Solution orchestration method, device, and storage medium

Abstract

The present application provides a solution orchestration method, a device, and a storage medium. The solution orchestration method applied to a first management device comprises: generating a target intent on the basis of received service requirement information; sending an intent negotiation request to a second management device on the basis of the target intent, and requesting to return candidate collaborative processing solution sets; and performing intent negotiation with the second management device on the basis of the candidate collaborative processing solution sets to determine a target collaborative processing solution set.

Description

Program arrangement method, device and storage medium Technical Field

[0001] The present application relates to the field of communication technology, and in particular to a scheme arrangement method, device and storage medium. Background Art

[0002] Against the backdrop of automated and intelligent network management, a large number of artificial intelligence / machine learning (AI / ML) functions have been proposed, including radio access network AI (RAN AI) capabilities, network data analytics functions (NWDAF), and management data analytics functions (MDAF). When specific business needs are put forward and network operators are faced with numerous network models, they need to balance computing resources with model training / inference performance. However, given the complex inference requirements and dynamic network computing resources, network operators are unable to determine the optimal collaborative processing solution based on the traditional manual parameter configuration and delivery method.

[0003] Summary of the Invention

[0004] In view of this, embodiments of the present application provide a solution arrangement method, device, and storage medium, which effectively determine the optimal collaborative processing solution.

[0005] An embodiment of the present application provides a solution arrangement method, applied to a first management device, including:

[0006] Generate target intent based on received business demand information;

[0007] Sending an intent negotiation request to the second management device based on the target intent, requesting the return of a set of candidate collaborative processing solutions;

[0008] An intent negotiation is performed with the second management device based on the candidate collaborative processing solution set to determine a target collaborative processing solution set.

[0009] An embodiment of the present application provides a solution arrangement method, which is applied to a second management device, including:

[0010] receiving an intention negotiation request sent by the first management device;

[0011] Determining a corresponding set of candidate collaborative processing solutions at least according to the intended negotiation request;

[0012] The candidate collaborative processing solution set is sent to the first management device, so that the first management device performs intent negotiation with the second management device based on the candidate collaborative processing solution set to determine a target collaborative processing solution set.

[0013] An embodiment of the present application provides a communication device, comprising: a memory, and one or more processors;

[0014] The memory is configured to store one or more programs;

[0015] When the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any one of the above embodiments.

[0016] An embodiment of the present application provides a storage medium storing a computer program. When the computer program is executed by a processor, the method described in any one of the above embodiments is implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] FIG1 is a schematic diagram of an implementation of a service-oriented management architecture provided by related technologies;

[0018] FIG2 is a configuration diagram of an intent-driven management service model provided by the related art;

[0019] FIG3 is a flow chart of a method for arranging a solution according to an embodiment of the present application;

[0020] FIG4 is a flow chart of another solution arrangement method provided in an embodiment of the present application;

[0021] FIG5 is an interactive diagram of a solution arrangement provided by an embodiment of the present application;

[0022] FIG6 is an interactive diagram of another solution arrangement provided in an embodiment of the present application;

[0023] FIG7 is a schematic diagram of an interaction arrangement of another solution provided in an embodiment of the present application;

[0024] FIG8 is a schematic diagram of an interaction arrangement according to another embodiment of the present application;

[0025] FIG9 is a structural block diagram of a program arrangement device provided in an embodiment of the present application;

[0026] FIG10 is a structural block diagram of another solution arrangement device provided in an embodiment of the present application;

[0027] FIG11 is a schematic structural diagram of a communication device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0028] The following describes the embodiments of the present application in conjunction with the accompanying drawings. The following describes the present application in conjunction with the accompanying drawings. The examples are only used to explain the present application and are not used to limit the scope of the present application.

[0029] Figure 1 is a schematic diagram of the implementation of a service-based management architecture provided by related technologies. As shown in Figure 1, the service-based management architecture (Service Based Management Architecture) includes: a business support system (BSS), a cross-domain management function unit (Cross Domain Management Function, CD-MnF), a domain management function unit (Domain Management Function, Domain-MnF) and a network element (NE). Among them, the cross-domain management function unit is used to manage one or more domain management function units. The domain management function unit can be used to manage one or more network elements. The following is a brief introduction to each unit.

[0030] Business support systems (BSSs) are designed for communications services and provide functions and management services such as billing, settlement, accounting, customer service, sales operations, network monitoring, communications service lifecycle management, and business intent translation. BSSs can be either carrier operations systems or vertical industry operations systems (VT Systems).

[0031] The cross-domain management function unit is also called the network management function unit (Network Management Function, NMF), which can be a network management entity such as the network management system (NMS), the network management service producer (MnS Producer), the network management service consumer (MnS Consumer) and the network function management service consumer (NFMS_C). Among them, the cross-domain management function unit provides one or more of the following management functions or management services: network lifecycle management, network deployment, network fault management, network performance management, network configuration management, network assurance, network optimization functions and translation of the network intent (intent from communication service provider, Intent-CSP) of the service producer. The network referred to in the above management functions or management services may include one or more network elements or subnetworks, or it may be a network slice. That is to say, the network management functional unit can be a network slice management functional unit (Network Slice Management Function, NSMF), or a cross-domain management data analysis functional unit (Management Data Analytical Function, MDAF), or a cross-domain self-organizing network function (Self-Organization Network function, SON Function) or a cross-domain intent management functional unit (Intent-Driven Management Service, Intent Driven MnS).

[0032] The domain management functional unit is also called the Network Subnet Management Function (NSMF) or the network element management functional unit. It can be a network element management entity such as the wireless automation engine (MAE), the element management system (EMS), the network function management service provider (NFMS_P), the network slice subnet management functional unit (NSSMF), the domain management data analysis functional unit (Domain MDAF), the domain self-organization network function (SON Function), the domain intent management functional unit, the MnS Producer, and the MnS Consumer. Domain management function units can be categorized as follows: by network type, they can be divided into: radio access network (RAN) domain management function unit (RAN domain MnF), core network domain management function unit (CN domain MnF), transport network domain management function unit (TN domain MnF), etc. It should be noted that the domain management function unit can also be a domain network management system that can manage one or more of the access network, core network, or transport network. By administrative region, they can be divided into: domain management function units for a certain region, such as the domain management function unit of city A and the domain management function unit of city B. Domain management function units provide one or more of the following functions or management services: lifecycle management of subnetworks or network elements, deployment of subnetworks or network elements, fault management of subnetworks or network elements, performance management of subnetworks or network elements, assurance of subnetworks or network elements, optimization of subnetworks or network elements, and translation of intents from network operators (Intent-NOPs) for subnetworks or network elements. The subnetwork here includes one or more network elements.A subnetwork can also include subnetworks, that is, one or more subnetworks form a larger subnetwork. The subnetwork here can also be a network slice subnetwork.

[0033] A network element is an entity that provides network services. Network elements may include core network elements, radio access network elements, or transport network elements. Specifically, core network elements may include, but are not limited to, Access and Mobility Management Function (AMF) entities, Session Management Function (SMF) entities, Policy Control Function (PCF) entities, Network Data Analysis Function (NWDAF) entities, Network Repository Function (NRF) entities, and gateways. The radio access network network element may include but is not limited to: various base stations (such as the next generation base station (Generation Node B, gNB), evolved base station (Evolved Node B, eNB), Central Unit Control Panel (CUCP), Central Unit (CU), Distributed Unit (DU), Centralized User Plane Unit (CUUP), etc. In this application, the network function NF is also referred to as the network element NE. The network element can provide one or more of the following management functions or management services: network element lifecycle management, network element deployment, network element fault management, network element performance management, network element assurance, network element optimization function, and network element intent translation.

[0034] The cross-domain management functional unit can be used for model training of AI / ML models and model reasoning of AI / ML models; the domain management functional unit can be used for model training of AI / ML models and model reasoning of AI / ML models; the network element is an entity that provides network services, including core network network elements and access network network elements, which can provide at least one of model training of AI / ML models and model reasoning of AI / ML models.

[0035] Figure 2 is a configuration diagram of an intent-driven management service model provided by related technologies. As shown in Figure 2, the intent-driven calendar service model includes: a management service consumer (MnS Consumer) and a management service producer (MnS Producer); among them, the management service consumer (MnS Consumer) is the intent owner (Intent Owner), and the management service producer (MnS Producer) is the intent processor.

[0036] If the management service is provided by a cross-domain management functional unit, the cross-domain management functional unit is the management service producer, and the business support system is the management service consumer. If the management service is provided by a domain management functional unit, the domain management functional unit is the management service producer, and the cross-domain management functional unit is the management service consumer. If the management service is provided by a network element, the network element is the management service producer, and the domain management functional unit is the management service consumer. An intent can be a desire for the intent producer (e.g., a network element) or the system in which the intent producer resides (e.g., a network or subnetwork), and can include requirements, goals, or constraints. The translation of an intent refers to the process of determining the intent. For example, a policy can be used to indicate conditions that do not meet the intent. For example, if the intent is energy saving, Policy A could be: If energy consumption exceeds a first threshold, energy consumption is abnormal (i.e., no energy saving); Policy B could be: If energy consumption exceeds a second threshold, energy consumption is abnormal (i.e., no energy saving). It is understood that even for the same intent, different policies may determine different solutions to meet the intent.

[0037] In order to facilitate the understanding of the implementation scheme of the present application, the parameters involved in the implementation scheme of the present application are explained as follows.

[0038] ML Entity: A manageable entity of an ML model.

[0039] NOTE 1: An ML entity may contain metadata related to an ML model. The metadata may include, for example, the applicable runtime context of the ML model.

[0040] AI Decision Entity: An entity that applies non-ML logic to make decisions and can be managed as a single composite entity.

[0041] ML Model: A mathematical algorithm that can be "trained" using data and human expert input as examples to replicate the decisions that an expert would make when given the same information.

[0042] ML Model Training: The process performed by the ML training function, which is to obtain training data, run it through the ML model, calculate the associated loss, and adjust the parameters of the ML model based on the calculated loss.

[0043] ML Initial Training: ML model training to generate the initial version of the ML entity.

[0044] ML Re-training: ML model training that generates new versions of trained ML entities using training data of the same type but different values ​​or distributions.

[0045] Note 3: A trained new version of an ML entity supports the same type of inference as the previous version of the ML entity, i.e., the data type of the inference input and the data type of the inference output remain unchanged between the two versions of the ML entity.

[0046] ML Training: refers to the end-to-end process that enables the ML training function to perform ML model initialization.

[0047] ML Training Function: A logical function that performs ML modeling training; also known as an MLT function.

[0048] AI / ML Inference Function: A logical function that uses ML models and / or AI decision-making entities to perform reasoning.

[0049] In view of this, an embodiment of the present application provides an orchestration scheme for collaborative processing solutions based on intent negotiation, which supports dynamic orchestration of collaborative processing solutions based on target intent, entity capability information and current computing power resource conditions generated according to business demand information.

[0050] It should be noted that the first management device in the embodiment of the present application may be a management service consumer, and the second management device may be a management service producer. For example, management service consumers may include but are not limited to: ML training management service consumers or AI / ML reasoning management service consumers; management service producers may include but are not limited to: ML training management service producers or AI / ML reasoning management service producers. In one example, when the ML training management service consumer is a cross-domain management functional unit, the ML training management service producer is a domain management functional unit; in one example, when the AI / ML reasoning management service consumer is a cross-domain management functional unit, the AI / ML reasoning management service producer is a domain management functional unit.

[0051] In one embodiment, FIG3 is a flowchart of a solution orchestration method provided by an embodiment of the present application. This embodiment is applied to the case of automatically orchestrating collaborative processing solutions. This embodiment can be executed by a first management device. As shown in FIG3, this embodiment includes: S310-S330.

[0052] S310: Generate target intent according to the received business requirement information.

[0053] The business requirement information is used to represent the requirement information that satisfies the business function. Generally speaking, the business requirement information can be in natural language, and the business requirement information can be converted to obtain the target intent of the structured language.

[0054] S320: Send an intent negotiation request to the second management device based on the target intent, requesting the return of a set of candidate collaborative processing solutions.

[0055] The set of candidate collaborative processing solutions may include one candidate collaborative processing solution or multiple candidate collaborative processing solutions. In one example, the candidate collaborative processing solutions may include but are not limited to at least one of the following: candidate collaborative training solutions; candidate collaborative reasoning solutions. For example, the candidate collaborative training solutions may include: candidate ML collaborative training solutions; candidate collaborative reasoning solutions may include: candidate AI / ML collaborative reasoning solutions. After the first management device receives the business requirement information, it converts the business requirement information, generates a target intent, and sends an intent negotiation request to the second management device based on the target intent, requesting the second management device to return a set of candidate collaborative processing solutions that matches the target intent.

[0056] In one example, after receiving the target intent sent by the first management device, the second management device can dynamically determine a corresponding set of candidate collaborative processing solutions based on the target intent, pre-acquired entity capability information, and current computing resource conditions. The current computing resource conditions are used to represent the usage of all computing resources managed by the second management device. For example, the second management device determines the set of collaborative processing solutions based on the computing resource conditions of all central processing units (CPUs) or servers contained in a region.

[0057] S330: Perform intent negotiation with the second management device based on the candidate collaborative processing solution set to determine a target collaborative processing solution set.

[0058] Among them, the target collaborative processing scheme set can be a subset of the candidate collaborative processing scheme set, that is, the number of collaborative processing schemes included in the target collaborative processing scheme set is less than or equal to the number of collaborative processing schemes included in the target collaborative processing scheme set. Among them, the target collaborative processing scheme set is used to characterize at least one target collaborative processing scheme determined by the first management device. After the second management device returns a candidate collaborative processing scheme set that matches the target intent, the first management device negotiates intent with the second management device based on the candidate collaborative processing scheme set, and composes one or more candidate collaborative processing schemes selected by the first management device from the candidate collaborative processing scheme set into a corresponding target collaborative processing scheme set, thereby achieving the effect that the first management device can automatically determine the target collaborative processing scheme based on the target intent. In one example, when the first management device does not select one of the candidate collaborative processing schemes from the candidate collaborative processing scheme set, the number of collaborative processing schemes included in the target collaborative processing scheme set is 0.

[0059] In one embodiment, generating a target intent based on received business requirement information includes converting the received business requirement information to generate a corresponding target intent, wherein the business requirement information includes at least an AI / ML requirement. In one example, the first management device converts the received AI / ML requirement (e.g., in natural language) to generate corresponding AI / ML requirement information (e.g., in structured language), and then identifies and extracts the intent from the AI / ML requirement information as the corresponding target intent.

[0060] In one embodiment, the category of the AI / ML requirement of the target intent includes at least one of the following: reasoning requirement; training requirement. In one example, the reasoning requirement may be an AI / ML reasoning requirement; and the training requirement may be an ML training requirement.

[0061] In one embodiment, the target intent includes at least one of the following parameters: reasoning category; model collaboration strategy; model version; expected output result; input data; used entity; model performance indicator;

[0062] Model performance metrics include at least one of the following parameters: accuracy; precision; recall; precision-recall (PR) curve; and F1 score. The inference category indicates the inference type of the entity and is generally determined by the inference or training requirements. For example, the inference category may include, but is not limited to, the Model Driven Architecture (MDA) type and the analysis ID (denoted as Analytic ID). The model collaboration strategy indicates the AI / ML model's runtime (e.g., 10 hours) or the model runtime period (e.g., time point A to time point B). It may also indicate the ML model's training duration (e.g., 11 hours) or the model training period (e.g., time point C to time point D). The model version indicates the version information of the AI / ML model or ML model. The expected output indicates the user's expected inference result and is generally determined by specific requirements. The input data indicates the input data or input data address provided by the user. The entity used indicates the AI / ML entity or ML entity specified by the user for use. Model performance metrics indicate the performance metrics of the AI / ML model or ML model.

[0063] In one embodiment, the target intent includes an intent negotiation category, which indicates the category of the intent negotiation request. In one example, the first management device can implicitly send the target intent to the second management device by including the target intent and the intent negotiation category in the intent negotiation request. The first management device then sends an intent negotiation request to the second management device based on the intent negotiation category, requesting the second management device to return a set of candidate collaborative processing solutions that match the intent negotiation category.

[0064] In one embodiment, the intent negotiation category includes at least one of the following: optimal parameter negotiation and specific strategy negotiation. In one example, optimal parameter negotiation is used to indicate the optimal value that can be achieved for at least one parameter in the target intent returned by the second management device, and its corresponding set of candidate collaborative processing solutions. For example, the optimal parameter may include, but is not limited to, one of the following: shortest training time; minimum training price; minimum energy consumption. In one example, the specific strategy is used to indicate the set of candidate collaborative processing solutions generated based on the target intent and a specific strategy returned by the second management device. For example, the specific strategy may include, but is not limited to, one of the following: maximum training accuracy strategy; optimal demand matching strategy.

[0065] In one embodiment, when the intent negotiation category is optimal parameter negotiation, the candidate collaborative processing solution set includes at least one of the following: the optimal value achieved by at least one parameter in the target intent; and the corresponding candidate collaborative processing solution. In one example, when the intent negotiation category is optimal parameter negotiation, the candidate collaborative processing solution set returned by the second management device to the first management device may include the optimal value achieved by at least one parameter in the target intent, or may include the optimal value achieved by at least one parameter in the target intent and its corresponding candidate collaborative processing solution.

[0066] In one embodiment, when the intended negotiation category is specific policy negotiation, the candidate collaborative processing solution set includes at least one candidate collaborative processing solution generated based on the specific policy.

[0067] In one embodiment, the intent negotiation request includes: an intent negotiation operation; sending the intent negotiation request to the second management device based on the target intent, requesting the return of a set of candidate collaborative processing solutions, including:

[0068] Based on the target intent, the first management device initiates an intent negotiation operation with the second management device, requesting the return of a set of candidate collaborative processing solutions. In one example, the first management device may also explicitly initiate an intent negotiation with the second management device, requesting the return of a set of candidate collaborative processing solutions.

[0069] In one embodiment, the intent negotiation operation includes at least one of the following: optimal parameter intent negotiation; specific policy intent negotiation. In one example, a first management device initiates optimal parameter intent negotiation with a second management device based on a target intent, requesting the second management device to return optimal parameters and their corresponding candidate collaborative processing solutions. In another example, the first management device initiates specific policy intent negotiation with a second management device based on the target intent, requesting the second management device to return at least one candidate collaborative processing solution.

[0070] In one embodiment, when the optimal parameter intent negotiation is initiated, the candidate collaborative processing solution set includes at least one of the following: an optimal value obtained by at least one parameter in the target intent; and a corresponding candidate collaborative processing solution.

[0071] In one embodiment, when negotiation of a specific policy intention is initiated, the candidate co-processing solution set includes at least one candidate co-processing solution generated based on the specific policy.

[0072] In one embodiment, the solution orchestration method applied to the first management device further includes: receiving solution validity information for each candidate collaborative processing solution in the set of candidate collaborative processing solutions. In one example, the solution validity information is used to indicate the solution validity period corresponding to each candidate collaborative processing solution. For example, when the AI / ML requirement category is an ML training requirement, the candidate collaborative processing solution may be a candidate collaborative training solution. Correspondingly, the solution validity information refers to the validity period of the shortest expected training time of the ML model.

[0073] In one embodiment, the solution validity period information corresponding to each candidate collaborative processing solution is unique; alternatively, the solution validity period information corresponding to each candidate collaborative processing solution is the same. In one example, the second management device may set an independent and unique solution validity period for each candidate collaborative processing solution, or may set a common solution validity period for all candidate collaborative processing solutions.

[0074] In one embodiment, the scheme arrangement method applied to the first management device further includes: sending an intent negotiation reply message to the second management device within the scheme validity period corresponding to the scheme validity period information; wherein the intent negotiation reply message at least includes a determined target collaborative processing scheme set; the target collaborative processing scheme set is at least one candidate collaborative processing scheme in the candidate collaborative processing scheme set. After the second management device returns the candidate collaborative processing scheme set to the first management device, the second management device starts the scheme validity period timer and waits for the first management device to send the intent negotiation reply message. If the first management device selects at least one candidate collaborative processing scheme from the candidate collaborative processing scheme set within the scheme validity period and generates a corresponding target collaborative processing scheme set, the first management device sends the intent negotiation reply message to the second management device within the scheme validity period, so that the second management device executes the target collaborative processing scheme in the target collaborative processing scheme set.

[0075] In one embodiment, the solution arrangement method applied to the first management device further includes: receiving intention negotiation result information returned by the second management device.

[0076] In one embodiment, the intended negotiation result information includes at least one of the following: negotiation result; negotiation success status information; negotiation failure status information and negotiation failure reason;

[0077] The negotiation failure status information includes at least one of the following: automatic execution of the plan; intention suspended; intention failed. In one example, the negotiation result may include one of the following: successful negotiation; failed negotiation. In one example, when the negotiation result is a successful negotiation, the intention negotiation result information also includes: negotiation success status information. For example, the negotiation success status information may include but is not limited to one of the following: the collaborative processing plan is being executed; the collaborative processing plan is to be executed. In one example, when the negotiation result is a failed negotiation, the intention negotiation result information also includes: negotiation failure status information and the reason for the negotiation failure. In one example, the negotiation failure status information may include one of the following: automatic execution of the plan; intention suspended; intention failed. In one example, the reason for the negotiation failure may include one of the following: no intention negotiation reply message was received within the validity period.

[0078] In one embodiment, when the negotiation failure status information indicates an intent suspension, the solution orchestration method applied to the first management device further includes: sending an intent recovery indication message to the second management device; receiving a new set of candidate collaborative processing solutions re-arranged by the second management device based on the intent recovery indication message, current computing power resource conditions, and entity capability information; and determining at least one new target collaborative processing solution set from the new set of candidate collaborative processing solutions. The intent recovery indication message is used to request the resumption of execution of the intent. If the second management device does not receive an intent negotiation reply message sent by the first management device within the validity period of the solution, the second management device may suspend the intent; then, the second management device may send a negotiation failure status message carrying the intent suspension to the first management device; then, the first management device may send an intent recovery indication message to the second management device to request the second management device to resume execution of the intent; the second management device re-determines a new set of candidate collaborative processing solutions based on the current computing power resource conditions, the intent recovery indication message, and entity capability information, and sends the new set of candidate collaborative processing solutions to the first management device, so that the first management device determines at least one new target collaborative processing solution set from the new set of candidate collaborative processing solutions. In one example, while the second management device sends a new candidate collaborative processing solution set to the first management device, the second management device may also send solution validity period information corresponding to each new candidate collaborative processing solution to the first management device.

[0079] In one embodiment, FIG4 is a flowchart of another solution orchestration method provided by an embodiment of the present application. This embodiment is applied to the case of automatically orchestrating collaborative processing solutions. This embodiment can be executed by a first management device. As shown in FIG4, this embodiment includes: S410-S430.

[0080] S410: Receive a negotiation intention request sent by a first management device.

[0081] S420: Determine a corresponding candidate collaborative processing solution set at least according to the intended negotiation request.

[0082] S430: Send the candidate collaborative processing solution set to the first management device, so that the first management device performs intent negotiation with the second management device based on the candidate collaborative processing solution set to determine a target collaborative processing solution set.

[0083] In one embodiment, the solution orchestration method applied to the second management device further includes: sending corresponding intent negotiation result information to the first management device. The intent negotiation result information indicates the result of the intent negotiation between the first management device and the second management device. The intent negotiation result information includes at least one of the following: a negotiation result; negotiation success status information; negotiation failure status information and a negotiation failure reason; wherein the negotiation failure status information includes at least one of the following: automatic solution execution; intent suspension; or intent failure.

[0084] In one embodiment, the solution arrangement method applied to the second management device further includes: acquiring entity capability information of each entity.

[0085] In one embodiment, entity capability information includes at least one of the following: entity identification; entity version; entity reasoning category; entity training context; entity performance indicators. In one example, the entity may include one of the following: AI / ML entity; ML entity. The entity identification is used to indicate the unique identification of each entity; the entity version is used to indicate the version information of each entity; the entity reasoning category is used to indicate the reasoning type of each entity, for example, it may include MDA type and analysis ID, etc.; the entity training context includes training-related information for each entity, for example, it may include ML entity training time information, ML entity training resource consumption information, and ML entity training data information, etc.; the entity performance indicators include performance indicators supported by each entity, such as accuracy, precision, recall, PR curve, and F1 score, etc.

[0086] In one embodiment, a corresponding set of candidate collaborative processing solutions is determined at least based on the intent negotiation request, including: determining a corresponding solution orchestration strategy based on the target intent in the intent negotiation request; and determining a corresponding set of candidate collaborative processing solutions based on the solution orchestration strategy, entity capability information, and current computing power resource conditions. In one example, the solution orchestration strategy is used to characterize the entity used by each candidate collaborative processing solution, as well as the computing power resources that can be adopted (i.e., the computing power resources allocated by the second management device to each candidate collaborative processing solution). In the case where the first management device sends the target intent to the second management device in an implicit manner, the target intent can be carried in the intent negotiation request, so that the second management device can determine the corresponding solution orchestration strategy based on the target intent, and determine the corresponding set of candidate collaborative processing solutions based on the solution orchestration strategy, entity capability information, and current computing power resource conditions.

[0087] In one embodiment, a corresponding set of candidate collaborative processing solutions is determined based at least on an intent negotiation request, including: determining a corresponding solution orchestration strategy based on the target intent and the intent negotiation operation in the intent negotiation request; and determining a corresponding set of candidate collaborative processing solutions based on the solution orchestration strategy, entity capability information, and current computing power resource conditions. In one example, the first management device can explicitly initiate intent negotiation, i.e., the first management device sends a target intent to the second management device and initiates an intent negotiation operation to the second management device, so that the second management device determines a corresponding solution orchestration strategy based on the target intent and the intent negotiation operation. Furthermore, a corresponding set of candidate collaborative processing solutions is determined based on the solution orchestration strategy, entity capability information, and current computing power resource conditions.

[0088] In one embodiment, the solution arrangement method applied to the second management device further includes: sending solution validity period information of each candidate collaborative processing solution in the candidate collaborative processing solution set to the first management device.

[0089] In one embodiment, the solution orchestration method applied to the second management device further includes: receiving, within the solution validity period corresponding to the solution validity period information, a negotiation intent reply message sent by the first management device; wherein the negotiation intent reply message at least includes a determined target collaborative processing solution set; the target collaborative processing solution set is at least one candidate collaborative processing solution in the candidate collaborative processing solution set;

[0090] At least one target co-processing solution in the target co-processing solution set is executed.

[0091] In one embodiment, the solution orchestration method applied to the second management device further includes releasing computing resources unrelated to the target collaborative processing solution. Upon receiving a negotiation intent reply message from the first management device during the validity period of the solution, the second management device may execute the target collaborative processing solution from the negotiated set of target collaborative processing solutions and release computing resources not involved in the target collaborative processing solution, thereby fully utilizing the computing resources.

[0092] In one embodiment, when no negotiation intention reply message sent by the first management device is received within the plan validity period corresponding to the plan validity period information, the plan arrangement method applied to the second management device further includes:

[0093] Generate corresponding negotiation failure status information based on a pre-configured solution orchestration policy; the negotiation failure status information includes at least one of the following: automatic solution execution; pending intent; or failed intent. If the second management device does not receive a negotiation intent reply message from the first management device within the solution validity period, the second management device may automatically execute the previously determined target collaborative processing solution, suspend the intent, or determine that the intent has failed, based on the pre-configured solution orchestration policy.

[0094] In one embodiment, when the negotiation failure status information is an intention to suspend, the solution arrangement method applied to the second management device also includes: receiving the intention to recover indication information sent by the first management device; re-arranging based on the intention to recover indication information, the current computing power resource situation and the entity capability information to generate a new set of candidate collaborative processing solutions; and sending the new set of candidate collaborative processing solutions to the first management device.

[0095] It should be noted that for the explanation of parameters such as target intent, candidate collaborative processing solution set, target collaborative processing solution set, intent negotiation result information, entity capability information, and intent recovery indication information in the embodiment corresponding to the solution orchestration method applied to the second management device, please refer to the description of the corresponding parameters in the above-mentioned embodiment corresponding to the solution orchestration method applied to the first management device, and will not be repeated here.

[0096] In one embodiment, FIG5 is an interactive diagram of a solution arrangement provided by an embodiment of the present application. Taking the first management device as the AI / ML reasoning management service consumer, the second management device as the AI / ML reasoning management service producer, the entity used as the AI / ML entity (including AI / ML entity 1, AI / ML entity 2...AI / ML entity N), the collaborative processing solution as the AI / ML collaborative reasoning solution, the business demand information as the AI / ML reasoning demand, and the target intent as the AI / ML reasoning demand information as an example, the arrangement process of the collaborative processing solution is described. As shown in FIG5, the arrangement process of the collaborative processing solution in this embodiment includes the following steps:

[0097] S510: Convert the received AI / ML reasoning requirement to generate AI / ML reasoning requirement information.

[0098] In an embodiment, an AI / ML inference management service consumer receives an AI / ML inference demand and generates AI / ML inference demand information.

[0099] S520: Send an intent negotiation request to request a report of an AI / ML collaborative reasoning solution for a specific strategy.

[0100] Specific strategies may include, but are not limited to, the highest training accuracy strategy and the best demand matching strategy. In an embodiment, the AI / ML reasoning management service consumer sends an intent negotiation request to the AI / ML reasoning management service producer. The intent negotiation request includes a target intent and an intent negotiation operation. The target intent may be AI / ML reasoning requirement information, and the intent negotiation operation is a specific strategy intent negotiation. The request requires the AI / ML reasoning management service producer to provide feedback on a feasible solution corresponding to the specific strategy. The AI / ML reasoning management service consumer sends the specific strategy intent negotiation and AI / ML reasoning requirement information to the AI / ML reasoning management service producer.

[0101] S530: Query entity capability information corresponding to the AI / ML entity.

[0102] In an embodiment, the AI / ML inference management service producer asks the AI / ML entity for entity capability information of its AI / ML entity.

[0103] S540: Receive entity capability information reported by the AI / ML entity.

[0104] In an embodiment, the AI / ML entity reports entity capability information of the AI / ML entity to the AI / ML inference management service producer.

[0105] S550. Determine a set of candidate AI / ML collaborative reasoning solutions based on the solution orchestration strategy determined by the intent negotiation request, entity capability information, and current computing resource conditions.

[0106] In one embodiment, an AI / ML inference management service producer determines a set of potentially feasible candidate AI / ML collaborative inference solutions based on AI / ML inference requirements, AI / ML entity capabilities, and current computing resources. The AI / ML inference management service producer orchestrates the AI / ML entities to produce different solutions, reserves relevant resources, and sets the validity period of the solutions.

[0107] S560: Report a set of candidate AI / ML collaborative reasoning solutions and solution validity period information.

[0108] In an embodiment, the AI / ML reasoning management service producer responds to the intent negotiation and reports a set of candidate AI / ML collaborative reasoning solutions and solution validity period information of each candidate AI / ML collaborative reasoning solution to the AI / ML reasoning management service consumer.

[0109] S570: Start the solution validity period timer and wait for the intention negotiation reply message.

[0110] In an embodiment, the AI / ML reasoning management service producer starts a solution validity timer and waits for an intent negotiation reply message from the AI / ML reasoning management service consumer.

[0111] S580: When the plan validity period counter expires, if the intention negotiation reply message is still not received, it is determined that the intention execution has failed.

[0112] In an embodiment, when the solution validity timer expires and the AI / ML reasoning management service producer does not receive the intent negotiation reply message, it can be determined that the intent execution has failed.

[0113] S590: Send intention negotiation result information.

[0114] In an embodiment, the AI / ML reasoning management service producer sends intent negotiation result information to the AI / ML reasoning management service consumer, wherein the intent negotiation result information includes: negotiation result, negotiation failure status information and negotiation failure reason; wherein the negotiation result is negotiation failure; the negotiation failure status is intent execution failure; the negotiation failure reason is that the intent negotiation reply message was not received within the validity period.

[0115] In one embodiment, Figure 6 is an interactive diagram of another solution arrangement provided by an embodiment of the present application. Taking the first management device as the AI / ML reasoning management service consumer, the second management device as the AI / ML reasoning management service producer, the entity used is the AI / ML entity (including AI / ML entity 1, AI / ML entity 2...AI / ML entity N), the collaborative processing solution is the AI / ML collaborative reasoning solution, the business demand information is the AI / ML reasoning demand, and the target intent is the AI / ML reasoning demand information as an example, the arrangement process of the collaborative processing solution is described. As shown in Figure 6, the arrangement process of the collaborative processing solution in this embodiment includes the following steps:

[0116] S610: Convert the received AI / ML reasoning requirement to generate AI / ML reasoning requirement information.

[0117] In an embodiment, an AI / ML inference management service consumer receives an AI / ML inference demand and generates AI / ML inference demand information.

[0118] S620: Send an intent negotiation request to request a report of an AI / ML collaborative reasoning solution for a specific strategy.

[0119] Specific strategies may include, but are not limited to, the highest training accuracy strategy and the best demand matching strategy. In an embodiment, the AI / ML reasoning management service consumer sends an intent negotiation request to the AI / ML reasoning management service producer. The intent negotiation request includes a target intent and an intent negotiation operation. The target intent may be AI / ML reasoning requirement information, and the intent negotiation operation is a specific strategy intent negotiation. The request requires the AI / ML reasoning management service producer to provide feedback on a feasible solution corresponding to the specific strategy. The AI / ML reasoning management service consumer sends the specific strategy intent negotiation and AI / ML reasoning requirement information to the AI / ML reasoning management service producer.

[0120] S630: Query entity capability information corresponding to the AI / ML entity.

[0121] In an embodiment, the AI / ML inference management service producer asks the AI / ML entity for entity capability information of its AI / ML entity.

[0122] S640: Receive entity capability information reported by the AI / ML entity.

[0123] In an embodiment, the AI / ML entity reports entity capability information of the AI / ML entity to the AI / ML inference management service producer.

[0124] S650. Determine a set of candidate AI / ML collaborative reasoning solutions based on the solution orchestration strategy determined by the intent negotiation request, entity capability information, and current computing resource conditions.

[0125] In one embodiment, an AI / ML inference management service producer identifies potentially feasible candidate AI / ML collaborative inference solutions based on AI / ML inference requirements, AI / ML entity capabilities, and current computing resources. The AI / ML inference management service producer orchestrates the AI / ML entities to produce different solutions, reserves relevant resources, and sets the validity period of the solutions.

[0126] S660: Report the candidate AI / ML collaborative reasoning solution set and solution validity period information.

[0127] In an embodiment, the AI / ML reasoning management service producer responds to the intent negotiation and reports a set of candidate AI / ML collaborative reasoning solutions to the AI / ML reasoning management service consumer, as well as solution validity period information for each candidate AI / ML collaborative reasoning solution.

[0128] S670: Start the solution validity period timer and wait for the intention negotiation reply message.

[0129] In an embodiment, the AI / ML reasoning management service producer starts a solution validity timer and waits for an intent negotiation reply message from the AI / ML reasoning management service consumer.

[0130] S680: When the plan validity period counter expires, if no intention negotiation reply message is received, the intention is suspended.

[0131] In this embodiment, if the proposal validity timer expires and the AI / ML reasoning management service producer does not receive a response from the AI / ML reasoning management service consumer, the intent is suspended (execution is temporarily suspended) and an intent suspension timer is set. If no intent negotiation response is received before the timer expires, the intent is deemed to have failed.

[0132] S690: Send intention negotiation result information.

[0133] In this embodiment, the AI / ML inference management service producer sends intent negotiation result information to the AI / ML inference management service consumer. The intent negotiation result information includes: negotiation result, negotiation failure status information, and negotiation failure reason. The negotiation result is negotiation failure; the negotiation failure status is intent pending; and the negotiation failure reason is failure to receive an intent negotiation response within the validity period. The intent pending timer is started to wait for a response from the AI / ML inference management service consumer.

[0134] S6100. Send recovery intention indication information.

[0135] In an embodiment, the AI / ML reasoning management service consumer sends an intent resumption indication message to the AI / ML reasoning management service producer to request to resume execution of the intent.

[0136] S6110. Rearrange the intention recovery indication information, current computing resource status and entity capability information to generate a new set of candidate AI / ML collaborative processing solutions.

[0137] In an embodiment, the AI / ML reasoning management service producer re-arranges and generates a new set of candidate AI / ML collaborative reasoning solutions based on the current computing resource situation and entity capability information.

[0138] S6120. Send a new set of candidate AI / ML collaborative processing solutions.

[0139] In an embodiment, the AI / ML reasoning management service producer initiates intent negotiation again, and reports a new set of candidate AI / ML collaborative reasoning solutions and solution validity period information of each new candidate AI / ML collaborative reasoning solution to the AI / ML reasoning management service consumer.

[0140] S6130: Start the solution validity period timer and wait for the intention negotiation reply message.

[0141] In an embodiment, the AI / ML reasoning management service producer starts a solution validity timer and waits for the AI / ML reasoning management service consumer to reply to the negotiation intention.

[0142] S6140. Send at least one new target AI / ML collaborative reasoning solution.

[0143] In an embodiment, the AI / ML reasoning management service consumer selects at least one new candidate AI / ML collaborative reasoning solution from the new candidate AI / ML collaborative reasoning solution set as the corresponding new target AI / ML collaborative reasoning solution.

[0144] S6150: Execute a new candidate AI / ML collaborative reasoning solution.

[0145] In an embodiment, the AI / ML inference management service producer executes the selected new candidate AI / ML collaborative inference solution and releases resources not involved in the solution.

[0146] S6160: Send the intention negotiation result information.

[0147] In an embodiment, the AI / ML reasoning management service producer sends intent negotiation result information to the AI / ML reasoning management service consumer, wherein the intent negotiation result information includes: negotiation result, negotiation failure status information and negotiation failure reason; wherein the negotiation result is successful negotiation; the negotiation success status information is that the AI / ML model collaborative reasoning solution is being executed.

[0148] In one embodiment, FIG7 is an interactive diagram of another solution arrangement provided by an embodiment of the present application. Taking the first management device as the ML training management service consumer, the second management device as the ML training management service producer, the entity used as the ML entity (including ML entity 1, ML entity 2...ML entity N), the collaborative processing solution as the ML collaborative training solution, the business requirement information as the ML training requirement, and the target intent as the ML training requirement information as an example, the arrangement process of the collaborative processing solution is described. As shown in FIG7, the arrangement process of the collaborative processing solution in this embodiment includes the following steps:

[0149] S710: Convert the received ML training requirement to generate ML training requirement information.

[0150] In an embodiment, an ML training management service consumer receives an ML model training requirement and generates ML model training requirement information.

[0151] S720: Send an intent negotiation request, requesting to report a candidate ML collaborative training solution with the shortest training time.

[0152] In an embodiment, an ML training management service consumer sends an intent negotiation request, which includes a target intent. The target intent includes an intent negotiation category, which is optimal parameter negotiation, i.e., the shortest training time. It can be understood that the target intent may include ML training requirement information and the shortest training time, and requires a candidate ML collaborative training solution with the shortest training time. The ML training management service consumer sends the ML training requirement information and the shortest training time to the ML training management service producer.

[0153] S730: Query the ML entity for corresponding entity capability information.

[0154] In an embodiment, the ML training management service producer asks each ML entity for entity capability information of its ML entity.

[0155] S740: Receive entity capability information reported by the ML entity.

[0156] In an embodiment, each ML entity reports entity capability information of the ML entity to the ML training management service producer.

[0157] S750: Determine a set of candidate ML collaborative training solutions based on the solution scheduling strategy determined by the intent negotiation request, entity capability information, and current computing resource conditions.

[0158] In one embodiment, the ML training management service producer determines a set of feasible candidate ML collaborative training solutions based on ML model training requirements, ML entity capability information, and current computing resources. The producer evaluates the training time of each candidate ML collaborative training solution and determines the candidate ML collaborative training solution with the shortest training time. The producer also reserves relevant resources and sets a validity period for the value, i.e., the shortest training time validity period.

[0159] S760: Report the candidate ML collaborative training solution set and solution validity period information.

[0160] In an embodiment, the ML training management service producer reports the shortest training time and its corresponding candidate ML collaborative training solution to the ML training management service consumer, including solution validity period information.

[0161] The management service producer supports feedback of the best achievable value of a parameter (such as a key performance indicator (KPI)) based on the request of the management service consumer. In this example, the ML training management service producer reports the shortest training time.

[0162] S770: Start the solution validity period timer and wait for the intention negotiation reply message.

[0163] In an embodiment, the ML training management service producer starts a solution validity timer and waits for a negotiation intention reply message from the ML training management service consumer.

[0164] S780: Determine and request execution of a candidate ML collaborative training solution with the shortest training time.

[0165] In an embodiment, the ML training management service consumer requests to execute the ML collaborative training plan corresponding to the shortest training time, and uses it as the target ML collaborative training plan.

[0166] S790: Execute the candidate ML collaborative training solution with the shortest training time.

[0167] In an embodiment, the ML training management service producer executes the candidate ML collaborative training solution with the shortest training time.

[0168] S7100: Send intention negotiation result information.

[0169] In an embodiment, the ML training management service producer sends intent negotiation result information to the ML training management service consumer, wherein the intent negotiation result information includes: negotiation result, negotiation failure status information and negotiation failure reason; wherein the negotiation result is successful negotiation; the negotiation success status information is that the ML model collaborative training plan is being executed.

[0170] In one embodiment, FIG8 is an interactive diagram of another solution arrangement provided by an embodiment of the present application. Taking the first management device as the ML training management service consumer, the second management device as the ML training management service producer, the entity used as the ML entity (including ML entity 1, ML entity 2...ML entity N), the collaborative processing solution as the ML collaborative training solution, the business demand information as the ML training demand, and the target intent as the ML training demand information as an example, the arrangement process of the collaborative processing solution is described. As shown in FIG8, the arrangement process of the collaborative processing solution in this embodiment includes the following steps:

[0171] S810: Convert the received ML training requirement to generate ML training requirement information.

[0172] In an embodiment, an ML training management service consumer receives an ML model training requirement and generates ML model training requirement information.

[0173] S820: Send an intent negotiation request, requesting to report a candidate ML collaborative training solution with the shortest training time.

[0174] In an embodiment, an ML training management service consumer sends an intent negotiation request, which includes a target intent. The target intent includes an intent negotiation category, which is optimal parameter negotiation, i.e., shortest training time. It can be understood that the target intent may include ML training requirement information and the shortest training time, and requests the report of a candidate ML collaborative training solution with the shortest training time. The ML training management service consumer sends the ML training requirement information and the shortest training time to the ML training management service producer.

[0175] S830: Query the ML entity for corresponding entity capability information.

[0176] In an embodiment, the ML training management service producer asks each ML entity for entity capability information of its ML entity.

[0177] S840: Receive entity capability information reported by the ML entity.

[0178] In an embodiment, each ML entity reports entity capability information of the ML entity to the ML training management service producer.

[0179] S850: Determine a set of candidate ML collaborative training solutions based on the solution scheduling strategy determined by the intent negotiation request, entity capability information, and current computing resource conditions.

[0180] In this embodiment, the ML training management service producer determines feasible ML model collaborative training solutions based on ML model training requirements, ML entity capability information, and current resource availability. The producer evaluates the training time of each solution, determines the shortest training time solution, and reserves relevant resources and sets a numerical validity period, namely, the shortest training time validity period.

[0181] S860: Report the candidate ML collaborative training solution set and solution validity period information.

[0182] In an embodiment, the ML training management service producer reports the shortest model training time and the corresponding training plan to the ML training management service consumer, including information on the validity period of the value.

[0183] S870: Start the solution validity period timer and wait for the intention negotiation reply message.

[0184] In an embodiment, the ML training management service producer starts a value validity timer and waits for a response to the intention negotiation from the ML training management service consumer.

[0185] S880: If no intention negotiation reply message is received within the validity period of the plan, the target ML collaborative training plan is automatically executed.

[0186] In an embodiment, if no intent negotiation response is received within the validity period of the value, the ML training management service producer automatically executes the ML model collaborative training plan corresponding to the shortest training time.

[0187] S890: Send the intention negotiation result information.

[0188] In an embodiment, the ML training management service producer sends intent negotiation result information to the AI / ML reasoning management service consumer, wherein the intent negotiation result information includes: negotiation result, negotiation failure status information and negotiation failure reason; wherein the negotiation result is negotiation failure; the negotiation failure status information is automatic execution of the ML model collaborative training plan; the negotiation failure reason is that no intent negotiation reply is received within the validity period.

[0189] In one embodiment, Figure 9 is a block diagram of a solution orchestration apparatus provided in an embodiment of the present application. This embodiment is applied to a first management device. As shown in Figure 9, the solution orchestration apparatus in this embodiment includes: a generation module 910, a sending negotiation module 920, and an intention negotiation module 930.

[0190] The generating module 910 is configured to generate a target intent according to the received business requirement information.

[0191] The sending negotiation module 920 is configured to send an intent negotiation request to the second management device based on the target intent, requesting the return of a set of candidate collaborative processing solutions.

[0192] The intention negotiation module 930 is configured to perform intention negotiation with the second management device based on the candidate collaborative processing solution set to determine a target collaborative processing solution set.

[0193] In one embodiment, the generation module 910 is configured to convert the received business requirement information to generate a corresponding target intent; wherein the business requirement information at least includes AI / ML requirements.

[0194] In one embodiment, the category of the AI / ML requirements of the target intent includes at least one of the following: reasoning requirements; training requirements.

[0195] In one embodiment, the target intent includes at least one of the following parameters: reasoning category; model collaboration strategy; model version; expected output result; input data; used entity; model performance indicator;

[0196] Among them, the model performance indicators include at least one of the following parameters: accuracy; precision; recall rate; precision-recall rate PR curve; F1 score.

[0197] In one embodiment, the target intent includes: an intent negotiation category; the intent negotiation category is used to indicate a category of an intent negotiation request.

[0198] In one embodiment, the intended negotiation category includes at least one of the following: optimal parameter negotiation, specific policy negotiation.

[0199] In one embodiment, when the intent negotiation category is optimal parameter negotiation, the candidate collaborative processing solution set includes at least one of the following: an optimal value obtained by at least one parameter in the target intent; and a corresponding candidate collaborative processing solution.

[0200] In one embodiment, when the intended negotiation category is specific policy negotiation, the candidate collaborative processing solution set includes at least one candidate collaborative processing solution generated based on the specific policy.

[0201] In one embodiment, the intention negotiation request includes: an intention negotiation operation; and a sending negotiation module 920 configured to initiate the intention negotiation operation to the second management device based on the target intention, and request to return a set of candidate collaborative processing solutions.

[0202] In one embodiment, the intent negotiation operation includes at least one of the following: optimal parameter intent negotiation; specific policy intent negotiation.

[0203] In one embodiment, when the optimal parameter intent negotiation is initiated, the candidate collaborative processing solution set includes at least one of the following: an optimal value obtained by at least one parameter in the target intent; and a corresponding candidate collaborative processing solution.

[0204] In one embodiment, when negotiation of a specific policy intention is initiated, the candidate co-processing solution set includes at least one candidate co-processing solution generated based on the specific policy.

[0205] In one embodiment, the plan arrangement apparatus applied to the first management device further includes:

[0206] The receiver is configured to receive solution validity period information of each candidate co-processing solution in the candidate co-processing solution set.

[0207] In one embodiment, the solution validity period information corresponding to each candidate collaborative processing solution is unique; or, the solution validity period information corresponding to each candidate collaborative processing solution is the same.

[0208] In one embodiment, the plan arrangement apparatus applied to the first management device further includes:

[0209] The transmitter is configured to send an intention negotiation reply message to the second management device within the validity period of the solution corresponding to the solution validity period information; wherein the intention negotiation reply message at least includes a determined target collaborative processing solution set; the target collaborative processing solution set is at least one candidate collaborative processing solution in the candidate collaborative processing solution set.

[0210] In one embodiment, the plan arrangement apparatus applied to the first management device further includes:

[0211] The receiver is further configured to receive the intention negotiation result information returned by the second management device.

[0212] In one embodiment, the intended negotiation result information includes at least one of the following: negotiation result; negotiation success status information; negotiation failure status information and negotiation failure reason;

[0213] The negotiation failure status information includes at least one of the following: automatic execution plan; intention suspension; intention failure.

[0214] In one embodiment, when the negotiation failure status information is an intention to suspend, the solution arrangement apparatus applied to the first management device further includes:

[0215] The transmitter is further configured to send restoration intention indication information to the second management device;

[0216] The receiver is further configured to receive a new set of candidate collaborative processing solutions rearranged by the second management device based on the intention to restore indication information, the current computing resource situation, and the entity capability information;

[0217] The determining module is configured to determine at least one new target collaborative processing solution set from the new candidate collaborative processing solution set.

[0218] The program arrangement device provided in this embodiment is configured to implement the program arrangement method applied to the first management device in the embodiment shown in FIG3 . The implementation principle and technical effects of the program arrangement device provided in this embodiment are similar and will not be described in detail here.

[0219] In one embodiment, FIG10 is a block diagram of another solution arrangement apparatus provided by an embodiment of the present application. This embodiment is applied to a second management device. As shown in FIG10 , the solution arrangement apparatus in this embodiment includes: a receiver 1010, a determination module 1020, and a transmitter 1030.

[0220] The receiver 1010 is configured to receive the intention negotiation request sent by the first management device.

[0221] The determination module 1020 is configured to determine a corresponding candidate collaborative processing solution set at least according to the intended negotiation request.

[0222] The transmitter 1030 is configured to send the candidate collaborative processing solution set to the first management device, so that the first management device performs intent negotiation with the second management device based on the candidate collaborative processing solution set to determine a target collaborative processing solution set.

[0223] In one embodiment, the plan arrangement apparatus applied to the second management device further includes:

[0224] The transmitter 1030 is further configured to send corresponding intention negotiation result information to the first management device.

[0225] In one embodiment, the plan arrangement apparatus applied to the second management device further includes:

[0226] The acquisition module is configured to acquire entity capability information of each entity.

[0227] In one embodiment, the determination module 1020 includes:

[0228] A first determining unit is configured to determine a corresponding solution arrangement strategy according to a target intention in the intention negotiation request;

[0229] The second determining unit is configured to determine a corresponding set of candidate collaborative processing solutions based on the solution scheduling strategy, entity capability information, and current computing resource conditions.

[0230] In one embodiment, the determination module 1020 includes:

[0231] A third determining unit is configured to determine a corresponding solution arrangement strategy according to the target intent and the intent negotiation operation in the intent negotiation request;

[0232] The fourth determining unit is configured to determine a corresponding set of candidate collaborative processing solutions based on the solution scheduling strategy, entity capability information, and current computing resource conditions.

[0233] In one embodiment, the entity capability information includes at least one of the following: entity identification; entity version; entity reasoning category; entity training context; entity performance indicator.

[0234] In one embodiment, the plan arrangement apparatus applied to the second management device further includes:

[0235] The transmitter 1030 is further configured to send solution validity period information of each candidate collaborative processing solution in the candidate collaborative processing solution set to the first management device.

[0236] In one embodiment, the plan arrangement apparatus applied to the second management device further includes:

[0237] The receiver 1010 is further configured to receive, within the validity period of the solution corresponding to the solution validity period information, a negotiation intention reply message sent by the first management device; wherein the negotiation intention reply message includes at least a determined target collaborative processing solution set; the target collaborative processing solution set is at least one candidate collaborative processing solution in the candidate collaborative processing solution set;

[0238] The executor is configured to execute at least one target collaborative processing solution in the target collaborative processing solution set.

[0239] In one embodiment, the solution arrangement apparatus applied to the second management device further includes: a release module configured to release computing resources unrelated to the target collaborative processing solution.

[0240] In one embodiment, when no negotiation intention reply message sent by the first management device is received within the plan validity period corresponding to the plan validity period information, the plan arrangement apparatus applied to the second management device further includes:

[0241] The generation module is configured to generate corresponding negotiation failure status information according to a pre-configured solution arrangement strategy; wherein the negotiation failure status information includes at least one of the following: automatic execution of the solution; intention suspension; intention failure.

[0242] In one embodiment, when the negotiation failure status information indicates an intention to suspend, the solution arrangement apparatus applied to the second management device further includes:

[0243] The receiver 1010 is further configured to receive the restoration intention indication information sent by the first management device;

[0244] The generation module is further configured to re-arrange and generate a new set of candidate collaborative processing solutions based on the intention recovery indication information, the current computing resource status and the entity capability information;

[0245] The transmitter 1030 is further configured to send a new candidate collaborative processing solution set to the first management device.

[0246] The program arrangement device provided in this embodiment is configured to implement the program arrangement method applied to the second management device in the embodiment shown in Figure 4. The implementation principle and technical effects of the program arrangement device provided in this embodiment are similar and will not be repeated here.

[0247] In one embodiment, Figure 11 is a schematic diagram of the structure of a communication device provided by an embodiment of the present application. As shown in Figure 11, the device provided by the present application includes: a processor 1110, a memory 1120, and a communication module 1130. The number of processors 1110 in the device can be one or more, and Figure 11 takes one processor 1110 as an example. The number of memories 1120 in the device can be one or more, and Figure 11 takes one memory 1120 as an example. The processor 1110, memory 1120, and communication module 1130 of the device can be connected via a bus or other means, and Figure 11 takes connection via a bus as an example. In this embodiment, the device can be a first management device or a second management device.

[0248] Memory 1120, as a computer-readable storage medium, can be configured to store software programs, computer-executable programs, and modules, such as program instructions / modules corresponding to the device of any embodiment of the present application (for example, the generation module 910, the sending negotiation module 920, and the intention negotiation module 930 in the solution arrangement device applied to the first management device). Memory 1120 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created based on the use of the device, etc. In addition, memory 1120 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state memory device. In some examples, memory 1120 may further include memory remotely located relative to processor 1110, and these remote memories may be connected to the device via a network. Examples of the above-mentioned networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0249] In the case where the communication device is the first management device, the device provided above can be configured to execute the solution arrangement method applied to the first management device provided in any of the above embodiments, and have corresponding functions and effects.

[0250] In the case where the communication device is a second management device, the device provided above can be configured to execute the solution arrangement method applied to the second management device provided in any of the above embodiments, and have corresponding functions and effects.

[0251] An embodiment of the present application also provides a storage medium containing computer-executable instructions, which, when executed by a computer processor, are used to execute a solution orchestration method applied to a first management device, the method comprising: generating a target intent based on received business demand information; sending an intent negotiation request to a second management device based on the target intent, requesting the return of a set of candidate collaborative processing solutions; and conducting intent negotiation with the second management device based on the set of candidate collaborative processing solutions to determine a target collaborative processing solution set.

[0252] An embodiment of the present application also provides a storage medium containing computer-executable instructions, which, when executed by a computer processor, are used to execute a solution orchestration method applied to a second management device, the method comprising: receiving an intent negotiation request sent by a first management device; determining a corresponding set of candidate collaborative processing solutions at least based on the intent negotiation request; and sending the set of candidate collaborative processing solutions to the first management device, so that the first management device conducts intent negotiation with the second management device based on the set of candidate collaborative processing solutions to determine a target collaborative processing solution set.

[0253] It will be appreciated by those skilled in the art that the term user equipment encompasses any suitable type of wireless user equipment, such as a mobile phone, a portable data processing device, a portable web browser or a car-mounted mobile station.

[0254] In general, various embodiments of the present application may be implemented in hardware or dedicated circuits, software, logic, or any combination thereof. For example, some aspects may be implemented in hardware, while other aspects may be implemented in firmware or software that can be executed by a controller, microprocessor, or other computing device, although the present application is not limited thereto.

[0255] Embodiments of the present application may be implemented by executing computer program instructions by a data processor of a mobile device, for example, in a processor entity, or by hardware, or by a combination of software and hardware. The computer program instructions may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages.

[0256] The block diagram of any logic flow in the drawings of the present application may represent program steps, or may represent interconnected logic circuits, modules and functions, or may represent a combination of program steps and logic circuits, modules and functions. A computer program may be stored on a memory. The memory may be of any type suitable for the local technical environment and may be implemented using any suitable data storage technology, such as, but not limited to, read-only memory (ROM), random access memory (RAM), optical storage devices and systems (digital versatile discs (DVD) or compact disks (CD)), etc. Computer-readable media may include non-transient storage media. A data processor may be of any type suitable for the local technical environment, such as, but not limited to, a general-purpose computer, a special-purpose computer, a microprocessor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), and a processor based on a multi-core processor architecture.

[0257] The above are merely optional embodiments of the present application and are not intended to limit the present application. Various modifications and variations are possible for those skilled in the art. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included within the scope of protection of the present application.

Claims

1. A solution arrangement method, applied to a first management device, comprising: Generate target intent based on received business demand information; Sending an intent negotiation request to the second management device based on the target intent, requesting the return of a set of candidate collaborative processing solutions; An intent negotiation is performed with the second management device based on the candidate collaborative processing solution set to determine a target collaborative processing solution set.

2. The method according to claim 1, wherein Generating a target intent according to the received business requirement information includes: The received business requirement information is converted to generate corresponding target intent; wherein the business requirement information at least includes artificial intelligence / machine learning AI / ML requirements.

3. The method according to claim 2, wherein: The categories of AI / ML requirements of the target intent include at least one of the following: reasoning requirements; training requirements.

4. The method according to claim 1, wherein The target intent includes at least one of the following parameters: reasoning category; model collaboration strategy; model version; expected output result; input data; used entity; model performance indicator; The model performance index includes at least one of the following parameters: accuracy; precision; recall; precision-recall PR curve; F1 score.

5. The method according to claim 1, wherein The target intent includes: an intent negotiation category; the intent negotiation category is used to indicate the category of the intent negotiation request.

6. The method according to claim 5, wherein: The intended negotiation category includes at least one of the following: optimal parameter negotiation and specific strategy negotiation.

7. The method according to claim 6, wherein: In a case where the intent negotiation category is optimal parameter negotiation, the candidate collaborative processing solution set includes at least one of the following: an optimal value obtained by at least one parameter in the target intent; and a corresponding candidate collaborative processing solution.

8. The method according to claim 6, wherein: In a case where the intended negotiation category is specific policy negotiation, the candidate collaborative processing solution set includes at least one candidate collaborative processing solution generated based on a specific policy.

9. The method according to claim 1, wherein: The intention negotiation request includes: an intention negotiation operation; sending the intention negotiation request to the second management device based on the target intention, requesting to return a set of candidate collaborative processing solutions, including: An intent negotiation operation is initiated with the second management device based on the target intent, requesting the return of a set of candidate collaborative processing solutions.

10. The method according to claim 9, wherein: The intention negotiation operation includes at least one of the following: optimal parameter intention negotiation; specific strategy intention negotiation.

11. The method according to claim 10, wherein: In the case of starting the optimal parameter intention negotiation, the candidate collaborative processing solution set includes at least one of the following: an optimal value obtained by at least one parameter in the target intention; and a corresponding candidate collaborative processing solution.

12. The method according to claim 10, wherein: In the case of starting negotiation of a specific policy intention, the candidate co-processing solution set includes at least one candidate co-processing solution generated based on a specific policy.

13. The method according to claim 1, further comprising: Receive solution validity period information of each candidate collaborative processing solution in the candidate collaborative processing solution set.

14. The method according to claim 13, wherein The solution validity period information corresponding to each candidate collaborative processing solution is unique; or, the solution validity period information corresponding to each candidate collaborative processing solution is the same.

15. The method according to claim 13, further comprising: During the validity period of the scheme corresponding to the scheme validity period information, an intention negotiation reply message is sent to the second management device; wherein the intention negotiation reply message at least includes a determined target collaborative processing scheme set; the target collaborative processing scheme set is at least one candidate collaborative processing scheme in the candidate collaborative processing scheme set.

16. The method according to any one of claims 1 to 15, further comprising: Receive the intention negotiation result information returned by the second management device.

17. The method according to claim 16, wherein: The intended negotiation result information includes at least one of the following: negotiation result; negotiation success status information; negotiation failure status information and negotiation failure reason; The negotiation failure status information includes at least one of the following: automatic execution of the plan; intention suspension; intention failure.

18. The method according to claim 17, wherein: In a case where the negotiation failure status information is an intention to suspend, the method further includes: Sending restoration intention indication information to the second management device; receiving a new set of candidate collaborative processing solutions rearranged by the second management device based on the restoration intention indication information, the current computing resource situation, and the entity capability information; At least one new target co-processing solution is determined from the new candidate co-processing solution set.

19. A solution arrangement method, applied to a second management device, comprising: receiving an intention negotiation request sent by the first management device; Determining a corresponding set of candidate collaborative processing solutions at least according to the intended negotiation request; The candidate collaborative processing solution set is sent to the first management device, so that the first management device performs intent negotiation with the second management device based on the candidate collaborative processing solution set to determine a target collaborative processing solution set.

20. The method according to claim 19, further comprising: Send corresponding intention negotiation result information to the first management device.

21. The method of claim 19, further comprising: Get entity capability information for each entity.

22. The method according to claim 21, wherein The determining a corresponding candidate collaborative processing solution set at least according to the intended negotiation request includes: Determine a corresponding solution arrangement strategy according to the target intention in the intention negotiation request; A corresponding set of candidate collaborative processing solutions is determined according to the solution scheduling strategy, the entity capability information and the current computing resource situation.

23. The method according to claim 21, wherein The determining a corresponding candidate collaborative processing solution set at least according to the intended negotiation request includes: Determine a corresponding solution arrangement strategy according to the target intent and the intent negotiation operation in the intent negotiation request; A corresponding set of candidate collaborative processing solutions is determined according to the solution scheduling strategy, the entity capability information and the current computing resource situation.

24. The method according to claim 21, wherein The entity capability information includes at least one of the following: entity identification; entity version; entity reasoning category; entity training context; entity performance indicator.

25. The method of claim 19, further comprising: Sending solution validity period information of each candidate collaborative processing solution in the candidate collaborative processing solution set to the first management device.

26. The method according to claim 25, further comprising: receiving, within the validity period of the solution corresponding to the solution validity period information, an intention negotiation reply message sent by the first management device; wherein the intention negotiation reply message at least includes a determined target collaborative processing solution set; the target collaborative processing solution set is at least one candidate collaborative processing solution in the candidate collaborative processing solution set; At least one target co-processing solution in the target co-processing solution set is executed.

27. The method according to claim 26, further comprising: Release computing resources that are not relevant to the target collaborative processing solution.

28. The method according to claim 25, wherein In a case where no negotiation intention reply message sent by the first management device is received within the validity period of the solution corresponding to the solution validity period information, the method further includes: Generate corresponding negotiation failure status information according to the pre-configured solution arrangement strategy; wherein the negotiation failure status information includes at least one of the following: automatic execution of the solution; intention suspension; intention failure.

29. The method according to claim 28, wherein In a case where the negotiation failure status information is an intention to suspend, the method further includes: Receiving restoration intention indication information sent by the first management device; Rearrange and generate a new set of candidate collaborative processing solutions based on the intended recovery indication information, current computing resource conditions, and entity capability information; The new candidate collaborative processing solution set is sent to the first management device.

30. A management device comprising: memory, and one or more processors; The memory is configured to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 18 or 19 to 29.

31. A storage medium storing a computer program, wherein when the computer program is executed by a processor, the method according to any one of claims 1 to 18 or 19 to 29 is implemented.

Images