An edge service autonomous deployment method for a service-intensive internet of things scene

Abstract

The application belongs to the technical field of edge computing, and discloses a kind of edge service autonomous deployment method for service-intensive Internet of Things scene, comprising: abstracting Internet of Things service as process component, planning data flow and service scheduling in network using various process modeling methods, and mathematically analyzing process model;Using service network model and distributed microservice data communication model to automatically monitor network data;According to the mobility of entity service node, set up appropriate service deployment algorithm to get the optimal service deployment scheme;Automatically deploy various microservices in code repository to Internet of Things service network.The application starts from process graphical modeling, autonomously completes process programming, process feasibility analysis, service network data and topology monitoring, deployment scheme calculation, and automatically deploys the whole life cycle content of running, combines the theoretical innovation of deployment algorithm with the realization of practical engineering scene application.

Description

A method for autonomous deployment of edge services in service-intensive IoT scenarios Technical Field

[0001] This invention belongs to the field of edge computing technology, and in particular relates to a method for autonomous deployment of edge services for service-intensive Internet of Things (IoT) scenarios. Background Technology

[0002] Currently, the Internet of Things (IoT) relies on mature and diverse network technologies to connect a vast and diverse network of sensors, enabling widespread interconnection between things and between things and people. Traditional IoT consists of a data center and numerous device-side service nodes. The data center serves as the storage and control center for the entire IoT system, while the device-side service nodes connect devices of various protocol types to the network, providing data for the entire network. However, as the IoT system grows increasingly large, the traditional IoT structure is revealing numerous problems: the ever-expanding network structure and number of nodes lead to a surge in data volume and service demands. The traditional network structure's practice of placing all data and services in a remote data center results in a sharp increase in service time consumption and network load, gradually failing to meet the needs of real-world scenarios.

[0003] For example, in scenarios such as fire alarms and medical alarms, the time consumed by traditional network frameworks for data transmission and service waiting is clearly unacceptable. However, in applications such as industrial sensor monitoring, most of the data generated is immediately used in short-range machine-to-machine communication, eliminating the need for time-consuming and laborious transmission and storage in data centers. Fully utilizing edge computing nodes and conducting reasonable service system planning, design, and management are crucial for the further development and promotion of IoT technology.

[0004] Therefore, the Internet of Things (IoT) system has gradually broken free from the traditional framework of direct connection between the edge and the data center, forming a new "edge-cloud" architecture that fully utilizes edge computing resources. The concept of edge-cloud collaboration has also been widely recognized and discussed in academia and industry. Edge-cloud collaborative computing systems have become a relatively mature concept in recent years, mainly involving comprehensive collaboration between edge nodes and data centers at various levels, including hardware, platforms, and services, to meet the needs of data traffic optimization, network load reduction, time consumption reduction, and service load balancing in various business scenarios.

[0005] Existing technologies for service deployment in edge / fog computing largely ignore the complexity of service processes in real-world scenarios such as the Industrial Internet and smart cities. They treat edge services and sensor data sources as independent entities and mostly rely on theoretical scenario assumptions, studying service deployment issues based on predefined scenarios and known data. They fail to truly integrate with engineering solutions, neglecting the complex network scenarios, service discovery issues, inter-service data interaction problems, and the dynamic changes in various basic environmental data encountered in actual engineering projects.

[0006] Based on the above analysis, the problems and defects of the existing technology are as follows: the service deployment scheme of the existing technology differs greatly from the actual scenario, cannot be combined with the engineering scheme, and cannot be matched with the actual data in real time. Summary of the Invention

[0007] To address the problems existing in the prior art, this invention provides a method for autonomous deployment of edge services in service-intensive IoT scenarios.

[0008] This invention is implemented as follows: a method for autonomous deployment of edge services in service-intensive IoT scenarios, comprising:

[0009] First, IoT services are abstracted into process components. Various process modeling methods are used to plan data flow and service scheduling in the network, and mathematical analysis is performed on the process model. A suitable process model coding scheme is selected, and the graphical process modeling is integrated into the automated deployment framework.

[0010] Secondly, it is used to construct a service network model and a data communication model for distributed microservices, and to automatically monitor network data using the service network model and the data communication model for distributed microservices.

[0011] Then, based on the various types of raw IoT data obtained, and according to the mobility of the physical service nodes, an appropriate service deployment algorithm is set to obtain the optimal service deployment scheme.

[0012] Finally, based on the generated server deployment plan, various microservices in the code repository are automatically deployed to the IoT service network.

[0013] Furthermore, the method for autonomous deployment of edge services in service-intensive IoT scenarios includes the following steps:

[0014] Step 1: Model the user's predetermined data flow through a graphical interface according to a standard process model protocol; map the constructed graphical model to a mathematical model, verify the security of the process, and convert the graphical model into readable data.

[0015] Step 2: Construct a service network and data communication model; define the network topology using the service network and data communication model, and determine the service discovery and service data interaction schemes in the network based on the defined network topology; simultaneously monitor the network infrastructure environment under the defined network topology, including the running status of various microservices, load balancing status, server hardware status, and dynamic changes in the network environment.

[0016] Step 3: Set up service deployment algorithms for fixed entity services applicable to various scenarios, and obtain the optimal service deployment scheme based on the original IoT data; use the Monte Carlo simulation method, with the fixed service optimization algorithm as the single sample target, to determine the best entity service movement scheme;

[0017] Step 4: Generate an automated deployment scheme based on Kubernetes. Package all individual microservices in the cloud code repository into Docker Images and store them in the cloud service image repository. Transform the obtained optimal service deployment scheme into XML information carried by the service deployment. Use containerization to automatically deploy IoT services to the edge server cluster according to a reasonable deployment scheme.

[0018] Furthermore, as described in step one, the present invention adopts the Business Process Modeling Notation 2.0 (BPMN 2.0) protocol. The present invention realizes the mutual conversion between models and data through the Activiti process execution engine, and realizes the progression of user requirements, graphical interface, process model and system available data.

[0019] Furthermore, the network topology definition in step two includes:

[0020] The IoT application system adopts an event-driven distributed edge computing architecture, where services are completely decoupled and have no dependencies. Business processes operate through event propagation; when a service receives data and a certain parameter reaches a threshold, it is triggered to perform an operation, generating data that flows to subsequent services. Message queue-based publish / subscribe communication perfectly supports this model. Furthermore, services are stateless; communication between services is solely dependent on the currently transmitted data or events, completely unrelated to historical interactions, and no historical interaction information is retained. This architectural design ensures high service efficiency and stability, and further guarantees the manageability and scalability of the entire service network.

[0021] Abstracting services and data: Abstracting the original service nodes into entity service nodes and event service nodes, and managing the service status and data of each sub-network structure through a multi-level service registry;

[0022] Step two, which involves constructing the service network and data communication model, includes:

[0023] (1) Construct a multi-level service registry based on the ZooKeeper distributed service system, and use the multi-level service registry to perform autonomous service discovery and registration, autonomous message flow and service load balancing.

[0024] (2) Use RocketMQ to build a message queue and perform multi-type message flow within sub-networks, between networks, one-to-one, and one-to-many based on message attributes;

[0025] (3) Based on ZABBIX, monitor the hardware attributes of edge servers in the entire network and describe the network and hardware status in real time.

[0026] Furthermore, the raw IoT data mentioned in step three includes the relevant data and models mentioned in steps 1-2: workflow information, service interaction information, service network topology, and edge server hardware information.

[0027] Step three includes the following specific implementation details:

[0028] 1) Obtain the deployment information of each event service node in the network by using the fixed location service deployment algorithm of the entity service node: take the target entity service node as the entity service node with a fixed location, and use the dominance-based multi-objective optimization heuristic algorithm with the deployment quantity and final network consumption as scheduling objectives to obtain the deployment information of each event service node in the network.

[0029] 2) Obtain a dataset of randomly distributed entity service points, use the fixed-location entity service node deployment algorithm as the objective function for a single sample, calculate the ideal optimization objective for a large number of samples based on the randomly distributed entity service point dataset, and obtain a reasonable deployment scheme for mobile entity service node deployment.

[0030] Another objective of this invention is to provide an edge service autonomous deployment system for service-intensive IoT scenarios, which implements the aforementioned edge service autonomous deployment method for service-intensive IoT scenarios. The edge service autonomous deployment system for service-intensive IoT scenarios includes:

[0031] The process design module is used to abstract IoT services into process components, plan data flow and service scheduling in the network using various process modeling methods, and perform mathematical analysis on the process model; at the same time, it selects and writes appropriate process model coding schemes and integrates process graphical modeling into the automated deployment framework.

[0032] The network environment monitoring module is used to construct a service network model and a data communication model for distributed microservices, and to automatically monitor network data using the service network model and the data communication model for distributed microservices.

[0033] The server deployment scheme generation module is used to obtain the optimal service deployment scheme by setting appropriate service deployment algorithms based on the mobility of physical service nodes according to the various types of raw IoT data obtained.

[0034] The automated deployment module is used to automatically deploy various microservices from the code repository to the IoT service network based on the generated server deployment plan.

[0035] Another object of the present invention is to provide a computer device including a memory and a processor, the memory storing a computer program, which, when executed by the processor, causes the processor to perform the steps of the edge service autonomous deployment method for service-intensive Internet of Things scenarios.

[0036] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the edge service autonomous deployment method for service-intensive Internet of Things scenarios.

[0037] Another objective of this invention is to provide an information data processing terminal, which is used to realize the edge service autonomous deployment system for service-intensive Internet of Things scenarios.

[0038] Based on the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solution to be protected by this invention are as follows:

[0039] First, addressing the technical problems existing in the prior art and the difficulty in solving them, this paper closely analyzes, in conjunction with the technical solution to be protected by this invention and the results and data obtained during the research and development process, how the technical solution of this invention solves the technical problems, and the inventive technical effects brought about by solving these problems. The specific description is as follows:

[0040] This invention constructs a complete business process model solution, which digitizes the relationships between service processes in service-intensive IoT scenarios, whether related or distributed independently, constrains the overall framework of IoT services, and establishes standardized and regulated process management. While ensuring the security of the entire IoT business service process, it completes the overall process from personalized user needs to programmatic solutions.

[0041] This invention proposes a service network model and a distributed microservice data communication model; it fully leverages the interconnectivity of IoT devices, supported by a multi-level service registry, and distributes centralized IoT services across edge servers to handle complex IoT microservice communication processes.

[0042] To address the challenges of large-scale services and complex data flow in service-intensive IoT scenarios, this invention designs a heuristic edge service deployment algorithm based on a service process model and multi-objective optimization. This algorithm improves the service quality of the system under complex service processes and network environments, while ensuring the system's scalability and load balancing.

[0043] This invention addresses the characteristics of service-intensive IoT scenarios, such as dense service requests, short and heterogeneous data, and time sensitivity. Building upon the previous study of a method adapted to mobile service scenarios, this invention further proposes offline deployment of data implementation services based on the analysis of existing mobile data source distribution samples, thereby improving the overall service quality of data source services in mobile scenarios from a system perspective.

[0044] Second, considering the technical solution as a whole or from a product perspective, the technical effects and advantages of the technical solution to be protected by this invention are specifically described as follows:

[0045] This invention starts with graphical process modeling and autonomously completes the entire lifecycle of process lifecycle management, including flowchart programming, process feasibility analysis, monitoring of service network hardware data and topology, deployment scheme calculation, and automatic deployment and operation. It combines theoretical innovation in deployment algorithms with practical engineering application. The system framework designed by this method has significant positive implications for the development and maintenance of related IoT systems, greatly reducing development costs and improving service quality.

[0046] Third, as supplementary evidence of the inventive step of the claims of this invention, it is also reflected in the following important aspects:

[0047] The technical solution of this invention overcomes technical bias:

[0048] Existing research on service deployment issues largely focuses on data-intensive services. This type of research treats each edge application as an independent entity, chain, or DAG, emphasizing how to fully utilize the bandwidth resources of the edge environment to transmit massive data streams. This invention focuses on typical IoT application scenarios, such as smart parks, smart hospitals, and smart factories—service-intensive scenarios. In these scenarios, the individual data points generated by service data sources such as sensors are relatively few, but service data production is frequent and service rules are complex. There are intricate relationships between sensors, between sensors and services, and between services themselves. After data sources generate data, they often produce different types of continuous and widespread impacts in the IoT network based on different thresholds. In this type of service scenario, applications can no longer be considered simple services or service chains, but rather a standardized and collaborative whole—a complete business process. This invention designs a service deployment algorithm tailored to the service and data characteristics of this scenario, improving the system's service quality while ensuring scalability and load balancing under complex service processes and network environments.

[0049] In scenarios where data source services can be mobile, most existing technologies rely on online real-time service deployment based on individual user service demand prediction. This invention addresses the characteristics of service-intensive IoT scenarios, such as dense service requests, short and heterogeneous data, and time sensitivity. Building upon previous research, it proposes a method adapted to mobile service scenarios. By analyzing existing mobile data source distribution sample data, it enables offline service deployment, thereby improving the overall service quality of data source services in mobile scenarios from a system perspective.

[0050] Most existing technologies only make theoretical scenario assumptions, studying service deployment issues based on predefined scenarios and known data. They fail to truly integrate with engineering solutions, neglecting the complex network scenarios, service discovery problems, inter-service data interaction issues, and the dynamic changes of various basic environmental data encountered in real-world engineering projects. This paper fully discusses the implementation scheme for real-time monitoring and acquisition of relevant data in engineering scenarios through service network models and distributed microservice data communication models, and designs specific engineering implementation methods within this system framework for service discovery and interaction methods. Attached Figure Description

[0051] Figure 1 is a flowchart of the autonomous deployment method for edge services in service-intensive IoT scenarios provided by an embodiment of the present invention.

[0052] Figure 2 is a diagram of the edge service autonomous deployment system architecture for service-intensive IoT scenarios provided by an embodiment of the present invention.

[0053] Figure 3 is a Petri net mapping example diagram provided in an embodiment of the present invention;

[0054] Figure 4 is a simplified network topology diagram of the service framework provided in an embodiment of the present invention;

[0055] Figure 5 is a flowchart of the entity service node location fixed service deployment algorithm provided in an embodiment of the present invention;

[0056] Figure 6 is a flowchart of the algorithm for generating the maximum network distance using the workflow sequence provided in an embodiment of the present invention.

[0057] Figure 7 shows the editing interface of the service process model of the system designed in this invention.

[0058] Figure 8 shows the service process model management page of the system designed in this invention.

[0059] Figure 9 shows the edge computing node monitoring interface of the system designed in this invention.

[0060] Figure 10 shows the unified management interface of the system service container designed in this invention.

[0061] Figure 11 shows the unified management interface of the computing nodes of the system designed in this invention.

[0062] Figure 12 shows the interface of the Smart Hospital BIM Management System, the application system of this invention.

[0063] Figure 13 shows the architecture of the smart hospital BIM management system of the application system of this invention.

[0064] Figure 14 is a comparison of the optimization curves of different algorithms under the same sample in the simulation experiment of the deployment algorithm of the present invention.

[0065] Figure 15 shows a comparison of the time consumption of various IoT applications in the application experiment of this invention. Detailed Implementation

[0066] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0067] As shown in Figure 1, the edge service autonomous deployment method for service-intensive IoT scenarios provided in this embodiment of the invention includes the following steps:

[0068] S101, the user's predetermined data flow is modeled through a graphical interface according to a standard process model protocol; the constructed graphical model is mapped to a mathematical model to verify the security of the process, and the scheme of the graphical model is converted into readable matrix data.

[0069] S102, Construct a service network and data communication model; Define the network topology using the service network and data communication model, and determine the service discovery and service data interaction schemes in the network based on the defined network topology; Simultaneously, monitor the network infrastructure environment under the defined network topology, including the running status of various microservices, load balancing status, server hardware status, and dynamic changes in the network environment.

[0070] S103, set up a service deployment algorithm for fixed entity services applicable to various scenarios, obtain the optimal service deployment scheme based on the original IoT data; use the Monte Carlo simulation method, with the fixed service optimization algorithm as the single sample target, to determine the best entity service movement scheme;

[0071] S104 generates an automated deployment scheme based on Kubernetes. It packages all microservices in the cloud code repository into Docker Images and stores them in the cloud service image repository. It transforms the obtained optimal service deployment scheme into XML information carried by the service deployment and uses containerization to automatically deploy IoT services to the edge server cluster according to a reasonable deployment scheme.

[0072] The edge service autonomous deployment method for service-intensive IoT scenarios provided in this embodiment of the invention specifically includes:

[0073] 1) Process modeling, validation, and datafication, specifically including:

[0074] IoT services are abstracted into process components. Various process modeling methods are used to plan data flow and service scheduling in the network, and mathematical analysis is performed on the process models to ensure their feasibility. A suitable process model coding scheme is selected, and the graphical process modeling is integrated into an automated deployment framework to provide inter-service process data for service deployment.

[0075] 2) Service network and data communication model design, and related automatic monitoring of network data, specifically including:

[0076] A service network model and a distributed microservice data communication model are proposed. By giving full play to the device-to-device connectivity of the Internet of Things (IoT), and supported by a multi-level service registry, centralized IoT services are distributed and deployed on edge servers to cope with complex IoT microservice communication processes.

[0077] 3) Service deployment algorithm design and service deployment scheme generation, specifically including:

[0078] Based on service process data and service network model data, a service deployment algorithm for fixed entity services applicable to various scenarios is designed to obtain the optimal service deployment scheme. Then, the Monte Carlo simulation method is used, with the above fixed service optimization algorithm as the single sample objective, to obtain the approximate optimal solution for the final entity service mobility problem.

[0079] Based on the workflow information, service interaction information, service network topology, edge server hardware information, and other raw information provided by the above two modules, the service deployment algorithm module obtains a reasonable automated deployment scheme, namely the deployment information of microservices to edge servers.

[0080] 4) Automated service deployment, specifically including:

[0081] Based on Kubernetes and other technologies, an automated deployment scheme is designed. All individual microservices in the cloud code repository are packaged into Docker Images and stored in the cloud service image repository. The deployment scheme provided by the above service deployment algorithm module is transformed into XML information carried by the service deployment. Relying on containerization technology, the service is automatically deployed to the edge server cluster according to a reasonable deployment scheme.

[0082] As shown in Figure 2, the specific settings of the edge service autonomous deployment system for service-intensive IoT scenarios provided in this embodiment of the invention are as follows:

[0083] <1> Process design module:

[0084] Based on the BPMN (Business Process Modeling Notation) modeling specification, the service process is graphically modeled; various services are abstracted into the following model nodes: Node =<Name,Task,Input Event,OutputEvent> This is to address process modeling. (Name is the node name identifier used to distinguish node type names; Task is the atomic task in the process; InputEven is the input event stream that triggers the atomic task; OutputEven is the output event stream after the task is completed).

[0085] The model is mapped to a Petri net, a mathematical representation of a discrete parallel system. Model nodes (Nodes) are mapped to the flow of Petri net place transitions; events are mapped to independent places. When a task T occurs, the node transitions from its original place to a new place, and a place for an event flow is output from that node. Figure 3 shows an example of Petri net mapping. This method maps the process model to a Petri net, and then the Petri net is analyzed using a reachable graph to verify process safety.

[0086] The goal is to achieve graphical modeling and coded design of service processes, which involves programmatically reading processes that conform to the BPMN specification, extracting key data such as service attributes, data exchange and relationships between services, etc.; based on this data, service deployment strategies are studied, and then automated service deployment is further implemented.

[0087] <2> Network environment monitoring module:

[0088] By adopting a microservices approach and abstracting services and data, this invention utilizes a novel multi-level service registry to design a rational service model. The original service nodes are abstracted into entity service nodes and event service nodes, and the multi-level service registry manages the service status and data of each sub-network structure.

[0089] Each sub-service network module relies on the message queue cluster of the multi-level service registry to conduct message data circulation and remote service invocation; it can also rely on the multi-level service registry to spread events to the entire service network according to the constraints of the event flow.

[0090] The entire service framework adopts a tree-like network topology. The data center manages the entire service network macroscopically through communication with service centers at all levels. This network structure can easily identify the spatial relationship between services by the number of forwards required by the multi-level service registry; the multi-level service registry masks the differences between all sub-service nodes, ensuring communication between services that are spatially distant. Figure 4 shows a simplified network topology of the service framework.

[0091] This system leverages the ZooKeeper distributed service system to design and implement a multi-level service registry, providing macro-level management of the entire network service content. This enables services to autonomously discover and register, messages to flow autonomously, and services to load balance. Utilizing RocketMQ, a message queue is implemented, allowing for various message types—within sub-networks, between networks, one-to-one, and one-to-many—based on message attributes. This achieves service decoupling, data flow peak shaping and limiting, and high system availability. Based on ZABBIX, the system monitors the hardware attributes of edge servers throughout the network. Ultimately, this information is aggregated into a service and data communication model integrated into the coding framework. This provides the service deployment algorithm with raw information such as network structure and edge server attributes, ensuring that services, after automatic deployment, autonomously register, run, and flow messages autonomously.

[0092] <3> Service deployment solution generation module:

[0093] The data prerequisites for further service deployment algorithms include the following aspects:

[0094] ① The network topology, hardware status of each edge server, and distribution of each edge server in the network, provided by the service network and data communication model;

[0095] ② Workflow data in the service network provided by the data results of process modeling, as well as the correlations between various services.

[0096] The service deployment algorithm mainly consists of the following two stages:

[0097] First, the algorithm for deploying services with fixed physical service node locations.

[0098] All target entity service nodes are fixed-location entity service nodes. A dominance-based multi-objective optimization heuristic algorithm is used, with the deployment quantity and final network consumption as scheduling objectives. The goal is to achieve the best system service quality while minimizing the deployment of event service nodes. Finally, the most reasonable scheme is selected from a series of suitable options as the final optimization objective, yielding information on the deployment status of each event service node in the network. The specific algorithm flow is shown in Figure 5.

[0099] Fitness calculation includes three objectives: best QoS, minimum resource consumption, and load balancing. The formula for best QoS is described below:

[0100]

[0101] The server network distance is D, where d is the distance from D to the server network. i,j The network topology distance from the i-th edge server to the j-th edge server is determined by the number of multi-level service center nodes traversed; the service workflow sequence S, where s i The description of the event flow generated by the i-th entity service node is a workflow sequence in the network, consisting of a tree-like structure of event service nodes a; the server load is Q, where q i This represents the load status of the i-th edge server.

[0102] Figure 6 shows the specific algorithm flow for the maximum network distance generated by the workflow sequence described in the above formula.

[0103] Second, the algorithm for deploying mobile services at the location of entity service nodes.

[0104] Based on the above problems, a dataset with randomly distributed entity service locations is obtained. A sample mean algorithm is used to address the randomness issue of mobile services. The aforementioned fixed-location service deployment algorithm for entity service nodes is used as the objective function for a single sample. Based on the randomly distributed entity service location dataset, the ideal optimization objective is calculated for a large number of samples, ultimately yielding a reasonable deployment scheme for mobile services at entity service node locations.

[0105] In this invention, the random distribution of mobile entity services approximates a Poisson distribution, and the independent variables only involve the random distribution of mobile entity services. The optimization variable is the distribution of event service nodes in the edge servers, which is a finite set. Therefore, single-sample extraction can be abstracted into the following formula:

[0106] g(x)=G(w,x)

[0107] In the above formula, w represents a randomly distributed sample of the mobile entity service, x represents the distribution of the target event service in the edge server, and g(x) represents the final optimization objective; the stochastic optimization problem including the expectation can be expressed as:

[0108]

[0109] Where W is a set of Poisson distributions W ~ P(λ) that follows λ, and x is an optimization variable belonging to a finite set x∈S. Finally, the expected function g(x) is obtained.

[0110] Based on the sample mean approximation method, the sample mean of the function g(x) can be defined as:

[0111]

[0112] Where N is the number of samples, the corresponding optimization problem is transformed into:

[0113]

[0114] The algorithm simulates the distribution of mobile entity services by sampling. Based on the distribution status of the mobile entity service distribution, it calculates the sample mean to approximate the target value, ultimately obtaining the approximate optimal solution x in the actual scenario. G(x,w) can be calculated using the fixed entity service deployment algorithm described above.

[0115] <4> Automated deployment module:

[0116] An automated deployment scheme based on Kubernetes is designed to package all individual microservices in the cloud code repository into Docker Images and store them in the cloud service image repository. The deployment scheme provided by the above service deployment algorithm module is transformed into XML information carried by the service deployment. Relying on containerization technology, the service is deployed to the edge server cluster according to a reasonable deployment scheme.

[0117] To demonstrate the inventiveness and technical value of the technical solution of this invention, this section provides specific product or related technology application examples of the technical solution claimed.

[0118] <1> Basic edge service autonomous deployment system for service-intensive IoT scenarios

[0119] This system is based on an edge service autonomous deployment method for service-intensive IoT scenarios, completing the entire lifecycle scheduling from process visualization modeling, dynamic environmental monitoring, deployment scheme generation, to automated deployment and operation. This method can serve as a basic system framework to assist other related types of IoT service systems in rapid development and deployment based on this patented method. The main modules of the method are demonstrated below:

[0120] (1) Process Model Management

[0121] The system platform shown is based on the Activiti engine to implement online BPMN modeling. As shown in Figure 7, users can directly design business process models graphically on the web, reducing the technical gap between users and the system and directly transforming users' abstract and complex service process requirements into data that the system can understand. As shown in Figure 8, the system provides unified management of user process models, offering functions such as process model retrieval, sharing, and exporting to BPMN files. The service process model can be applied directly with one click. Based on Kubernetes, the system automates the deployment of service containers according to the service deployment plan, and can deploy various services to the entire edge computing network based on the service process model data.

[0122] (2) Environmental monitoring

[0123] The system platform, based on ZABBIX, monitors relevant parameters of the entire edge environment and dynamically acquires the hardware status of the entire edge computing environment. As shown in Figure 9, the system provides monitoring of various hardware data for each computing node in the cluster, including disk and memory space data, CPU data, network data, etc., and can also view historical hardware data statistics for different nodes as needed.

[0124] (3) Service Cluster Management

[0125] The system, based on Kubernetes, enables automatic service deployment and monitors and manages the runtime status of related service containers. Figure 10 shows the system's unified management of service container resources and its monitoring of the runtime status of service containers on each node. Figure 11 shows the system's unified management of the node cluster and its monitoring of the status of service containers on each compute node.

[0126] The aforementioned automated deployment system framework for IoT services significantly improves service quality by fully utilizing edge computing node resources. The distributed architecture also enhances system stability, ensuring that most nodes can still provide basic services even when some network nodes or even the cloud service center are offline. The standardization and protocolization of service process models and service containers promotes system standardization, improving overall management and operational efficiency while reducing the workload of development and maintenance personnel. The graphical modeling and management interfaces also greatly reduce the technical costs for management and maintenance personnel. Currently, supported by this system framework, several practical application projects have been implemented, including smart parks, smart buildings, smart hospitals, smart sentinels, and smart roads, fully demonstrating the system's effectiveness.

[0127] <2> Application Examples of Smart Park Related Systems

[0128] The intelligent building operation and maintenance management system uses a 3D model and a rule engine to intelligently manage the IoT devices throughout the building, aiming to improve user experience and reduce overall building resource consumption. Figure 12 shows the system's main interface. The 3D model provides a clear preview of the real-time status of devices in various areas of the building, and allows for remote batch control and management of these devices. The rule engine is a practical engineering application of the system's workflow. When certain parameters of data generated by some devices reach a certain threshold, the system autonomously controls other devices to perform specific operations, further generating corresponding events. This intelligent control system significantly reduces the need for both physical and human resources throughout the building.

[0129] This scenario involves a wide variety of IoT devices, such as lights, air conditioners, curtains, and temperature sensors. The specific service rules, or service process models, are defined by the administrator based on user needs. This diversity of user requirements leads to numerous service types and complex service processes. The system framework described in this paper significantly improves the response efficiency of various IoT application services within the system. Particularly in time-sensitive services such as access control and identity recognition, it greatly enhances the user experience.

[0130] <3> Application Examples of Smart Hospital Related Systems

[0131] The smart hospital BIM management system combines the concepts of "smart service" and "smart management," integrating a data service platform for users with an IoT management platform for hospital administrators. By integrating various IoT systems, including parking management system, monitoring system, access control system, security alarm system, personnel positioning system, building control system, elevator system, ward and call system, it achieves real-time monitoring and effective management of various hospital resources. Figure 12 shows the main interface of the system.

[0132] This system offers a wide variety of services, has widely distributed computing nodes, and a complex network environment. Therefore, a reasonable service deployment and interaction scheme is essential for integrating and fully scheduling various IoT systems. Figure 13 shows further system design based on the system framework described in this paper, which greatly improves the system's development efficiency and ultimately achieves coordinated collaboration among the various systems, making the entire platform a highly unified whole and ensuring the efficient and stable operation of the entire system.

[0133] It should be noted that embodiments of the present invention can be implemented in hardware, software, or a combination of both. The hardware portion can be implemented using dedicated logic; the software portion can be stored in memory and executed by a suitable instruction execution system, such as a microprocessor or dedicated-design hardware. Those skilled in the art will understand that the above-described devices and methods can be implemented using computer-executable instructions and / or included in processor control code, for example, such code provided on a carrier medium such as a disk, CD, or DVD-ROM, a programmable memory such as read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The devices and modules of the present invention can be implemented by hardware circuitry such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field-programmable gate arrays, programmable logic devices, etc., or by software executed by various types of processors, or by a combination of the above-described hardware circuitry and software, such as firmware.

[0134] <1> Comparison of Deployment Algorithms for Fixed Location Service Nodes

[0135] The edge computing node distribution in this experiment is based on the EUA dataset, which includes the actual edge computing node (communication base station) distribution status in the Melbourne CBD, Australia, published by the Australian Communications and Media Authority (ACMA); the service process model adopts the strategy of randomizing the workflow of distributed heterogeneous systems proposed by Stavrinides et al.

[0136] This paper quantitatively analyzes the Pareto solution set of the corresponding multi-objective algorithm by evaluating the convergence, evenness, and spread of the solution set.

[0137] Table 1. Quantitative Analysis Indicators of Pareto Solutions

[0138]

[0139] Table 1 further provides a quantitative analysis of the Pareto solutions of different algorithms. Regarding solution set convergence, the solution set coverage of the proposed algorithm exceeds that of the NSGAII-based algorithm by more than 8 times. In terms of solution set uniformity, the minimum distance standard deviation of the proposed algorithm's solution set is only 77.8% of that of the NSGAII-based algorithm. Regarding solution set breadth, the continuous solution distance of the proposed algorithm's solution set is only 41.9% of that of the NSGAII-based algorithm. Quantitative descriptions of convergence, uniformity, and breadth show that the proposed algorithm outperforms the NSGAII-based algorithm in all three aspects, demonstrating its applicability in this type of scenario.

[0140] <2> Comparison of Entity Service Node Location Mobile Service Deployment Algorithms

[0141] To verify the suitability of the proposed mobile edge deployment algorithm in communication-intensive IoT scenarios, we compared the fitness calculation results of the RandomPlacement (RP) algorithm, the GASP algorithm (based on a genetic algorithm for offline deployment of mobile edge services), and the GenDoc algorithm (based on a greedy approach) under the aforementioned environments. The comparison results are shown in Figure 14.

[0142] The algorithm in this paper stabilizes around generation 70, with a final sample fitness of around 2500, representing a 39.1% improvement over the initial service distribution fitness of 4100. The GASP algorithm's final sample fitness stabilizes around 2800, meaning the algorithm in this paper reduces fitness calculation by 10.7% in the same scenario. The GenDoc algorithm achieves a sample fitness of around 2900, resulting in a 13.7% reduction in fitness calculation for the same scenario. Random algorithms consistently fail to yield effective results; in 200 generations of random selection, the best-fit solution corresponds to a sample fitness of around 3200, a reduction of 21.8% for the algorithm in this paper. This aligns with the expectations for the algorithm in this scenario, fully demonstrating its rationality and efficiency in addressing mobile service deployment problems in service-intensive scenarios.

[0143] <3> Comparison of application time consumption in real-world IoT scenarios

[0144] Figure 15 shows a comparison of the response speeds of selected IoT applications in the above system with those in the traditional framework. It is evident that the proposed method significantly reduces the average access time for each IoT application in the aforementioned practical project, demonstrating the correctness and usability of the algorithm and system framework presented in this paper.

[0145] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions, and improvements made by those skilled in the art within the scope of the technology disclosed in the present invention, and within the spirit and principles of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for autonomous deployment of edge services in service-intensive IoT scenarios, characterized in that, The edge service autonomous deployment method for service-intensive IoT scenarios includes: First, abstracting IoT services into process components, planning data flow and service scheduling in the network using various process modeling methods, and performing mathematical analysis on the process model; selecting a suitable process model coding scheme and integrating the graphical process modeling into the automated deployment framework; Second, constructing a service network model and a distributed microservice data communication model, and using the service network model and distributed microservice data communication model for automatic monitoring of network data; Then, based on the acquired raw IoT data, setting a suitable service deployment algorithm according to the mobility of physical service nodes to obtain the optimal service deployment scheme; Finally, automatically deploying various microservices in the code repository to the IoT service network according to the generated server deployment scheme; The method also includes the following steps: Step 1, using a graphical interface to process the user's predetermined data flow according to the standard process model protocol. The process involves four steps: 1) Modeling: Mapping the constructed graphical model onto a mathematical model to verify the security of the process and converting the graphical model's solution into readable matrix data; 2) Constructing a service network and data communication model: Defining the network topology using the service network and data communication model, and determining the service discovery and service data interaction schemes within the network based on the defined topology; Simultaneously, monitoring the network infrastructure environment under the defined topology for various microservice operating states, load balancing states, server hardware states, and dynamic changes in the network environment; 3) Setting service deployment algorithms suitable for fixed entity services in various scenarios, obtaining the optimal service deployment scheme based on raw IoT data; Using Monte Carlo simulation, with the fixed service optimization algorithm as the single-sample objective, determining the best entity service movement scheme; 4) Generating an automated deployment scheme based on Kubernetes, packaging all individual microservices in the cloud code repository into Docker containers. Images are stored in a cloud service image repository. The optimal service deployment scheme is transformed into XML information carried by the service deployment. The IoT service is automatically deployed to the edge server cluster according to a reasonable deployment scheme using a containerization scheme. Step three includes: 1) Obtaining the deployment information of each event service node in the network using a fixed-location service deployment algorithm for entity service nodes: taking the target entity service node as a fixed-location entity service node, using a dominance-based multi-objective optimization heuristic algorithm with the deployment quantity and final network consumption as scheduling objectives, and filtering to obtain the deployment information of each event service node in the network; 2) Obtaining a dataset of randomly distributed entity service points, using the fixed-location service deployment algorithm for entity service nodes as the objective function of a single sample, calculating the ideal optimization objective of a large number of samples based on the randomly distributed entity service point dataset, and obtaining a reasonable deployment scheme for mobile service deployment of entity service nodes.

2. The edge service autonomous deployment method for service-intensive IoT scenarios as described in claim 1, characterized in that, The construction of the service network and data communication model includes: Step (1) Building a multi-level service registry based on the ZooKeeper distributed service system, and using the multi-level service registry to perform autonomous service discovery and registration, autonomous message flow and service load balancing; Step (2) Building a message queue using RocketMQ, and performing multi-type flow of messages within sub-networks, between networks, one-to-one, and one-to-many according to message attributes; and monitoring the hardware attributes of edge servers in the entire network based on ZABBIX; Step (3) Integrating the information from Step (1) to Step (2) into a service and data communication model into the coding framework.

3. The edge service autonomous deployment method for service-intensive IoT scenarios as described in claim 1, characterized in that, The raw IoT data includes: workflow information, service interaction information, service network topology, and edge server hardware information.

4. The edge service autonomous deployment method for service-intensive IoT scenarios as described in claim 1, characterized in that, The definition of network topology using service networks and data communication models includes: adopting a microservice service model to abstract services and data; abstracting original service nodes into entity service nodes and event service nodes; managing the service status and data of each sub-network structure through a multi-level service registry; enabling message data flow and remote service invocation among each sub-service network structure based on the message queue cluster of the multi-level service registry; and, based on the constraints of event flow, disseminating events to the entire service network according to the multi-level service registry; wherein the entire service network is a tree-shaped network topology.

5. A system for autonomous deployment of edge services in service-intensive IoT scenarios, implementing the autonomous deployment method for edge services in service-intensive IoT scenarios as described in claim 1, characterized in that, The edge service autonomous deployment system for service-intensive IoT scenarios includes: a process design module, used to abstract IoT services into process components, plan data flow and service scheduling in the network using various process modeling methods, and perform mathematical analysis on the process model; simultaneously, it selects and writes appropriate process model coding schemes, integrating graphical process modeling into the automated deployment framework; it models the user-defined data flow according to standard process model protocols through a graphical interface; it maps the constructed graphical model to a mathematical model, verifies the security of the process, and converts the graphical model scheme into readable matrix data; and a network environment monitoring module, used to construct a service network and data communication model; and utilize the service network and data communication model... The system defines the network topology and determines the service discovery and service data interaction schemes within the network based on this topology. Simultaneously, it monitors the network infrastructure environment under the defined topology, including the running status of various microservices, load balancing status, server hardware status, and dynamic changes in the network environment. A server deployment scheme generation module sets service deployment algorithms suitable for various scenarios, obtaining the optimal service deployment scheme based on raw IoT data. Using Monte Carlo simulation with the fixed service optimization algorithm as a single-sample objective, it determines the best entity service movement scheme. An automated deployment module generates automated deployment schemes for Kubernetes, packaging all individual microservices in the cloud code repository into Docker images and storing them in the cloud service image repository. It converts the obtained optimal service deployment scheme into XML information carried by the service deployment, and uses containerization to automatically deploy IoT services to the edge server cluster according to a reasonable deployment scheme.

6. A computer device, characterized in that, The computer device includes a memory and a processor. The memory stores a computer program, which, when executed by the processor, causes the processor to perform the steps of the edge service autonomous deployment method for service-intensive IoT scenarios as described in any one of claims 1-4.

7. A computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the edge service autonomous deployment method for service-intensive Internet of Things scenarios as described in any one of claims 1-4.

8. An information data processing terminal, characterized in that, The information data processing terminal is used to implement the edge service autonomous deployment system for service-intensive IoT scenarios as described in claim 5.

Images