Industrial Internet T-shaped distributed intelligent interconnection cluster architecture, 3M implementation method and software platform

Abstract

This invention provides an industrial internet T-shaped distributed intelligent interconnected cluster architecture, a 3M implementation method, and a software platform, including: a horizontal platform, composed of a master cluster and slave clusters, with data communication between the master and slave clusters achieved through physical link redundancy technology, cluster metadata synchronization mechanism, hierarchical message system, and edge computing; a vertical platform, including production execution systems and industrial equipment, used for industrial data acquisition and control command execution; and distributed intelligent interconnected cluster enterprise standards, including internal standards of the horizontal platform, internal standards of the vertical platform, and standards between the horizontal and vertical platforms, used to standardize service calls, resource access, and communication protocols. This invention's integrated T-shaped architecture deeply integrates internet-sensitive services with stable industrial production, achieving comprehensive connectivity across all elements, the entire industrial chain, and the entire value chain of cross-industry, cross-level, and cross-regional collaborative capabilities, realizing full-scenario integration from manufacturing to services.

Description

Technical Field

[0001] This invention relates to the field of industrial internet system software technology, specifically to an industrial internet T-shaped distributed intelligent interconnection cluster architecture, a 3M implementation method, and a software platform. Background Technology

[0002] The T-shaped distributed intelligent interconnected cluster architecture and the 3M method of the Industrial Internet have been proven feasible in ultra-large-scale group enterprises, significantly improving cross-industry collaboration efficiency, system stability, and security. The software platform (ePlat / iPlat), as the core supporting tool, has achieved deep integration of horizontal service integration and vertical manufacturing connectivity. Future development requires further optimization of edge intelligence algorithms and protocol ecosystems to adapt to a wider range of industrial scenarios.

[0003] As the types of industrial equipment increase, the protocol adaptation layer needs to be continuously expanded (such as adding 5G industrial protocols). The false alarm rate of the model under complex working conditions still needs to be further reduced. In extreme network fluctuation scenarios, metadata synchronization is occasionally delayed (probability <0.1%).

[0004] Patent document CN114979189B's unified access spatiotemporal big data REST service across industrial internet systems, considering dedicated high-speed fiber optic cables, does not address data encryption and data signing. Patent document CN114979189B's spatiotemporal big data REST microservice does not cover security details. Patent document CN114979189B's spatiotemporal big data REST microservice does not cover stability assurance details. Patent document CN107147728B is only applicable to the specific system of object storage systems and cannot be applied to industrial internet systems.

[0005] Summary of common defects in existing technologies: Most technologies require high-speed fiber optic support and cannot adapt to the low-speed network environment of the Internet under the "one headquarters, multiple bases" model; the existing architecture focuses on vertical control (such as within a single system) and lacks the ability to coordinate all elements across industries and regions; it generally lacks mechanisms such as multi-tenant isolation, data encryption, rate limiting and circuit breaking, making it difficult to cope with the high concurrency and high reliability requirements of ultra-large-scale clusters; the protocols are single and the architecture is rigid, which cannot support the access of multiple types of industrial equipment and the dynamic expansion and contraction requirements. Summary of the Invention

[0006] To address the shortcomings of existing technologies, the purpose of this invention is to provide an industrial internet T-shaped distributed intelligent interconnection cluster architecture, a 3M implementation method, and a software platform.

[0007] The industrial internet T-shaped distributed intelligent interconnection cluster architecture provided by the present invention includes:

[0008] The horizontal platform consists of a master cluster and slave clusters. The master cluster and slave clusters communicate with each other through physical link redundancy technology, cluster metadata synchronization mechanism, hierarchical message system and edge computing.

[0009] The vertical platform, which includes production execution systems and industrial equipment, is used for industrial data acquisition and control command execution.

[0010] Enterprise standards for distributed intelligent interconnected clusters include internal standards for horizontal platforms, internal standards for vertical platforms, and standards between horizontal and vertical platforms, which are used to standardize service calls, resource access, and communication protocols.

[0011] Preferably, the physical link redundancy technology includes:

[0012] At least two independent physical links are established between the master cluster and the slave cluster, and link redundancy is achieved through dual network interface card bonding technology.

[0013] When the network packet loss rate exceeds a preset threshold, it automatically switches to a backup link. The link selection strategy is based on the preset operator priority.

[0014] Preferably, the cluster metadata synchronization mechanism includes:

[0015] Event-driven synchronization: Device status change events are pushed to the message queue in real time, and the message format follows the CloudEvents standard;

[0016] Batch synchronization: Synchronize historical data via SFTP at preset time points, mark the difference data as sync_status:pending, and update it to synced after the synchronization is completed;

[0017] Data versioning: Using Delta Lake to preserve historical versions, supporting backtracking queries by timestamp.

[0018] Preferably, the hierarchical messaging system includes three types of standardized messages:

[0019] Command message: Control commands issued by the master cluster, compressed in Avro format, including command type, target cluster ID and timestamp;

[0020] Data messages: Production data packets transmitted between master and slave clusters are required to include a CRC32 checksum, and a retransmission mechanism is triggered when the checksum fails.

[0021] Control messages: Used for cluster health monitoring. The message body contains node status codes and is synchronized at a preset frequency via heartbeat packets.

[0022] Preferably, the edge computing includes:

[0023] Deploy the model for anomaly detection; the anomaly determination formula is as follows:

[0024]

[0025] Where n is the total number of sensors, w i μ is the weight of the sensor. i σ is the historical mean. i x is the standard deviation. i It is the real-time measurement value of the i-th sensor.

[0026] The 3M implementation method of the T-shaped distributed intelligent interconnection cluster architecture of the industrial Internet provided by the present invention includes: tenant isolation, unified and decentralized integration, and multi-sided cloud access;

[0027] Tenant isolation refers to ensuring that the data, resources and services of different tenants are independent of each other in the shared infrastructure, and that the enterprise standard for distributed intelligent interconnection clusters is followed within the tenant and between tenants.

[0028] The so-called unified and decentralized combination means that the distributed intelligent cluster includes multiple horizontal platforms, one of which is the central control node and the other horizontal platforms are edge nodes; the horizontal platforms internally follow the enterprise standard of the distributed intelligent cluster, and the horizontal platforms between each other also follow the enterprise standard of the distributed intelligent cluster.

[0029] The term "multi-sided cloud access" refers to a horizontal platform connecting to multiple vertical platforms. The vertical platforms include various types of production execution systems and various types of industrial equipment. The vertical platforms internally follow the distributed intelligent cluster enterprise standard, and the horizontal and vertical platforms also follow the distributed intelligent cluster enterprise standard.

[0030] Preferably, the tenant isolation includes:

[0031] Technical architecture layer isolation: Each tenant is allocated an independent logical resource pool through a virtual network, and network policy isolation is achieved using SDN technology;

[0032] Policy management: Define tenant role permissions based on the RBAC model to restrict the scope of resource operations;

[0033] Security hardening: Sensitive data is encrypted using AES-256-GCM, with the key dynamically allocated by cluster-level KMS and updated according to a preset cycle.

[0034] Preferably, the combination of centralized and decentralized management includes:

[0035] Dynamic resource allocation algorithm: Automatically adjusts resource allocation based on cluster load. The calculation formula is as follows:

[0036]

[0037] Hierarchical access control: The central control node is responsible for global policies, while edge nodes autonomously adjust the priority of local resources.

[0038] The software platform provided by the present invention includes:

[0039] ePlat: Supports horizontal service integration and provides master-slave cluster management, tenant isolation, and multilateral protocol conversion functions;

[0040] iPlat: Supports vertical manufacturing integration, enabling real-time data synchronization between the production execution system and industrial equipment.

[0041] Preferably, the ePlat's API gateway performs the following functions: converting the REST API of the horizontal platform into OPC UA method calls of the vertical platform; data mapping uses JSONPath expression parsing, and the mapping relationship table contains field mappings and data type casting rules.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] (1) By adopting the T-shaped distributed intelligent cluster architecture, the problems of low vertical integration efficiency, lack of horizontal integration capability of the industrial chain, and lack of support for distributed mode in the vertical five-layer architecture of the ISA-95 "Information System and Control System Integration" international standard were solved.

[0044] (2) By adopting a horizontal and vertical integrated "T" structure, the problem of a distributed deployment mode with one headquarters and multiple bases has been solved. The horizontal and vertical integrated "T" architecture deeply integrates industrial steady-state production and Internet agile services, and realizes the full-element, full-industry chain and full-value chain of cross-industry, cross-level and cross-regional collaborative capabilities of the entire group, achieving remarkable results.

[0045] (3) By adopting the 3M architecture of tenant isolation architecture (Multi-Tenant), unified and decentralized architecture (Master-Slave), and multi-edge cloud architecture (Multi-Edge), the key technical problems of high concurrency and high reliability of massive industrial big data have been solved. It has realized ultra-large-scale processing, ultra-complex scenario application, cross-industry and cross-regional business collaboration under the distributed manufacturing mode of one headquarters and multiple bases. It supports the uninterrupted operation of nearly 100,000 concurrent users and millions of online users in the Baowu Steel ecosystem, as well as massive high-concurrency access in complex transaction consistency scenarios. Attached Figure Description

[0046] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0047] Figure 1 It is a T-shaped distributed intelligent interconnected cluster architecture;

[0048] Figure 2 This is a schematic diagram of the tenant isolation structure;

[0049] Figure 3 This is a schematic diagram of a unified and decentralized structure;

[0050] Figure 4 This is a schematic diagram of a multi-sided cloud access structure;

[0051] Figure 5 Flowchart of the 3M implementation method. Detailed Implementation

[0052] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0053] Example

[0054] This invention provides a T-shaped distributed intelligent interconnection cluster architecture, a 3M implementation method, and a software platform, specifically:

[0055] like Figure 1 The T-shaped distributed intelligent cluster architecture addresses the problems of excessive layers, lack of flatness, low vertical integration efficiency, and insufficient horizontal integration capabilities inherent in the ISA-95 international standard's five-layer vertical architecture for "Information Systems and Control Systems Integration." The architecture includes: a horizontal platform, a vertical platform, and a distributed intelligent cluster enterprise standard.

[0056] The horizontal platform includes a master cluster and slave clusters. The T-shaped distributed intelligent interconnected cluster architecture comprises one master cluster and any number of slave clusters. The master cluster internally adheres to the enterprise standard for distributed intelligent interconnected clusters. The slave clusters internally adhere to the enterprise standard for distributed intelligent interconnected clusters. The master cluster and slave clusters communicate with each other in accordance with the enterprise standard for distributed intelligent interconnected clusters.

[0057] The data communication process between the master cluster and slave clusters includes physical link redundancy technology, cluster metadata synchronization mechanism, hierarchical message system between clusters, and edge computing.

[0058] Physical link redundancy technology: Dual NIC bonding technology is employed, establishing at least two independent physical links between the primary and secondary clusters. Two independent physical links are provided by two unconnected operators selected from China Telecom, China Mobile, and China Unicom. The backup channel is automatically switched when network jitter is detected (e.g., packet loss rate > 5%).

[0059] Cluster metadata synchronization mechanism: The master cluster and slave clusters must ensure metadata synchronization. Slave clusters periodically send heartbeat packets to the master cluster and send change information when service status, user status, tenant status, or internal cluster failures occur, in order to maintain cluster metadata and ensure global state consistency. The synchronization process must follow the message format defined in the "Horizontal Platform Internal Intelligent Connectivity Enterprise Standard".

[0060] Layered messaging system between clusters: Defines three types of standardized messages, which are transmitted using Avro format compression.

[0061] 1. Command message: Commands issued by the master cluster;

[0062] 2. Data messages: Data packets transmitted between the master cluster and slave clusters;

[0063] 3. Control messages: Cluster health status monitoring and fault notification.

[0064] Edge computing: Through the Multi-Edge feature, slave clusters can directly handle 80% of localized requests on the edge side, and only upload the 20% of data that requires global collaboration to the master cluster, reducing cross-cluster traffic.

[0065] The vertical platform includes a production execution system (MES) and industrial equipment. The MES is used for industrial manufacturing. The MES collects data from the industrial equipment and sends control commands to it. The industrial equipment sends industrial data to the MES and simultaneously receives control commands from it.

[0066] The aforementioned distributed intelligent interconnection cluster enterprise standard is used to replace the ISA-95 "Information System and Control System Integration" international standard, including: intelligent interconnection enterprise standard within horizontal platforms, intelligent interconnection enterprise standard within vertical platforms, and intelligent interconnection enterprise standard between horizontal and vertical platforms.

[0067] The horizontal platform's internal intelligent interconnection enterprise standards are designed to ensure efficient collaboration and data consistency between the master and slave clusters. These standards include data format standards, communication protocol specifications, and access control and security.

[0068] Data format standards include a unified data model and metadata specifications.

[0069] Unified data model: Adopts a Protobuf-based schema definition to ensure cross-cluster data parsing compatibility.

[0070] Metadata specifications: It is mandatory for data packets to include metadata such as source_cluster_id (cluster ID) and data_version (version number) for conflict detection and tracing.

[0071] Communication protocol specifications include RPC frameworks and message queue protocols.

[0072] RPC framework: The gRPC framework must be used, and the interface definition between the master cluster and the slave cluster must conform to the IDL (Interface Description Language) specification.

[0073] Message Queue Protocol: In asynchronous communication scenarios, the message body is required to conform to the CloudEvents standard and support Kafka as the transport layer.

[0074] Access and security include the RBAC model and data encryption.

[0075] RBAC model: Cluster nodes must pass two-way TLS authentication (mTLS) and API access permissions are restricted based on roles.

[0076] Data encryption: Sensitive data must be encrypted using AES-256-GCM, and the key is managed uniformly by the cluster-level KMS.

[0077] The vertical platform's internal intelligent connectivity standards aim to achieve seamless integration and real-time control of production execution systems and industrial equipment. These standards include device access specifications, production system interaction, security and reliability, and disaster recovery and backup.

[0078] Device access specifications include protocol compatibility and data sampling frequency.

[0079] Protocol compatibility: Forced support for OPC UA, Modbus TCP, and MQTT protocols. Data reported by the device must include mandatory fields such as the device unique identifier (device_id) and timestamp. The timestamp must conform to the ISO8601 standard.

[0080] Data sampling frequency: The data reporting interval of sensor devices shall not exceed 1 second, and the control command response delay shall be less than 50ms.

[0081] Production system interaction includes instruction standardization and status synchronization.

[0082] Instruction standardization: Instructions issued by the production execution system must conform to the ISA-88 batch control model.

[0083] Status synchronization: The device needs to provide real-time feedback of operation_state and error_code, and the production execution system needs to update the local database within 200ms.

[0084] Security and reliability include physical isolation and data integrity.

[0085] Physical isolation: Critical equipment (such as PLCs) must be isolated through industrial firewalls, allowing only whitelisted IPs to access designated ports (such as port 4840 of OPC UA).

[0086] Data integrity: Data reported by the device must be accompanied by a CRC32 checksum. If the production execution system fails to verify the data, a retransmission mechanism will be triggered.

[0087] Disaster recovery and backup include local caching and version compatibility.

[0088] Local caching: The production execution system needs to store the device data of the most recent 24 hours at the edge node and continue to execute the preset control logic when the network is down.

[0089] Version compatibility: Device firmware upgrades must support dual versions coexisting (A / B partitions), and rollback time must not exceed 5 minutes.

[0090] The enterprise standard for intelligent connectivity between horizontal and vertical platforms aims to streamline data and control flows between production execution systems and cross-cluster business systems. This standard includes protocol conversion rules, security isolation mechanisms, data synchronization strategies, and auditing and traceability.

[0091] Protocol conversion rules include API gateway and data mapping table.

[0092] API Gateway: REST APIs on horizontal platforms need to be automatically converted to OPC UA method calls on vertical platforms.

[0093] Data mapping table: Defines field mapping relationships (e.g., horizontal mapping of order_quantity to vertical mapping of Batch.size), and supports JSONPath expression parsing.

[0094] Security isolation mechanisms include DMZ deployment and access control.

[0095] DMZ Deployment: Horizontal and vertical platforms must communicate through the API gateway in the DMZ zone; direct connection to the production network is prohibited.

[0096] Access control: The vertical platform only allows access from specific service accounts on the horizontal platform, with permissions refined to the device group level.

[0097] Data synchronization strategies include event-driven architecture and batch synchronization.

[0098] Event-driven architecture: Device state change events on the vertical platform need to be pushed to the message queue of the horizontal platform in real time.

[0099] Batch synchronization: The horizontal platform triggers a batch task every day at midnight to synchronize historical production data (CSV format) of the vertical platform via SFTP. Difference data must be marked with sync_status:pending.

[0100] Auditing and tracing include log correlation and data versioning.

[0101] Log association: Cross-platform operations require the generation of a globally unique trace_id that conforms to the OpenTracing standard and supports tracing the entire chain (such as order creation → device execution → result feedback) through Jaeger.

[0102] Data versioning: Data synchronized in both directions must retain historical versions (based on Delta Lake) and support querying and backtracking by timestamp.

[0103] like Figure 5 The 3M implementation method includes: multi-tenant isolation, master-slave integration, and multi-edge cloud access.

[0104] like Figure 2 The aforementioned multi-tenant isolation uses technical means to ensure that the data, resources, and services of different tenants are independent of each other within a shared infrastructure. The horizontal platform contains multiple tenants, and all resources owned by each tenant, including but not limited to users, user groups, data, etc., are isolated from each other. Within a tenant, the enterprise standard for distributed intelligent interconnected clusters is followed. Between tenants, the enterprise standard for distributed intelligent interconnected clusters is also followed. The specific implementation process includes technical architecture layer isolation, policy management and access control, and security hardening.

[0105] Technical architecture layer isolation includes virtualization technology, network isolation, and storage isolation.

[0106] Virtualization technology: Allocates an independent logical resource pool to each tenant through virtual networks and virtualized instances.

[0107] Network isolation: Software-defined networking (SDN) technology is used to configure independent network policies for different tenants.

[0108] Storage isolation: Using logical volume partitioning, distributed storage, and database sharding technologies, ensure that tenant data is physically or logically independent.

[0109] Policy management and access control include RBAC (role-based access control), quota management, and dynamic policy allocation.

[0110] RBAC (Role-Based Access Control): Defines tenant-specific roles and permissions to restrict the scope of tenant operations on resources.

[0111] Quota management: Set resource quotas to prevent a single tenant from over-consuming resources.

[0112] Dynamic policy allocation: In software-defined wide area networks (SD-WAN), traffic priorities or security rules are dynamically adjusted according to business needs.

[0113] Security hardening includes encryption and tunneling technologies, as well as security auditing and monitoring.

[0114] Encryption and tunneling technology: TLS is used to encrypt tenant data transmission channels, ensuring data privacy.

[0115] Security audit and monitoring: Ensure the effectiveness of isolation mechanisms through logging, intrusion detection systems (IDS), and regular penetration testing.

[0116] like Figure 3 The aforementioned master-slave architecture refers to a distributed intelligent cluster comprising multiple horizontal platforms, one of which serves as the central control node, while the others act as edge nodes. Each horizontal platform adheres to the enterprise standards for distributed intelligent clusters. Inter-platform communication also follows these enterprise standards. The implementation process includes architecture design, resource integration and flexible allocation, and decentralized governance of permissions and services.

[0117] The architecture design includes a unified management platform and a layered execution mechanism.

[0118] Unified management platform: Establish a central control node responsible for global resource scheduling, policy distribution and monitoring.

[0119] Layered execution mechanism: Localized execution is achieved at edge nodes, reducing the pressure on the central node.

[0120] Resource integration and flexible allocation include resource pooling and dynamic allocation.

[0121] Resource pooling: Virtualize physical resources into a shared pool and allocate them to different levels or tenants as needed.

[0122] Dynamic allocation: Automatically adjusts resource allocation based on load.

[0123] Decentralized governance of permissions and services includes permission layering and service differentiation.

[0124] Hierarchical access control: The central control node is responsible for strategic planning and security compliance, while edge nodes or tenants have operational autonomy.

[0125] Service Differentiation: Provide customized services for each tenant through a multi-tenant architecture.

[0126] like Figure 4The term "multi-edge cloud access" refers to a system that is numerous, with horizontal platforms capable of connecting to multiple vertical platforms; and diverse in type, with vertical platforms encompassing various types of production execution systems and industrial equipment. Vertical platforms internally adhere to the enterprise standard for distributed intelligent interconnected clusters. Similarly, horizontal and vertical platforms communicate using the same enterprise standard. The implementation process includes a protocol adaptation layer, edge computing, and elastic scaling mechanisms.

[0127] The protocol adaptation layer includes an industrial protocol conversion engine that supports interoperability with 15+ standard protocols. It converts industrial protocols from vertical platforms into target protocols for horizontal platforms, supports dynamic loading of the industrial protocol conversion engine, and achieves hot deployment in ≤50ms.

[0128] Edge computing includes implementing traffic tiering strategies and edge-side anomaly detection.

[0129] Implement a traffic tiering strategy, which divides data into three levels: real-time data, key metrics, and global data. Different strategies are applied to each level of data, described using Mermaid syntax as follows:

[0130] graph LR

[0131] A [Edge Device] --> | 60% Real-Time Data | B (Local Rule Engine)

[0132] A --> | 30% Key Metrics | C (Subordinate Cluster Preprocessing)

[0133] A --> |10% Global Data|D (Main Cluster Decision Center)

[0134] Edge-side anomaly detection involves deploying the DeepSeek-R1 deep thinking model to complete 80% of anomaly detection tasks at the edge using artificial intelligence.

[0135] Elastic scaling mechanisms include horizontal platform scaling mechanisms, vertical platform scaling mechanisms, and dynamic bandwidth allocation algorithms.

[0136] The scaling mechanism of a horizontal platform means that a horizontal platform can connect to one vertical platform or multiple vertical platforms.

[0137] The scaling mechanism of a vertical platform means that a vertical platform can connect to one or multiple production execution systems, and can connect to one type of industrial equipment or multiple types of industrial equipment.

[0138] Dynamic bandwidth allocation algorithms refer to strategies that limit communication between vertical and horizontal platforms, shape traffic, and ensure the system stability of horizontal platforms. For example, the core logic can be represented in Python code as follows, where throughput represents traffic and latency represents delay:

[0139]

[0140] The software platform includes ePlat and iPlat. ePlat is used to implement the master cluster and all slave clusters. iPlat is used to implement the production execution system.

[0141] Those skilled in the art will understand that, in addition to implementing the system, apparatus, and their modules provided by this invention in purely computer-readable program code, the same program can be implemented in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system, apparatus, and their modules provided by this invention can be considered a hardware component, and the modules included therein for implementing various programs can also be considered structures within the hardware component; alternatively, modules for implementing various functions can be considered both software programs implementing the method and structures within the hardware component.

[0142] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A 3M implementation method for an Industrial Internet T-shaped distributed intelligent interconnected cluster architecture, characterized in that, The industrial internet T-shaped distributed intelligent interconnection cluster architecture includes: The horizontal platform consists of a master cluster and slave clusters. The master cluster and slave clusters communicate with each other through physical link redundancy technology, cluster metadata synchronization mechanism, hierarchical message system and edge computing. The vertical platform, which includes production execution systems and industrial equipment, is used for industrial data acquisition and control command execution. Enterprise standards for distributed intelligent interconnected clusters include standards within horizontal platforms, standards within vertical platforms, and standards between horizontal and vertical platforms, which are used to standardize service calls, resource access, and communication protocols. The 3M implementation methods include: tenant isolation, centralized and decentralized integration, and multi-sided cloud access; Tenant isolation refers to ensuring that the data, resources and services of different tenants are independent of each other in the shared infrastructure, and that the enterprise standard for distributed intelligent interconnection clusters is followed within the tenant and between tenants. The so-called unified and decentralized combination means that the distributed intelligent cluster includes multiple horizontal platforms, one of which is the central control node and the other horizontal platforms are edge nodes; the horizontal platforms internally follow the enterprise standard of the distributed intelligent cluster, and the horizontal platforms between each other also follow the enterprise standard of the distributed intelligent cluster. The term "multi-sided cloud access" refers to a horizontal platform connecting to multiple vertical platforms. The vertical platforms include various types of production execution systems and various types of industrial equipment. The vertical platforms internally follow the distributed intelligent cluster enterprise standard, and the horizontal and vertical platforms also follow the distributed intelligent cluster enterprise standard.

2. The 3M implementation method of the T-shaped distributed intelligent interconnection cluster architecture of the Industrial Internet according to claim 1, characterized in that, The physical link redundancy technology includes: At least two independent physical links are established between the master cluster and the slave cluster, and link redundancy is achieved through dual network interface card bonding technology. When the network packet loss rate exceeds a preset threshold, it automatically switches to a backup link. The link selection strategy is based on the preset operator priority.

3. The 3M implementation method of the T-shaped distributed intelligent interconnection cluster architecture of the Industrial Internet according to claim 1, characterized in that, The cluster metadata synchronization mechanism includes: Event-driven synchronization: Device status change events are pushed to the message queue in real time, and the message format follows the CloudEvents standard; Batch synchronization: Synchronize historical data via SFTP at preset time points, mark the difference data as sync_status:pending, and update it to synced after the synchronization is completed; Data versioning: Using Delta Lake to preserve historical versions, supporting backtracking queries by timestamp.

4. The 3M implementation method of the T-shaped distributed intelligent interconnection cluster architecture of the Industrial Internet according to claim 1, characterized in that, The hierarchical messaging system includes three types of standardized messages: Command message: Control commands issued by the master cluster, compressed in Avro format, including command type, target cluster ID and timestamp; Data messages: Production data packets transmitted between master and slave clusters are required to include a CRC32 checksum, and a retransmission mechanism is triggered when the checksum fails. Control messages: Used for cluster health monitoring. The message body contains node status codes and is synchronized at a preset frequency via heartbeat packets.

5. The 3M implementation method of the T-shaped distributed intelligent interconnection cluster architecture of the Industrial Internet according to claim 1, characterized in that, The edge computing includes: Deploy the model for anomaly detection; the anomaly determination formula is as follows: Where n is the total number of sensors, w i μ is the weight of the sensor. i σ is the historical mean. i x is the standard deviation. i It is the real-time measurement value of the i-th sensor.

6. The 3M implementation method of the T-shaped distributed intelligent interconnection cluster architecture of the Industrial Internet according to claim 1, characterized in that, The tenant isolation includes: Technical architecture layer isolation: Each tenant is allocated an independent logical resource pool through a virtual network, and network policy isolation is achieved using SDN technology; Policy management: Define tenant role permissions based on the RBAC model to restrict the scope of resource operations; Security hardening: Sensitive data is encrypted using AES-256-GCM, with the key dynamically allocated by cluster-level KMS and updated according to a preset cycle.

7. The 3M implementation method of the T-shaped distributed intelligent interconnection cluster architecture of the Industrial Internet according to claim 1, characterized in that, The integration of centralized and decentralized management includes: Dynamic resource allocation algorithm: Automatically adjusts resource allocation based on cluster load. The calculation formula is as follows: Hierarchical access control: The central control node is responsible for global policies, while edge nodes autonomously adjust the priority of local resources.

8. A software platform, characterized in that, A 3M implementation method comprising the industrial internet T-type distributed intelligent interconnected cluster architecture according to any one of claims 1 to 7, including: ePlat: Supports horizontal service integration and provides master-slave cluster management, tenant isolation, and multilateral protocol conversion functions; iPlat: Supports vertical manufacturing integration, enabling real-time data synchronization between the production execution system and industrial equipment.

9. The software platform according to claim 8, characterized in that, The ePlat API gateway performs the following functions: converting the REST API of the horizontal platform into OPC UA method calls of the vertical platform; data mapping uses JSONPath expression parsing, and the mapping relationship table contains field mapping and data type casting rules.

Images