network element is less than the preset alarm threshold, a performance alarm operation is performed.
4. The method according to claim 3, characterized in that, The step of performing a performance alarm operation when the predicted performance value of the network element is less than the preset alarm threshold includes: If the predicted performance value of the network element is less than the preset alarm threshold, the target network element will be marked as an early warning node in the multi-dimensional knowledge graph. Output early warning information for the target network element.
5. The method according to claim 1, characterized in that, Based on the historical performance data in the multi-dimensional knowledge graph, a time series prediction model is used to perform univariate probability prediction to obtain network element performance prediction information for future periods, including: Obtain historical performance data from a multi-dimensional knowledge graph; Based on the historical performance data, construct time series data; Extract lag features from the time series data; The lag features and static covariates are used to form a label for a univariate time series at a given time step; By inputting the label of the univariate time series at the given time step into the time series prediction model, network element performance prediction information for future periods can be obtained.
6. The method according to claim 1, characterized in that, The converged network element includes at least one of the following: AMF-MME-SGSN, SMF-PGWC-SGWC, UPF-PGWU-SGWU, PCF-PCRF, and UDM-HSS.
7. A network element performance prediction device, characterized in that, include: The first construction module is used to identify and aggregate converged network elements that carry the functions of multiple generations of core networks based on the network element names of each network element in the core network, and to construct a first-level data model centered on network element objects. The second construction module is used to construct a second-level spatial resource data model based on the logical topological relationship between network elements, resource pools, data centers and network regions. The third construction module is used to associate the performance index information of each network element in the core network with the PIM data index and VIM data index through the network element object, and construct the third-level virtualization resource data model. The fourth construction module is used to construct a multi-dimensional knowledge graph based on the first-level data model, the second-level spatial resource data model, and the third-level virtualization resource data model. The prediction module is used to perform univariate probability prediction using a time series prediction model based on historical performance data in the multi-dimensional knowledge graph, so as to obtain network element performance prediction information for future periods.
8. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to execute instructions to implement the steps of the method as described in any one of claims 1-6.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1-6.
10. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method according to any one of claims 1-6.