An intent decomposition method, apparatus and electronic device

Abstract

The application provides an intention decomposition method, device and electronic equipment. The method comprises the following steps: obtaining a target intention to be decomposed and state information corresponding to the target intention; decomposing the target intention based on the state information to obtain a plurality of sub-intentions and an execution sequence of the plurality of sub-intentions; and executing the plurality of sub-intentions based on the execution sequence of the plurality of sub-intentions. The application can improve the processing effect of the intention.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to an intent decomposition method, apparatus and electronic device. Background Technology

[0002] In existing solutions, intent processing mainly includes processes such as intent parsing, installation, and execution. For simple intents, the above processes can basically complete the design and operation of the intent. However, for complex intents, such as applying multi-layered complex intents to slice-related businesses, the intent processing effect is poor. Summary of the Invention

[0003] This application provides an intent decomposition method, apparatus, and electronic device to solve the problem of poor intent processing performance.

[0004] In a first aspect, embodiments of this application provide an intent decomposition method, including:

[0005] Obtain the target intent to be decomposed, and the state information corresponding to the target intent;

[0006] Based on the state information, the target intent is decomposed to obtain multiple sub-intents and the execution order of the multiple sub-intents;

[0007] The multiple sub-intents are executed according to their execution order.

[0008] Secondly, embodiments of this application also provide an intent decomposition apparatus, comprising:

[0009] The acquisition module is used to acquire the target intent to be decomposed, and the state information corresponding to the target intent;

[0010] The decomposition module is used to decompose the target intent based on the state information to obtain multiple sub-intents and the execution order of the multiple sub-intents;

[0011] An execution module is used to execute the multiple sub-intents based on their execution order.

[0012] Thirdly, embodiments of this application also provide an electronic device, including: a transceiver, a memory, a processor, and a program stored in the memory and executable on the processor; the processor is configured to read the program in the memory to implement the steps in the method described in the first aspect of embodiments of this application.

[0013] Fourthly, embodiments of this application also provide a readable storage medium storing a program that, when executed by a processor, implements the steps of the method described in the first aspect of embodiments of this application.

[0014] In this embodiment, a target intent to be decomposed and state information corresponding to the target intent are obtained. Based on the state information, the target intent is decomposed to obtain multiple sub-intents and the execution order of the multiple sub-intents. Based on the execution order of the multiple sub-intents, the multiple sub-intents are executed. That is, the target intent to be decomposed can be decomposed into multiple sub-intents and executed according to the execution order of the multiple sub-intents, thereby realizing the processing of the target intent and improving the execution effect of intent processing. Attached Figure Description

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

[0016] Figure 1 This is a flowchart illustrating an intent decomposition method provided in an embodiment of this application;

[0017] Figure 2 This is a schematic diagram of a layered communication network structure provided in an embodiment of this application;

[0018] Figure 3 This is an architecture diagram of an intent management function provided in an embodiment of this application;

[0019] Figure 4 This is a flowchart illustrating an intent orchestration method provided in an embodiment of this application;

[0020] Figure 5 This is a flowchart illustrating another intent decomposition method provided in an embodiment of this application;

[0021] Figure 6 This is a flowchart illustrating another intent decomposition method provided in an embodiment of this application;

[0022] Figure 7 This is a schematic diagram of a sub-intent arrangement provided in an embodiment of this application;

[0023] Figure 8 This is a schematic diagram of another sub-intent arrangement provided in the embodiments of this application;

[0024] Figure 9 This is a schematic diagram of the structure of an intent decomposition device provided in an embodiment of this application;

[0025] Figure 10 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0026] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0027] The terms "first," "second," etc., used in the embodiments of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or devices. Additionally, the use of "and / or" in this application indicates at least one of the connected objects, such as A and / or B and / or C, representing seven possibilities: including A alone, B alone, C alone, both A and B present, both B and C present, both A and C present, and A, B, and C present.

[0028] With the emergence and development of Network Functions Virtualization (NFV) and Software-defined Networking (SDN) technologies, and the continuous maturation of artificial intelligence, network automation has gradually increased. Consequently, the application of intent in communication networks has naturally evolved, leading to network technologies that configure Internet Technology (IT) infrastructure based on business intent (service requests from network administrators, without any manual intervention). This allows for the continuous provision of critical network insights based on business intent and the ongoing adjustment of hardware configurations to ensure that intent is met, thereby transforming the network management model from a device-centric model to a service-centric model.

[0029] Network slicing allows operators to create multiple virtual end-to-end networks on a unified infrastructure through on-demand networking. Each network slice is logically isolated from the radio access network to the core network to adapt to various types of applications. A network slice can be divided into three parts: radio network sub-slices, bearer network slices, and core network slices, and is managed uniformly through an end-to-end management system. On a common physical infrastructure, different functional modules are deployed according to business needs; that is, management function blocks can be deployed on multiple physical entities.

[0030] Please see Figure 1 , Figure 1 This is a schematic diagram of an intent decomposition process provided in an embodiment of this application, such as... Figure 1 As shown, it includes the following steps:

[0031] Step 101: Obtain the target intent to be decomposed, and the state information corresponding to the target intent.

[0032] The aforementioned target intent to be decomposed can be the intent input by a communication service client in the hierarchical structure of the communication network. Specifically, the hierarchical structure of the communication network can be divided into communication service clients, communication service providers, and network operators. The intent of a communication service client allows it to indicate which attributes of the communication service it can request from the communication service provider, without needing to know how to manage the communication service in detail. For example, the intent of a communication service client could be "to enable VIP communication service for a group of vehicles at a specific time and location." The intent from a communication service provider allows it to express what it wants to do for the network, without needing to know how to manage the network in detail. For example, the intent of a communication service provider could be "to provide network service supporting VIP communication for Highway 417, capable of supporting 500 vehicles simultaneously." The intent of a network operator allows it to provide the management and control it wants over a group of network elements (i.e., subnetworks), without needing to know how to manage the network elements in detail. For example, the intent of a network operator could be "to provide wireless network service to meet specified coverage requirements and user equipment (UE) throughput requirements in a specific area."

[0033] It can be understood that the state information corresponding to the above-mentioned target intent is the state information of the system or domain related to the above-mentioned target intent, such as network state information, resource state information, etc.

[0034] Step 102: Based on the state information, decompose the target intent to obtain multiple sub-intents and the execution order of the multiple sub-intents.

[0035] In a hierarchical structure of a communication network, if the target intent is a complex intent involving multiple levels, it can be decomposed into multiple sub-intents based on the relevant state information, and the execution order of the multiple sub-intents can be determined. Thus, by executing the multiple sub-intents respectively, the target intent can be processed.

[0036] Step 103: Execute the multiple sub-intents based on their execution order.

[0037] This application can be applied to any management function block in a communication network structure, and implements intent management function by decomposing and executing the intent of the previous level or the intent input by the user.

[0038] It should be noted that this application introduces general intent management functions in the intent layering of communication networks, such as... Figure 2 As shown, on the one hand, it enables the management functions at all levels of the communication network to support intent functions; on the other hand, each level of intent management function can act as both the intent processor of the intent management function at the higher level and the intent owner of the intent management function at the lower level. Through these intent management functions, the hierarchical decomposition of intents in the communication network is supported.

[0039] In this embodiment, a target intent to be decomposed and state information corresponding to the target intent are obtained. Based on the state information, the target intent is decomposed to obtain multiple sub-intents and the execution order of the multiple sub-intents. Based on the execution order of the multiple sub-intents, the multiple sub-intents are executed. That is, the target intent to be decomposed can be decomposed into multiple sub-intents and executed according to the execution order of the multiple sub-intents, thereby realizing the processing of the target intent and improving the execution effect of intent processing.

[0040] Optionally, step 102, which involves decomposing the target intent based on the state information to obtain multiple sub-intents and their execution order, includes:

[0041] Based on the state information, the target intent is decomposed to obtain multiple sub-intents and the initial execution order of the multiple sub-intents;

[0042] Obtain the intent and orchestration goals;

[0043] The multiple sub-intents are sorted based on the intent orchestration target to obtain the execution order of the multiple sub-intents.

[0044] The initial execution order can be predetermined. For example, multiple sub-intents under a parent intent and their execution order can be predetermined based on experience, such as determining the first sub-intent to be executed and the subsequent intents of each sub-intent. In this way, after decomposing the target intent, the multiple sub-intents obtained from the decomposition and their initial execution order can be determined. Alternatively, the initial execution order can also be determined through model training. For example, a pre-trained intent decomposition model can be used, and the determined parent intent can be input into the model to obtain multiple output sub-intents and their initial execution order. This application does not impose any restrictions on this.

[0045] It should be noted that the aforementioned intention orchestration goals can originate from the intention owner's desired outcome. For example, the orchestration goal could be to select the sub-intention orchestration order with the shortest execution time or the one that consumes the fewest resources. Furthermore, in addition to the desired outcomes, the intention orchestration goals can also exclude certain sub-intention execution orders using constraints, such as resource limitations or mutually exclusive conditions between intention owners. By orchestrating multiple sub-intentions to be orchestrated to obtain various execution orders, and based on the aforementioned orchestration goals and constraints, the final execution order is determined.

[0046] The process of sorting multiple sub-intents based on the aforementioned intent orchestration objective to obtain the execution order of these multiple sub-intents can be understood as rearranging the initial execution order of the multiple sub-intents to obtain the final execution order of the sub-intents.

[0047] The execution order of sub-intents can be orchestrated based on pre-defined orchestration experience or through model training. For example, an orchestration scheme, such as a commonly used workflow scheduling algorithm, can be pre-defined in an orchestration experience library. By using this scheme and the aforementioned intent orchestration goals, the execution order of the multiple sub-intents can be obtained. The intent orchestration goals may also include orchestration constraints and information about the intent objects to be orchestrated. Alternatively, an intent orchestration model can be pre-trained based on historical learning experience, and the training results can be stored in the orchestration experience library to enrich its data and improve the accuracy of intent orchestration. The above examples are optional methods for sorting multiple sub-intents based on intent orchestration goals. Other specific methods for sorting multiple sub-intents based on intent orchestration goals can also be used, and this application does not impose any restrictions.

[0048] In this embodiment, based on the state information, the target intent is decomposed to obtain multiple sub-intents and their initial execution order; and the multiple sub-intents are sorted according to the intent orchestration target to obtain their execution order. That is, the execution order of the multiple sub-intents can be determined according to the intent orchestration target, which can be used to measure the execution effect of the sub-intents, thereby obtaining an execution order that conforms to the intent orchestration target and improving the execution effect of the multiple sub-intents.

[0049] Optionally, the step of sorting the plurality of sub-intents based on the intent orchestration target to obtain the execution order of the plurality of sub-intents includes:

[0050] Obtain multiple candidate execution orders for the multiple sub-intents;

[0051] The intent orchestration parameters for the various candidate execution orders are calculated respectively, and the intent orchestration parameters are determined based on the intent orchestration target;

[0052] Among the various candidate execution sequences, the execution sequence that satisfies the preset conditions for the intended orchestration parameters is determined.

[0053] The aforementioned intent orchestration parameters can be used to measure the execution effect of the multiple sub-intents under various execution orders. For example, when the intent orchestration goal is to select the sub-intent execution order with the shortest execution time, the execution time of each candidate execution order can be used as the intent orchestration parameter. In this way, the aforementioned preset conditions can be set to determine the candidate execution order with the shortest execution time (i.e., the smallest intent orchestration parameter), so that the multiple sub-intents can be executed according to the candidate execution order with the shortest execution time to achieve the aforementioned intent orchestration goal.

[0054] Specifically, for the multiple sub-intents obtained from the decomposition, various candidate execution orders can be freely arranged. If the above intent arrangement target also includes constraints, the corresponding parameters used to represent the above constraints can also be used as the above intent arrangement parameters.

[0055] In this embodiment, intent orchestration parameters are calculated for the various candidate execution orders, and the intent orchestration parameters are determined based on the intent orchestration target. Among the various candidate execution orders, the execution order in which the intent orchestration parameters satisfy preset conditions is determined. The execution effect of the various candidate execution orders is quantified in the form of the intent orchestration parameters, thereby improving the accuracy of determining the execution order of the multiple sub-intents.

[0056] Optionally, the step of decomposing the target intent based on the state information to obtain multiple sub-intents and the initial execution order of the multiple sub-intents includes:

[0057] Based on the state information, the target intent is decomposed to obtain the multiple sub-intents and the topological relationship of the multiple sub-intents;

[0058] Based on the topological relationship of the multiple sub-intents, the initial execution order of the multiple sub-intents is determined.

[0059] It should be noted that the topological relationship of the above-mentioned multiple sub-intents can represent multiple sub-intents under the same parent intent, as well as the execution order among the multiple sub-intents. For example, the first sub-intent, the last sub-intent, and the successor intent of each sub-intent can be determined. The first sub-intent is the first sub-intent to be executed, the last sub-intent is the last sub-intent to be executed, and the successor intent is the next sub-intent to be executed after each sub-intent is completed. The successor intent of the last sub-intent can be the parent intent, so as to realize the topological relationship closed loop of the parent intent and its multiple intents. After the multiple sub-intents under the parent intent are executed (that is, after the last sub-intent is executed), the last sub-intent points to the parent intent in the topological relationship to indicate that the multiple sub-intents under the parent intent have been completed.

[0060] In this embodiment, the initial execution order of the multiple sub-intents is determined based on their topological relationship, which can enable the multiple sub-intents to be executed in the order corresponding to their topological relationship, thereby improving the execution effect of the sub-intents.

[0061] Optionally, step 102, which involves decomposing the target intent based on the state information to obtain multiple sub-intents and their execution order, includes:

[0062] Based on the state information, a target decomposition scheme for the target intent is determined;

[0063] If the intent experience base includes the target decomposition scheme, the target decomposition scheme is obtained from the intent experience base;

[0064] If the intent experience base does not include the target decomposition scheme, the target intent and the state information are input into a pre-trained model to predict the decomposition scheme in order to obtain the target decomposition scheme.

[0065] Based on the target decomposition scheme, the plurality of sub-intents and the execution order of the plurality of sub-intents are determined.

[0066] It is understood that the aforementioned target decomposition scheme is used to determine the aforementioned multiple sub-intents and their execution order, and the aforementioned target decomposition scheme may be included among the multiple decomposition schemes stored in the aforementioned intent experience base. The aforementioned intent experience base may pre-store decomposition schemes for multiple intents; for example, multiple intents may be used as parent intents, listing multiple sub-intents under the parent intent and their execution order.

[0067] The specific form and implementation language of the pre-trained model can be arbitrary. Machine learning is used to predict the decomposition scheme, thereby improving the accuracy of the prediction scheme.

[0068] In addition, if the intent experience base does not include the above-mentioned target decomposition scheme, the generated decomposition scheme can be updated into the intent experience base. By continuously enriching the experience base, the accuracy of obtaining the above-mentioned decomposition scheme can be improved.

[0069] In this implementation, if the intent experience base includes the target decomposition scheme, the target decomposition scheme is obtained from the intent experience base; if the intent experience base does not include the target decomposition scheme, the target intent and the state information are input into a pre-trained model to predict the decomposition scheme to obtain the target decomposition scheme. This allows the decomposition scheme to be obtained according to different situations, improving the efficiency of obtaining the target decomposition scheme when the intent experience base includes it, and ensuring the accuracy of obtaining the target decomposition scheme even when the intent experience base does not include it.

[0070] Optionally, the execution of the multiple sub-intents based on the execution order of the multiple sub-intents in step 103 includes:

[0071] Based on the multiple sub-intents and their execution order, an intent instance is obtained;

[0072] Based on the intent instance, execute the plurality of sub-intents.

[0073] It is understood that the above intent instances are obtained during the execution of the intent in order to determine the specific parameter information of the above multiple sub-intents during the execution process.

[0074] Optionally, obtaining an intent instance based on the plurality of sub-intents and their execution order includes:

[0075] Based on the multiple sub-intents and their execution order, query the intent instances of the multiple sub-intents in the intent instance library;

[0076] If the intent instance library includes intent instances of the plurality of sub-intents, then the intent instances of the plurality of sub-intents are obtained from the intent instance library;

[0077] If the intent instance library does not include intent instances of the multiple sub-intents, then intent instances of the multiple sub-intents are generated.

[0078] After obtaining the intent instance based on the plurality of sub-intents and their execution order, the method further includes:

[0079] The intent instance library is updated based on the intent instance.

[0080] It is understandable that during the execution of the aforementioned multiple sub-intents, corresponding intent instances can be generated and saved to the aforementioned intent instance library. If the aforementioned intent instance library stores intent instances with identical parameters, the aforementioned multiple sub-intents can also directly reuse the corresponding intent instances during execution.

[0081] Thus, if the intent instance library includes intent instances of the multiple sub-intents, the intent instances in the intent instance library can be directly obtained; if the intent instance library does not include intent instances of the multiple sub-intents, intent instances that do not exist in the intent instance library can be updated to the intent instance library.

[0082] Optionally, one sub-intention corresponds to one sub-network slice; the execution of the multiple sub-intentions based on their execution order includes:

[0083] Based on the execution order of the multiple sub-intents, the multiple sub-intents are executed in multiple subnet slices respectively.

[0084] The aforementioned subnet slices can be core network-side slices, bearer network-side slices, or wireless-side slices, or they can be further slices under the aforementioned slices, so that the corresponding sub-intents can be executed in the aforementioned multiple subnet slices respectively.

[0085] In this embodiment, based on the execution order of the multiple sub-intents, the multiple sub-intents are executed in multiple subnet slices respectively, thereby realizing the decomposition and execution of the complex target intent.

[0086] For ease of understanding, a specific example is as follows:

[0087] To support the decomposition and combination of intents across multiple layers in a communication network, intent management functionality needs to be overlaid on network management functions at each layer. This allows each layer's network management function to support intent functionality and simultaneously assume both the management roles of intent owner and intent processor in different contexts. For example, a communication service provider's intent is an intent processor for communication service customers, but an intent owner for network operator intents. Conversely, a network operator's intent is an intent processor for communication service provider intents, but an intent owner for network service-related intents. A diagram illustrating the introduction of a general intent management function in the communication network intent layering is shown below. Figure 2 As shown.

[0088] The Intent Owner is the source of the intent and is responsible for managing the lifecycle of the intent object, including its creation, necessary modifications, and eventual deletion. Only the Intent Owner is allowed to create, modify, or delete the intent.

[0089] Once an Intent Handler accepts a specific Intent from an Intent Owner, it is obligated to perform the necessary actions, based on the available resources and solutions within its managed domain, to satisfy the Intent as much as possible, and to report the Intent processing status and success rate to the Intent Owner. For Intents that cannot be satisfied, the Intent Handler cannot directly modify the Intent, but can refuse to accept it, or negotiate with the Intent Owner through "Intent Exploration" to persuade the Intent Owner to change its initial intention, with both parties accepting the revised Intent.

[0090] By introducing common intent management functional modules at each layer, the overall architecture is as follows: Figure 3 As shown.

[0091] Specifically, the decision-making module can be used to perform the following functions:

[0092] Interacting with the intent implementation experience base allows you to obtain intent decomposition solutions and save dynamically generated intent decomposition experience to the intent implementation experience base.

[0093] Based on the information input from the intent experience base and the cognitive module, the intent is decomposed into sub-intents, and the relevant information of each sub-intent is defined;

[0094] It interacts with the training module, distributing the data required for training (intent input information, collected system information, intent decomposition information, etc.) to the training module for experience learning and continuous optimization of the intent decomposition model;

[0095] Interact with the orchestration module to obtain the optimal execution order of all sub-intents;

[0096] The intent decomposition information (including the composition of sub-intents, the execution order of sub-intents, and the information transmission between parent and child intents) is submitted to the execution module.

[0097] Specifically, the execution module can be used to perform the following functions:

[0098] Receive intent decomposition information sent by the decision-making module;

[0099] Send the information of the intent and its sub-intent instances to the intent instance management module;

[0100] Determine if an existing intent instance can be reused;

[0101] Sub-intent instances are generated and executed according to the order of sub-intent execution.

[0102] Specifically, the data acquisition and analysis module can perform the following functions:

[0103] Monitor the operational status of the target domain or domain and perform analysis;

[0104] Collect the relevant information according to the requirements of the cognitive system module.

[0105] Specifically, the intent instance management module can be used to perform the following functions:

[0106] Save the information of the intent instances generated by the execution module to the intent instance library;

[0107] Perform lifecycle management operations on intent instances, such as adding, deleting, and modifying them, and update and manage intent instances.

[0108] Preserve the relationship between the intentions of father and son, and the intentions of predecessors and successors.

[0109] Specifically, the training module can be used to perform the following functions:

[0110] Collect intent input information, system information, and intent decomposition information after intelligent decision-making, generate an experience base for intent decomposition schemes, and continuously optimize and iterate. The generation and optimization of this experience base can use technologies such as deep learning and machine learning as needed.

[0111] Specifically, the intent orchestration module can be used to perform the following functions:

[0112] The intent orchestration module orchestrates the intents decomposed by the decision module, determining the topological execution order of sub-intents. The intent execution order obtained in the intent decision phase is not the optimal execution order of the intents. The intent orchestration module can obtain the optimal execution order of all sub-intents. The intent orchestration module is an optional step.

[0113] The intended orchestration process is as follows: Figure 4 As shown, the specific process may include the following:

[0114] The orchestration module obtains the original execution order of the intent objects after intent decomposition from the intent owner decision module;

[0115] Obtain the orchestration goal. This information can come from the intention owner's expectations. For example, the orchestration goal could be to select the sub-intent object orchestration scheme with the shortest execution time, or it could be to select the sub-intent object orchestration scheme with the least resource consumption.

[0116] Obtain restriction information, which can come from the intentions and expectations of the intention owners, such as resource restrictions or mutually exclusive information between certain intention owners;

[0117] The orchestration goals, constraints, and information on all objects to be orchestrated are used as input to the orchestration scheme acquisition module. This module can obtain corresponding orchestration schemes based on an existing orchestration experience database, such as a commonly used workflow scheduling algorithm. An orchestration scheme selection strategy model training module can also be introduced here. This module can train a strategy selection model based on historical learning experience and save the training results to the orchestration experience database. The database can be continuously enriched to improve the accuracy of scheme acquisition.

[0118] Based on the above information, the selected orchestration scheme is implemented to generate the optimal execution order of intent objects.

[0119] Specifically, the intent experience base can be used to store intent information, intent decomposition schemes for intelligent decision-making, etc. The input information of the intent can be stored in the form of a general intent representation model; the specific form and implementation language of the model are not limited here. The knowledge in the intent experience base comes from two sources: one is the experience pre-set by the designer, and the other is the experience dynamically generated during runtime.

[0120] Intent Experience Pre-setting: Intent designers can decompose and combine intents during the design phase, and pre-set the designed solutions as experience in the intent experience library. The specific design content mainly includes the following three parts:

[0121] The design of parent-child intention relationships mainly involves listing all child intentions under this intention;

[0122] The design of the first child intent: mainly involves listing the first child intent under the parent intent, and this child intent will be executed first during the subsequent intent execution process;

[0123] Subsequent intent design: If the intent is also a child intent of a parent intent, its successor child intent must also be specified. At the same time, for all child intents under the parent intent, starting from the first child intent, its successor intent must be specified, and the successor intent of the last intent points to its parent intent.

[0124] Dynamic generation of intent experience: Intent experience can be dynamically generated by the decision-making module and the training module. The main process is as follows:

[0125] The training module continuously iterates and optimizes the decomposition model based on pre-set experience, as well as the intention input information, collected system information, and intention decomposition information after intelligent decision-making during actual operation, and enriches the content of the experience base.

[0126] When a user's newly input intent cannot be directly decomposed from the intent experience base, a training module can dynamically generate an intent decomposition scheme based on existing decomposition models, intent input information, and data collection system information, and save it to the intent experience base.

[0127] Specifically, the intent instance library can be used to store information about intent instances and the running status of intents.

[0128] In the hierarchical intent execution process of a communication network, intent decomposition mainly involves two levels of decomposition:

[0129] The communication service provider's intent is decomposed into sub-intents of various domains of the network operator, including the network operator's radio side sub-intent, the network operator's bearer network side sub-intent, and the network operator's core network side sub-intent.

[0130] The network operator sub-intents for each domain are further broken down into sub-intents related to network services in each domain.

[0131] Based on the above two levels of decomposition, the process of decomposing and orchestrating intents in a communication network is as follows:

[0132] Figure 5 A flowchart illustrating the intent of a communication network provider, such as Figure 5 As shown, the specific process includes the following:

[0133] (1) The communication service customer, the higher-level intent owner of the communication service provider, sends its intent to the communication service provider's intent management function to the cognitive module. This intent may include expectations regarding bandwidth, reliability, Service Level Agreement (SLA) (e.g., latency, packet loss, jitter), inter-site link requirements (e.g., topology, access protocol), and Quality of Service (QoS) policy information (e.g., platinum policy) for leased line services. At this point, the communication service provider's intent management module becomes the intent processor for the communication service customer's intent.

[0134] (2) The communication service provider's intent recognition module obtains the status of the system or domain related to the intent through the collection and analysis module. For example, information such as the network status and resource status of the wireless side, core network side, and bearer network side in the existing system;

[0135] (3) The communication service provider intent cognition module sends the intent input information and the current system or domain status to the communication service provider intent decision module;

[0136] (4) The communication service provider's intent decision module obtains existing experience from the intent experience base and decomposes the intent into multiple sub-intents based on the experience. A training module related to intent decomposition can also be introduced here. Based on information such as historical learning experience, the decomposition-related model can be trained, and the training results can be saved to the intent experience base.

[0137] (5) The communication service provider intent decision module submits the intent decomposition information to the communication service provider intent execution module. Here, the intent is decomposed into network operator wireless side sub-intent, network operator bearer network side sub-intent, and network operator core network side intent.

[0138] (6) Generate intent instances based on intent information obtained from the execution module, and save the intent instance information to the intent instance library through the intent instance management module;

[0139] (7) The communication service provider intent execution module further sends sub-intent information about each domain (radio side, bearer network side, core network side) of the network operator to the intent management function of each domain. Taking the intent of the bearer network side of the network operator as an example, the sub-intent information here may include, but is not limited to, service type (L3 Virtual Private Network (VPN)), access point device and interface identifier, user network interface (UNI) side configuration information, bandwidth (e.g., 100M), SLA (latency, packet loss, jitter), service topology, QoS policy information (e.g., platinum policy), etc.

[0140] Figure 6 This is a schematic diagram illustrating the decomposition process of the sub-intent of the wireless network operator, such as... Figure 6 As shown, the specific process includes the following:

[0141] (1) The network operator's wireless side intent, which is the intent owner of the communication service provider at the next higher level, sends its intent regarding the wireless network to the cognitive module of the wireless side intent management function. This intent may include coverage area information (e.g., geographical region), wireless setting parameter sets (e.g., frequency information, gNB (Gateway NodeB) identification document (ID) range, Physical Cell Identifier (PCI) range, cell ID range, nRTAC range), transmission setting parameters (e.g., OM transmission information (e.g., OM local IP address, Optical Multi-mode (OM) remote IP address, OM Next HopInfo) and NG transmission information (e.g., NG local IP address list, NG remote IP address list)), and supported network capacity information (e.g., maximum number of UEs) and network performance information (e.g., uplink (UL) / downlink (DL) throughput). Here, NG is the interface between the radio access network and the core network. In this case, the wireless side intent management module becomes the intent processor for the communication service provider's intent.

[0142] (2) The wireless intent recognition module obtains the status of the wireless domain through the data acquisition and analysis module. For example, information such as the status of wireless network services and resources in the existing system;

[0143] (3) The wireless side intent cognition module sends the intent input information and the current system or domain status to the wireless side intent decision module;

[0144] (4) The wireless side intent decision module obtains existing experience from the intent experience base and decomposes the intent into multiple sub-intents based on the experience. For example, here it can be decomposed into multiple network service detection sub-intents, network service deployment sub-intents, network service security isolation sub-intents, and network service assurance sub-intents.

[0145] (5) The wireless intent decision module sends the sub-intent information and orchestration objectives (which may include objectives such as minimum resource consumption, shortest implementation time, and anti-affinity) to the orchestration module. The orchestration module then uses an orchestration algorithm to obtain the optimal execution order of all sub-intents. Here, the optimal execution order can be to execute the probe sub-intent first, then the deployment sub-intent, and finally the security isolation sub-intent and service assurance sub-intent in parallel. If multiple network services exist, the orchestration module can be used to obtain the optimal execution order of each sub-intent for each network service.

[0146] (6) The wireless side intent decision module sends the corresponding data (intent input information, collection system information, and intent decomposition information after intelligent decision) to the training module for model training, and the training module sends the trained results to the intent experience base.

[0147] (7) The wireless side intent decision module submits intent decomposition information (including sub-intent composition, sub-intent execution order, parent-child intent information transmission) to the wireless side intent execution module;

[0148] (8) If an existing intent instance of the same type and with the same parameters is found in the intent obtained in the previous step, it will be reused directly. Otherwise, an intent instance will be generated, and the information of the intent instance will be saved to the intent instance library through the intent instance management module;

[0149] (9) Wireless side intent execution module, which further sends sub-intent information about multiple network services to the intent management function of each network service.

[0150] This application also provides an intent decomposition method, which ultimately decomposes the intent of a communication service customer into the core network intent of the network operator in a communication network. The network operator implements the core network intent using a service slicing technique, as illustrated in the following example:

[0151] like Figure 7 As shown, the slice intents on the user's core network side can be divided into the following sub-intents according to business type and process. These sub-intent divisions and their corresponding models can be saved in the intent experience library in the form of pre-set experiences by the designers.

[0152] Core network side slice probe sub-intent segmentation:

[0153] Sub-intentions of detection technology: including but not limited to digital twin detection sub-intentions and analysis and evaluation detection sub-intentions;

[0154] Sub-intents for probing content dimensions include, but are not limited to, resource pool information probing sub-intents and isolation information probing sub-intents.

[0155] Core network side-slice deployment sub-intent design:

[0156] Sub-slice architecture selects sub-intents, including but not limited to core network element-specific sub-intents, global data network element-shared sub-intents, session-type network element-specific sub-intents, and user plane-specific sub-intents.

[0157] Slice and DNN correspondence selection of sub-intents: including but not limited to one-to-one, one-to-many, many-to-one, etc.;

[0158] Slicing isolation sub-intentions include, but are not limited to, resource pool fully shared networking sub-intentions, server isolation networking intentions, and server and network device fully isolated networking intentions;

[0159] Sliced ​​network isolation sub-intentions: including but not limited to public network slice isolation sub-intentions and industry network slice isolation sub-intentions.

[0160] Core network side slice guarantee sub-intent design:

[0161] Sub-intents for maintaining technical dimensions: sub-intents for alarm monitoring technology, sub-intents for closed-loop technology, and sub-intents for autonomous machine learning;

[0162] Sub-intents for service protection: Premium service protection sub-intent, Exclusive service protection sub-intent, VIP service protection sub-intent, etc.

[0163] The premium service model may have the following requirements, such as dedicated coverage or dedicated frequency, ensuring service quality, supporting service isolation, low latency requirements, data not leaving the premises, edge computing, and uplink enhancement capabilities.

[0164] In the intended operation process, when a premium user requests the creation of a core network subnet slice, it is done through... Figure 8 The illustrated intent decomposition and orchestration process yields a possible execution order for the layered decomposition and orchestration of intents, such as... Figure 8 As shown:

[0165] During the intent execution phase, if a customer requests a slice of the premium business model, and we lack prior experience with this model during the design phase, we can utilize the aforementioned intent decomposition and orchestration process. Based on the model training module introduced during intent decomposition and orchestration, we can dynamically obtain the intent decomposition scheme for the premium business model during the execution phase and save it to the intent experience library. Figure 8 It is an illustration of dynamically combining the intent of preferential services during the operation phase.

[0166] This application proposes a holistic solution for decomposing and orchestrating intents in a communication network. Through the architecture and operational flow described in this application, multi-layered intents can be decomposed into sub-intents, and the optimal intent execution order can be obtained through intent orchestration. The final decomposition and orchestration scheme is then solidified as experience, enabling the hierarchical decomposition and execution of complex intents. Furthermore, a model training module is introduced during the intent decomposition and orchestration process, and machine learning-based model training during the operational flow improves the maturity of the decomposition and orchestration models, thereby enhancing the efficiency of intent decomposition and orchestration.

[0167] See Figure 9 , Figure 9This is a schematic diagram of an intent decomposition device provided in an embodiment of this application. Figure 9 As shown, the intentional decomposition device 900 includes:

[0168] The acquisition module 901 is used to acquire the target intent to be decomposed, and the state information corresponding to the target intent;

[0169] The decomposition module 902 is used to decompose the target intent based on the state information to obtain multiple sub-intents and the execution order of the multiple sub-intents;

[0170] The execution module 903 is used to execute the multiple sub-intents based on the execution order of the multiple sub-intents.

[0171] Optionally, the decomposition module 902 may specifically include:

[0172] The decomposition unit is used to decompose the target intent based on the state information to obtain multiple sub-intents and the initial execution order of the multiple sub-intents;

[0173] The first acquisition unit is used to acquire the intent orchestration target;

[0174] The sorting unit is used to sort the plurality of sub-intents based on the intent orchestration target to obtain the execution order of the plurality of sub-intents.

[0175] Optionally, the sorting unit may specifically include:

[0176] The first acquisition subunit is used to acquire multiple candidate execution orders of the multiple sub-intents;

[0177] A calculation subunit is used to calculate the intent orchestration parameters of the multiple candidate execution orders, wherein the intent orchestration parameters are determined based on the intent orchestration target;

[0178] The first determining subunit is used to determine, among the multiple candidate execution orders, the execution order in which the intended orchestration parameters satisfy a preset condition.

[0179] Optionally, the decomposition unit may specifically include:

[0180] The decomposition subunit is used to decompose the target intent based on the state information to obtain the multiple sub-intents and the topological relationship of the multiple sub-intents;

[0181] The second determining subunit is used to determine the initial execution order of the multiple sub-intents based on their topological relationships.

[0182] Optionally, the decomposition module 902 may specifically include:

[0183] The first determining unit is used to determine the target decomposition scheme of the target intent based on the state information;

[0184] The second acquisition unit is used to acquire the target decomposition scheme from the intent experience base when the intent experience base includes the target decomposition scheme.

[0185] The third acquisition unit is used to input the target intent and the state information into a pre-trained model to predict the decomposition scheme when the intent experience base does not include the target decomposition scheme, so as to obtain the target decomposition scheme.

[0186] The second determining unit is used to determine the plurality of sub-intents and the execution order of the plurality of sub-intents based on the target decomposition scheme.

[0187] Optionally, the execution module 903 may specifically include:

[0188] The fourth acquisition unit is used to acquire an intent instance based on the plurality of sub-intents and the execution order of the plurality of sub-intents;

[0189] An execution unit is configured to execute the plurality of sub-intents based on the intent instance.

[0190] Optionally, the fourth acquisition unit may specifically include:

[0191] The query subunit is used to query the intent instances of the multiple sub-intents in the intent instance library based on the multiple sub-intents and the execution order of the multiple sub-intents;

[0192] The second acquisition subunit is used to acquire intent instances of the plurality of sub-intents from the intent instance library when the intent instance library includes intent instances of the plurality of sub-intents.

[0193] A generation subunit is used to generate intent instances of the plurality of sub-intents when the intent instance library does not include intent instances of the plurality of sub-intents;

[0194] The intention decomposition device 900 also includes:

[0195] The update module is used to update the intent instance library based on the intent instance.

[0196] Optionally, one sub-intent corresponds to one sub-network slice; the execution module 903 may specifically include:

[0197] An execution unit is configured to execute the multiple sub-intents in multiple subnet slices, respectively, based on the execution order of the multiple sub-intents.

[0198] The intention decomposition device 900 can realize the embodiments of this application. Figure 1 The various processes in the method embodiments, and the ways to achieve the same beneficial effects, will not be repeated here to avoid repetition.

[0199] This application also provides an electronic device. Because the principle by which the electronic device solves the problem is similar to that in the embodiments of this application... Figure 1 The intent decomposition method shown is similar; therefore, the implementation of this electronic device can be found in the method implementation, and repeated details will not be elaborated upon. For example... Figure 10 As shown, the electronic device in this application embodiment includes a memory 1020, a transceiver 1010, and a processor 1000;

[0200] The memory 1020 is used to store computer programs; the transceiver 1010 is used to send and receive data under the control of the processor 1000; the processor 1000 is used to read the computer program in the memory 1020 and perform the following operations:

[0201] Obtain the target intent to be decomposed, and the state information corresponding to the target intent;

[0202] Based on the state information, the target intent is decomposed to obtain multiple sub-intents and the execution order of the multiple sub-intents;

[0203] The multiple sub-intents are executed according to their execution order.

[0204] Among them, Figure 10 In this context, the bus architecture may include any number of interconnected buses and bridges, specifically linking various circuits together, represented by one or more processors (processor 1000) and memory (memory 1020). The bus architecture may also link together various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. The bus interface provides an interface. The transceiver 1010 may be multiple elements, including transmitters and transceivers, providing a unit for communicating with various other devices over a transmission medium. The processor 1000 is responsible for managing the bus architecture and general processing, and the memory 1020 may store data used by the processor 1000 during operation.

[0205] The processor 1000 can be a central processing unit (CPU), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a complex programmable logic device (CPLD). The processor can also adopt a multi-core architecture.

[0206] Optionally, the step of decomposing the target intent based on the state information to obtain multiple sub-intents and the execution order of the multiple sub-intents includes:

[0207] Based on the state information, the target intent is decomposed to obtain multiple sub-intents and the initial execution order of the multiple sub-intents;

[0208] Obtain the intent and orchestration goals;

[0209] The multiple sub-intents are sorted based on the intent orchestration target to obtain the execution order of the multiple sub-intents.

[0210] Optionally, the step of sorting the plurality of sub-intents based on the intent orchestration target to obtain the execution order of the plurality of sub-intents includes:

[0211] Obtain multiple candidate execution orders for the multiple sub-intents;

[0212] The intent orchestration parameters for the various candidate execution orders are calculated respectively, and the intent orchestration parameters are determined based on the intent orchestration target;

[0213] Among the various candidate execution sequences, the execution sequence that satisfies the preset conditions for the intended orchestration parameters is determined.

[0214] Optionally, the step of decomposing the target intent based on the state information to obtain multiple sub-intents and the initial execution order of the multiple sub-intents includes:

[0215] Based on the state information, the target intent is decomposed to obtain the multiple sub-intents and the topological relationship of the multiple sub-intents;

[0216] Based on the topological relationship of the multiple sub-intents, the initial execution order of the multiple sub-intents is determined.

[0217] Optionally, the step of decomposing the target intent based on the state information to obtain multiple sub-intents and the execution order of the multiple sub-intents includes:

[0218] Based on the state information, a target decomposition scheme for the target intent is determined;

[0219] If the intent experience base includes the target decomposition scheme, the target decomposition scheme is obtained from the intent experience base;

[0220] If the intent experience base does not include the target decomposition scheme, the target intent and the state information are input into a pre-trained model to predict the decomposition scheme in order to obtain the target decomposition scheme.

[0221] Based on the target decomposition scheme, the plurality of sub-intents and the execution order of the plurality of sub-intents are determined.

[0222] Optionally, executing the plurality of sub-intents based on the execution order of the plurality of sub-intents includes:

[0223] Based on the multiple sub-intents and their execution order, an intent instance is obtained;

[0224] Based on the intent instance, execute the plurality of sub-intents.

[0225] Optionally, obtaining an intent instance based on the plurality of sub-intents and their execution order includes:

[0226] Based on the multiple sub-intents and their execution order, query the intent instances of the multiple sub-intents in the intent instance library;

[0227] If the intent instance library includes intent instances of the plurality of sub-intents, then the intent instances of the plurality of sub-intents are obtained from the intent instance library;

[0228] If the intent instance library does not include intent instances of the multiple sub-intents, then intent instances of the multiple sub-intents are generated.

[0229] The processor 1000 is also used to read the program from the memory 1020 and perform the following steps:

[0230] The intent instance library is updated based on the intent instance.

[0231] Optionally, one sub-intention corresponds to one sub-network slice; the execution of the multiple sub-intentions based on their execution order includes:

[0232] Based on the execution order of the multiple sub-intents, the multiple sub-intents are executed in multiple subnet slices respectively.

[0233] The electronic device provided in this application embodiment can perform the above-described functions. Figure 1 The method embodiments shown are similar in principle and technical effect, and will not be described again here.

[0234] This application also provides a readable storage medium storing a program that, when executed by a processor, implements the following... Figure 1 The various processes in the Chinese method embodiment can achieve the same technical effect, and will not be described again here to avoid repetition.

[0235] In the several embodiments provided in this application, it should be understood that the disclosed methods and apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0236] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can be physically included separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0237] The integrated units implemented as software functional units described above can be stored in a computer-readable storage medium. These software functional units, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute some steps of the transmission and reception methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0238] The above description is the preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principles described in this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. An intent decomposition method, characterized in that, include: Obtain the target intent to be decomposed, and the state information corresponding to the target intent; Based on the state information, the target intent is decomposed to obtain multiple sub-intents and the execution order of the multiple sub-intents; Execute the multiple sub-intents according to their execution order; The status information corresponding to the target intent includes network status information or resource status information of the system or domain related to the target intent; The step of decomposing the target intent based on the state information to obtain multiple sub-intents and the execution order of the multiple sub-intents includes: Based on the state information, a target decomposition scheme for the target intent is determined; If the intent experience base includes the target decomposition scheme, the target decomposition scheme is obtained from the intent experience base; If the intent experience base does not include the target decomposition scheme, the target intent and the state information are input into a pre-trained model to predict the decomposition scheme in order to obtain the target decomposition scheme. Based on the target decomposition scheme, the plurality of sub-intents and the execution order of the plurality of sub-intents are determined.

2. The method as described in claim 1, characterized in that, The step of decomposing the target intent based on the state information to obtain multiple sub-intents and the execution order of the multiple sub-intents includes: Based on the state information, the target intent is decomposed to obtain multiple sub-intents and the initial execution order of the multiple sub-intents; Obtain the intent and orchestration goals; The multiple sub-intents are sorted based on the intent orchestration target to obtain the execution order of the multiple sub-intents.

3. The method as described in claim 2, characterized in that, The step of sorting the multiple sub-intents based on the intent orchestration target to obtain the execution order of the multiple sub-intents includes: Obtain multiple candidate execution orders for the multiple sub-intents; The intent orchestration parameters for the various candidate execution orders are calculated respectively, and the intent orchestration parameters are determined based on the intent orchestration target; Among the various candidate execution sequences, the execution sequence that satisfies the preset conditions for the intended orchestration parameters is determined.

4. The method as described in claim 2, characterized in that, The step of decomposing the target intent based on the state information to obtain multiple sub-intents and the initial execution order of the multiple sub-intents includes: Based on the state information, the target intent is decomposed to obtain the multiple sub-intents and the topological relationship of the multiple sub-intents; Based on the topological relationship of the multiple sub-intents, the initial execution order of the multiple sub-intents is determined.

5. The method as described in claim 1, characterized in that, The execution of the multiple sub-intents based on their execution order includes: Based on the multiple sub-intents and their execution order, an intent instance is obtained; Based on the intent instance, execute the plurality of sub-intents.

6. The method as described in claim 5, characterized in that, The step of obtaining an intent instance based on the plurality of sub-intents and their execution order includes: Based on the multiple sub-intents and their execution order, query the intent instances of the multiple sub-intents in the intent instance library; If the intent instance library includes intent instances of the plurality of sub-intents, then the intent instances of the plurality of sub-intents are obtained from the intent instance library; If the intent instance library does not include intent instances of the multiple sub-intents, then intent instances of the multiple sub-intents are generated. After obtaining the intent instance based on the plurality of sub-intents and their execution order, the method further includes: The intent instance library is updated based on the intent instance.

7. The method according to any one of claims 1 to 6, characterized in that, One sub-intention corresponds to one sub-network slice; the execution of the multiple sub-intentions based on their execution order includes: Based on the execution order of the multiple sub-intents, the multiple sub-intents are executed in multiple subnet slices respectively.

8. An intention decomposition device, characterized in that, include: The acquisition module is used to acquire the target intent to be decomposed, and the state information corresponding to the target intent; The decomposition module is used to decompose the target intent based on the state information to obtain multiple sub-intents and the execution order of the multiple sub-intents; An execution module is used to execute the multiple sub-intents based on their execution order. The status information corresponding to the target intent includes network status information or resource status information of the system or domain related to the target intent; The decomposition module includes: The first determining unit is used to determine the target decomposition scheme of the target intent based on the state information; The second acquisition unit is used to acquire the target decomposition scheme from the intent experience base when the intent experience base includes the target decomposition scheme. The third acquisition unit is used to input the target intent and the state information into a pre-trained model to predict the decomposition scheme when the intent experience base does not include the target decomposition scheme, so as to obtain the target decomposition scheme. The second determining unit is used to determine the plurality of sub-intents and the execution order of the plurality of sub-intents based on the target decomposition scheme.

9. An electronic device, characterized in that, include: A transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor; characterized in that, The processor is configured to read a program from memory to implement the steps of the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium for storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 7.

Images