US-12621227-B2 - Service assurance in 5G networks using key performance indicator navigation tool

US12621227B2US 12621227 B2US12621227 B2US 12621227B2US-12621227-B2

Abstract

A navigation tool using a visual language is configured to interoperate with a curated catalog of KPIs that enables users associated with 5G mobile operators to implement service assurance in a graphical manner based on a unique ontological model of an operator's 5G network. The graphical navigation tool provides visually-based context to the catalog to streamline KPI selection while leveraging the cognitive benefits of the visual language to facilitate discovery, grouping, and connecting of the KPIs in a meaningful way to express essential aspects of 5G network performance.

Inventors

Alejandro Jose MIGUEL
William Lee LABOR, JR.

Assignees

MICROSOFT TECHNOLOGY LICENSING, LLC

Dates

Publication Date: 20260505
Application Date: 20220930

Claims (12)

1 . One or more hardware-based non-transitory computer-readable memory devices storing computer-executable instructions which, in response to execution by one or more processors disposed in a computing device, cause the computing device to: access a machine learning system configured for creating a semantic data model of a software-defined mobile network, the semantic data model representing network elements in the software-defined mobile network and data interfaces among the network elements, the machine learning system comprising a multi-layer neural network, wherein: the semantic data model comprises an ontology having a plurality of nodes representing the network elements and point-to-point interfaces among the network elements, the network elements comprising physical and logical systems and subsystems in a deployed instance of the software-defined mobile network; and one or more of the network elements support a virtualized network function (VNF); train the semantic data model by inputting representations of a training dataset of network elements and interfaces into the machine learning system and adjusting weights in one or more layers of the neural network to minimize an error function at an output of the machine learning system; populate the semantic data model by inputting representations of an unknown dataset of network elements and data interfaces into the machine learning system to create labeled output; generate a plurality of different versions of the ontology of the software-defined mobile network and perform semantic mapping between the different versions; and utilize the populated semantic data model to create a graph of key performance indicators (KPIs) applicable to the mobile network, the graph identifying ontological relationships for KPIs among specific instances of network elements and data interfaces in the software-defined mobile network.
2 . The one or more hardware-based non-transitory computer-readable memory devices of claim 1 in which the machine learning system comprises a convolutional neural network.
3 . The one or more hardware-based non-transitory computer-readable memory devices of claim 1 in which software-defined mobile network comprises a fifth generation (5G) mobile network.
4 . The one or more hardware-based non-transitory computer- readable memory devices of claim 1 in which the ontology further comprises point-to-point data interfaces between the network elements, the point-to-point interfaces described by 3GPP (3 rd Generation Partnership Project) TS 23 501.
5 . A computer-implemented method for service assurance in a mobile network comprising: accessing a machine learning system configured for creating a semantic data model of a software-defined mobile network, the semantic data model representing network elements in the software-defined mobile network and data interfaces among the network elements, the machine learning system comprising a multi-layer neural network, wherein: the semantic data model comprises an ontology having a plurality of nodes representing the network elements and point-to-point interfaces among the network elements, the network elements comprising physical and logical systems and subsystems in a deployed instance of the software-defined mobile network; and one or more of the network elements support a virtualized network function (VNF); training the semantic data model by inputting representations of a training dataset of network elements and interfaces into the machine learning system and adjusting weights in one or more layers of the neural network to minimize an error function at an output of the machine learning system; populating the semantic data model by inputting representations of an unknown dataset of network elements and data interfaces into the machine learning system to create labeled output; generating a plurality of different versions of the ontology of the software-defined mobile network and performing semantic mapping between the different versions; and utilizing the populated semantic data model to create a graph of key performance indicators (KPIs) applicable to the mobile network, the graph identifying ontological relationships for KPIs among specific instances of network elements and data interfaces in the software-defined mobile network.
6 . The method of claim 5 , wherein the machine learning system comprises a convolutional neural network.
7 . The method of claim 5 , wherein software-defined mobile network comprises a fifth generation (5G) mobile network.
8 . The method of claim 5 , wherein the ontology further comprises point-to-point data interfaces between the network elements, the point-to-point interfaces described by 3GPP (3 rd Generation Partnership Project) TS 23 501.
9 . A computing device used by a key performance indicator (KPI) service, comprising: one or more processors; memory in electronic communication with the one or more processors; a user interface (UI); and one or more hardware-based non-transitory computer-readable storage devices having computer-executable instructions stored thereon which, when executed by the one or more processors, cause the computing device to perform operations comprising: access a machine learning system configured for creating a semantic data model of a software-defined mobile network, the semantic data model representing network elements in the software-defined mobile network and data interfaces among the network elements, the machine learning system comprising a multi-layer neural network, wherein: the semantic data model comprises an ontology having a plurality of nodes representing the network elements and point-to-point interfaces among the network elements, the network elements comprising physical and logical systems and subsystems in a deployed instance of the software-defined mobile network; and one or more of the network elements support a virtualized network function (VNF); train the semantic data model by inputting representations of a training dataset of network elements and interfaces into the machine learning system and adjusting weights in one or more layers of the neural network to minimize an error function at an output of the machine learning system; populate the semantic data model by inputting representations of an unknown dataset of network elements and data interfaces into the machine learning system to create labeled output; generate a plurality of different versions of the ontology of the software-defined mobile network and perform semantic mapping between the different versions; and utilize the populated semantic data model to create a graph of key performance indicators (KPIs) applicable to the mobile network, the graph identifying ontological relationships for KPIs among specific instances of network elements and data interfaces in the software-defined mobile network.
10 . The computing device of claim 9 , wherein the machine learning system comprises a convolutional neural network.
11 . The computing device of claim 9 , wherein software-defined mobile network comprises a fifth generation (5G) mobile network.
12 . The computing device of claim 9 , wherein the ontology further comprises point-to-point data interfaces between the network elements, the point-to-point interfaces described by 3GPP (3 rd Generation Partnership Project) TS 23 501.

Description

BACKGROUND Fifth-generation (5G) mobile networks can be implemented using hybrid cloud technologies comprising a combination of private- and public-cloud infrastructure that supports software defined networking (SDN) in a service-based architecture (SBA). Using SDN and SBA, mobile operators can select and manage virtualized network functions (VNFs) and/or cloud-native network functions (CNFs) to deploy self-optimizing networks that can heal, defend, and provision themselves. Key performance indicators (KPIs) are quantifiable measures of performance for specific network metrics that describe the effectiveness of the 5G networks in meeting goals of end-users and mobile operators in various ways. KPIs in 5G networks are defined by the 3rd Generation Partnership Project (3GPP) in ETSI (European Telecommunications Standards Institute) TS 128 554. Testing and verifying KPIs can be challenging for mobile operators because real-world deployments of 5G networks are typically complex and dynamic using new and legacy infrastructure that support a variety of VNFs as new services and device types are constantly introduced. A need exists for effective tools and methodologies for 5G network services to have assured performance to thereby enable mobile operator resources and efforts to be concentrated on providing maximum benefits for end-users while meeting network operating targets. SUMMARY A navigation tool for key performance indicators (KPIs) is configured to support a visual language to interoperate with a curated catalog of KPIs that enables users associated with 5G mobile operators to implement service assurance in a graphical manner. The curated KPI catalog is based on a graph of an operator's 5G network that is created using a unique ontology that provides a standardized semantic model applicable across diverse 5G networks which can be complex with widely varying deployment characteristics. The navigation tool enables context for the catalog entries to be surfaced to users to streamline KPI selection while leveraging the cognitive benefits of the visual language to facilitate discovery, grouping, and connecting of the KPIs for users to better understand key aspects of 5G network performance. In various illustrative examples, the navigation tool is supported by a KPI discovery engine configured to identify relevant parameters for cataloged KPIs including data sources and network interfaces that constitute the data points from which the ontology is populated. The discovery engine includes a machine learning system adapted to build an ontologically-based KPI graph for a specific deployment of a 5G network of a mobile operator in an automated manner. An ontology is defined using multiple nodes representing network elements that include physical and logical systems and subsystems in the deployed 5G network and point-to-point interfaces between the elements known as reference points according to 3GPP TS 23.501 (ETSI TS 123 501). Employing suitable training data and algorithms, the machine learning system can identify data used to form KPIs along with relationships and connections in the KPI graph among nodes, including indirect connections among distant nodes. KPI graphs can be dynamically modified, and new graphs can be created on the fly using the machine learning system to reflect changes in 5G network configuration and new network deployments. The navigation tool may be extended to interoperate with conventional data visualization tools to expose KPI data using standard (i.e., out-of-the-box) or custom visuals. The integrity and reliability of visual KPI data is improved because the underlying connections, sources, and calculations for the data are comprehensively known from the ontology-based knowledge graph for the deployed 5G network. This knowledge can provide context to the KPI data which can aid in interpretation and analysis of 5G network performance for service assurance and other purposes. The increased context facilitated by the navigation tool enables users to fill in gaps in understanding and leverages user knowledge to provide a holistic view of the entirety of the 5G network with enhanced connectedness across the generated KPI visuals and reports. Use of the navigation tool provides numerous technical advantages in addition to improvement in KPI data integrity and reliability. Navigation of KPIs is streamlined, for example, as the tool enables users to select a KPI from the catalog and then quickly visually locate other KPIs and associated data of interest that are relevant to a given 5G network deployment. Using the navigation tool, the user can efficiently work through the physical and logical network systems and subsystems by following visual cues to discover relevant relationships and connections. Links may be readily established between sources and visualizations to provide context for the KPIs to thereby gain additional insights and understanding of network state and performance. The complexity