Data center equipment out-of-band remote operation and management control system and method based on fusion of 5g / 4g iot private network

Abstract

The present application relates to the technical field of remote operation and maintenance of data center equipment, in particular to a data center equipment out-of-band remote operation and maintenance management and control system and method combined with 5G / 4G Internet of Things private network, comprising the following steps: an out-of-band state acquisition unit; a 5G / 4G Internet of Things private network transmission unit; a remote operation and maintenance management and control processing unit; and an out-of-band instruction execution unit. The present application performs protocol conversion, field attribute annotation and standardized packaging on multi-source heterogeneous out-of-band data, constructs a state occurrence time envelope in combination with acquisition timeliness, transmission timeliness and link state metadata, and generates a state effective domain and an event time envelope according to the field physical type; further, fault evidence access is implemented through the common intersection of the state effective domain and the three-state event sequence constraint, forming a classified operation and maintenance management and control instruction. Thus, the state time sequence relationship can be accurately restored under the conditions of link switching, data retransmission and delay fluctuation, and non-synchronous state miscombination and event sequence misjudgment can be reduced.

Description

Technical Field

[0001] This invention relates to the field of remote operation and maintenance technology for data center equipment, and more specifically, to an out-of-band remote operation and maintenance management system and method for data center equipment that integrates 5G / 4G IoT private networks. Background Technology

[0002] As data centers evolve towards large-scale, distributed, and edge computing, the demand for remote operation and maintenance of servers, storage devices, and network equipment continues to grow. Out-of-band remote operation and maintenance systems collect device status, operation logs, and alarm information through independent management interfaces and communication links, and perform control operations such as restarts and power-on / off cycles. They have become an important technological foundation for ensuring business continuity and unattended operation of data centers.

[0003] Currently, out-of-band (O&M) technologies widely used in the industry mainly fall into two categories. The first category is a centralized O&M solution based on a wired management network. This typically consists of a device-side board management controller, a dedicated management switch, and a centralized management platform. It collects device temperature, voltage, power supply, fan status, startup status, and operation logs via IPMI, Redfish, SNMP, or vendor-specific protocols, and issues control commands through an independent Ethernet management channel. This solution is technologically mature, has a relatively complete interface system, and is suitable for data centers with stable network conditions and centralized device deployment. The second category is a remote O&M access solution based on 4G or 5G communication links. This typically involves configuring industrial gateways, cellular communication modules, or wireless CPEs on the data center side. It establishes an independent communication channel through operator private networks, dedicated APNs, VPNs, or network slicing to transmit device status, alarm logs, and control commands to a remote O&M platform. This solution offers flexible deployment, reduces reliance on business networks and fixed lines, and is suitable for edge data centers, unattended sites, and cross-regional device O&M scenarios.

[0004] The two types of solutions mentioned above still face several common technical problems in engineering applications. First, heterogeneous state data differ in acquisition cycles, timestamp accuracy, data retention characteristics, and link transmission latency. Remote platforms struggle to accurately reconstruct the actual occurrence range and effective continuity of each state on the device side. Especially under conditions of link switching, data retransmission, or clock skew, discrepancies easily arise between the platform's reception time and the actual occurrence time of the state, making it difficult to reliably correlate state records from different sources and of different types on a unified time scale. Second, complex faults typically require comprehensive judgment based on multiple state variables and log events. However, existing solutions often rely on discrete sampled values, log timestamps, or data arrival order to establish correlations, making it difficult to confirm whether multiple states truly exist simultaneously, and difficult to reliably determine their chronological order when event times overlap or are uncertain. This lack of effective constraints on the composability and credibility of fault evidence leads to incorrect splicing of abnormal states, misjudgment of the sequence of critical events, and further, false alarms, missed alarms, delayed handling, or mis-issuance of high-risk control commands, limiting the application of existing solutions in cross-regional, weak network, and high-reliability operation and maintenance scenarios. In view of this, we propose an out-of-band remote operation and maintenance management system and method for data center equipment that integrates 5G / 4G IoT private networks. Summary of the Invention

[0005] The purpose of this invention is to provide an out-of-band remote operation and maintenance management system and method for data center equipment that integrates 5G / 4G IoT private networks. This addresses the problems mentioned in the background art, such as the difficulty in accurately reconstructing the actual occurrence range and effective continuity of states due to differences in collection time base, field attributes, and network transmission latency of heterogeneous out-of-band state data, as well as the difficulty in reliably correlated multiple states with log events in complex fault scenarios, leading to insufficient credibility of fault evidence combinations and low accuracy and security of operation and maintenance management decisions.

[0006] To address the aforementioned technical problems, one objective of this invention is to provide an out-of-band remote operation and maintenance management system for data center equipment that integrates 5G / 4G IoT private networks, comprising: The out-of-band status acquisition unit receives the out-of-band status, operation logs and collected metadata of the equipment to be maintained in the data center, and outputs a standardized out-of-band status dataset with collection time and field attribute identifiers after protocol conversion and standardized encapsulation. The 5G / 4G IoT private network transmission unit receives standardized datasets uplink, performs private network slicing and multi-link redundancy scheduling, attaches transmission timeliness and link status metadata, and encrypts and transmits it to the remote operation and maintenance management unit; it receives classified operation and maintenance management instruction sets downlink, encrypts them using the same mechanism, and transmits them to the out-of-band instruction execution unit. The remote operation and maintenance management processing unit receives a standardized dataset with attached transmission metadata, constructs a state occurrence time envelope by combining the collection and transmission timeliness, generates a state valid domain and an event time envelope according to field attributes, completes fault access determination based on the intersection of the state valid domains and the three-state event sequence constraints, and generates a classified operation and maintenance management instruction set. The out-of-band instruction execution unit receives a set of classified operation and maintenance management instructions, and after device-side protocol mapping and permission verification, issues out-of-band operations to the target device.

[0007] As a further improvement to this technical solution, the out-of-band status acquisition unit includes a data access and protocol adaptation module, a field attribute and timeliness labeling module, and a standardized encapsulation output module, wherein: The data access and protocol adaptation module receives out-of-band hardware status signals, operation logs and collected metadata from each device to be maintained in the data center, performs protocol conversion processing on the multi-source heterogeneous out-of-band raw data, and obtains an out-of-band status raw dataset with a unified data format. The field attribute and timeliness labeling module receives the out-of-band status raw dataset in a unified data format, extracts the field physical attributes and collection time information corresponding to each status record, and adds field attribute identifiers and collection timeliness identifiers to each out-of-band status record. The standardized encapsulation output module receives the labeled out-of-band status records, performs standardized encapsulation processing on the labeled out-of-band status records, and generates a standardized out-of-band status dataset carrying the collection time and field attribute identifiers.

[0008] As a further improvement to this technical solution, the 5G / 4G IoT private network transmission unit includes a private network slicing scheduling and data offloading module, a transmission link metadata mounting module, and a bidirectional encrypted forwarding module, wherein: The private network slice scheduling and data splitting module receives the standardized out-of-band status dataset with collection time and field attribute identifiers output by the out-of-band status acquisition unit, and the classified operation and maintenance management instruction set output by the remote operation and maintenance management processing unit. Based on the business attributes of the two types of data, it matches the corresponding private network slice resources and completes the transmission path allocation and redundant link reservation for the two types of data respectively. The transmission link metadata mounting module receives a standardized out-of-band status dataset carrying collection time and field attribute identifiers after transmission path allocation, records the transmission time, link identifier and transmission status information of the corresponding transmission link, generates transmission time and link status metadata and mounts it to the corresponding dataset. The bidirectional encrypted forwarding module receives a standardized out-of-band status dataset carrying collection time and field attribute identifiers after attaching transmission time and link status metadata, as well as a classified operation and maintenance management instruction set allocated by the transmission path. After performing encryption and encapsulation processing on the two types of data respectively, the standardized out-of-band status dataset carrying collection time and field attribute identifiers is forwarded to the remote operation and maintenance management processing unit, and the classified operation and maintenance management instruction set is forwarded to the out-of-band instruction execution unit.

[0009] As a further improvement to this technical solution, the remote operation and maintenance management processing unit includes a state occurrence time envelope construction module, a state valid domain and event envelope generation module, a simultaneous state reconstruction and event sequence determination module, and a fault evidence access and instruction generation module, wherein: The state occurrence time envelope construction module receives a standardized out-of-band state dataset carrying the collection time and field attribute identifiers, as well as the corresponding transmission time and link status metadata, generates the state occurrence time envelope of each out-of-band state record, and outputs it to the state valid field and event envelope generation module. The state valid domain and event envelope generation module receives the state occurrence time envelope and the field attribute identifiers in the standardized out-of-band state dataset carrying the collection time and field attribute identifiers, generates the corresponding state valid domain and event time envelope and outputs them to the synchronous state reconstruction and event sequence determination module. The synchronous state reconstruction and event sequence determination module receives multiple sets of valid state domains and event time envelopes corresponding to a preset fault scenario, generates synchronous observable state or non-synchronous state identifiers, and three-state event sequence constraints. The three-state event sequence constraints refer to three types of event sequence determination results obtained by classifying the event time envelopes corresponding to multiple log events according to the relative position and overlap relationship between the event time envelopes. These results are: determined forward sequence, determined reverse sequence, and uncertain sequence. These are used to verify whether the event sequence relationship required by the preset fault scenario has temporal evidence to support it. The fault evidence admission and instruction generation module receives simultaneous observable states, non-simultaneous state identifiers and three-state event sequence constraints, performs fault evidence admission judgment, and generates a set of classified operation and maintenance management instructions for the target equipment.

[0010] As a further improvement to this technical solution, the state occurrence time envelope construction process of the state occurrence time envelope construction module includes the following steps: S31.1. Based on the transmission timeliness and link status metadata, the lower and upper bounds of the delay of the corresponding transmission link are statistically determined. The upper bound of the delay is extended by combining the retransmission status and link switching status of the current status record to obtain the extended upper bound of the delay of the corresponding status record. The original lower bound of the delay and the extended upper bound of the delay together constitute the upper and lower bounds of the delay of the status record. S31.2 Based on the upper and lower bounds of the delay recorded in this state, calculate the offset range of the source acquisition time base relative to the remote analysis time base; S31.3 Combining the collection time identifier in the standardized out-of-band state dataset carrying collection time and field attribute identifiers with the offset interval, the collection time of each out-of-band state record is mapped to a state occurrence time envelope with upper and lower time boundaries.

[0011] As a further improvement to this technical solution, the state valid domain and event envelope generation module includes a continuous state valid domain generation submodule, a sustained state valid domain generation submodule, and an instantaneous event envelope generation submodule. These three submodules process the input data in parallel based on field attribute identifiers, wherein: The continuous state valid domain generation submodule receives the state occurrence time envelope and field attribute identifier corresponding to the continuous class state, and extends the valid retention time in combination with the update rule of the continuous state to generate the state valid domain corresponding to the continuous state. The state validity field generation submodule receives the state occurrence time envelope and field attribute identifier corresponding to the state of the maintained class, and extends the validity duration by combining the state duration rule of the maintained state to generate the state validity field corresponding to the maintained state. The instantaneous event envelope generation submodule receives the state occurrence time envelope corresponding to the instantaneous event class log and directly determines the state occurrence time envelope as the event time envelope.

[0012] As a further improvement to this technical solution, the synchronous state reconstruction and event sequence determination module includes a synchronous state reconstruction submodule and a three-state event sequence determination submodule. The two submodules process in parallel and independently and jointly output the determination result, wherein: The synchronous state reconstruction submodule receives the state effective domains of multiple necessary states corresponding to a preset fault scenario and solves the common intersection of each state effective domain. When the intersection is not empty, a synchronous observable state is constructed. When the intersection is empty, a non-synchronous state identifier is generated, which corresponds to the uncombinable state of the fault evidence, and the corresponding state is prohibited from being combined into joint fault evidence. The three-state event sequence determination submodule receives event time envelopes corresponding to multiple logs and generates three-state event sequence constraints based on the relative positions and overlapping relationships of the event time envelopes, which determine the forward sequence, the reverse sequence, and the uncertain sequence.

[0013] As a further improvement to this technical solution, the fault evidence access and classification operation and maintenance management instruction generation process of the fault evidence access and instruction generation module includes the following steps: S34.1 Using the simultaneous observable state, non-simultaneous state identifiers and three-state event sequence constraints as pre-entry conditions, verify whether the common intersection of the state effective domains of the necessary states corresponding to the preset fault scenario exists, and whether the event sequence relationship matches the preset fault scenario requirements. S34.2. Based on the verification results, the fault evidence is divided into three categories: confirmable, unverifiable, and uncombinable. If the verification results show that the event sequence is reversed from the preset fault scenario requirements, the fault scenario is directly determined to be invalid and is not included in the above three categories of fault evidence status. S34.3. For the three types of fault evidence status, generate fault handling instructions, status verification instructions, log replenishment instructions or operation suppression instructions respectively, to form a set of classified operation and maintenance management instructions for the target device.

[0014] As a further improvement to this technical solution, the out-of-band instruction execution unit includes an instruction permission verification module, a device-side protocol mapping module, and an out-of-band control operation distribution module, wherein: The instruction permission verification module receives the classified operation and maintenance management instruction set output by the 5G / 4G IoT private network transmission unit, performs operation permission verification on the classified operation and maintenance management instruction set, and outputs the classified operation and maintenance management instruction set that has passed the permission verification. The device-side protocol mapping module receives the categorized operation and maintenance control instruction set that has passed the permission verification, performs device-side protocol mapping processing on the categorized operation and maintenance control instruction set according to the out-of-band management protocol specification of the target device, and generates control instructions that are adapted to the protocol format of the target device. The out-of-band control operation sending module receives control instructions adapted to the target device's protocol format and sends corresponding out-of-band control operations to the target device.

[0015] The second objective of this invention is to provide an out-of-band remote operation and maintenance management method for data center equipment that integrates 5G / 4G IoT private networks. Based on the above-mentioned out-of-band remote operation and maintenance management system for data center equipment that integrates 5G / 4G IoT private networks, the method includes the following steps: S1, receiving out-of-band hardware status signals, operation logs and collected metadata of each device to be operated and maintained in the data center, performing protocol conversion and standardized encapsulation processing on the multi-source heterogeneous out-of-band raw data to obtain a standardized out-of-band status dataset carrying the collection time and field attribute identifiers; S2. Using private network slicing and multi-link redundancy scheduling technology, link optimization processing is performed on the standardized out-of-band status dataset carrying collection timeliness and field attribute identifiers. After adding transmission timeliness and link status metadata, encrypted transmission processing is performed. S3. Combine the collection time and field attribute identifier, transmission time and link status metadata to construct the status occurrence time envelope of each out-of-band status record; S4. Based on the physical type of the status field, generate the valid status field and event time envelope for each out-of-band status record. S5. Match the preset fault scenario, solve the common intersection of the state effective domains of the corresponding necessary states, generate simultaneous observable state or non-simultaneous state identifiers, and determine the three-state event order constraints of forward sequence, reverse sequence, and uncertain order. S6. Using simultaneous observable state, non-simultaneous state identifiers and three-state event sequence constraints as prerequisite admission conditions, perform fault evidence admission judgment and generate a set of classified operation and maintenance control instructions for the target equipment. S7. After performing link optimization processing on the classified operation and maintenance management instruction set using private network slicing and multi-link redundancy scheduling technology, encrypted transmission processing is performed. S8. Perform device-side protocol mapping and operation permission verification on the classified operation and maintenance management instruction set, and issue corresponding out-of-band control operations to the target device.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention constructs a state occurrence time envelope for out-of-band state records by performing protocol conversion, field attribute annotation, and standardized encapsulation on multi-source heterogeneous out-of-band data, and combining this with acquisition timeliness, transmission timeliness, and link status metadata. Simultaneously, based on the physical type of the state field, it generates continuous state valid fields, sustained state valid fields, and instantaneous event time envelopes. Therefore, it can uniformly express the actual occurrence range and effective continuity relationship of state records under conditions of link switching, data retransmission, clock skew, and transmission delay fluctuations, reducing the correlation deviation of data from different sources and of different types on the time scale, and improving the remote platform's ability to reconstruct device operating status and the reliability of time series analysis. 2. This invention addresses a preset fault scenario by solving for the common intersection of the effective domains of multiple necessary states. Based on the relative positions and overlaps between the time envelopes of multiple events, it forms a three-state event order constraint: a definite forward sequence, a definite reverse sequence, and an uncertain sequence. Furthermore, using the synchronic relationship of states and the temporal relationship of events as the admission criteria for fault evidence, the fault evidence is divided into three categories: confirmable, unverifiable, and non-combinable. Fault handling, state verification, log replenishment, or operation suppression instructions are generated accordingly. This reduces erroneous combinations of non-synchronous states and the forced determination of uncertain event sequences, improves the composability and reliability of complex fault evidence, reduces the possibility of misjudgment, missed judgment, and erroneous issuance of high-risk control instructions, and enhances the accuracy and security of remote out-of-band maintenance decisions. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the overall method steps of the present invention; Figure 2This is a schematic diagram of the overall system structure framework of the present invention; The meanings of the labels in the diagram are as follows: 1. Out-of-band status acquisition unit; 11. Data access and protocol adaptation module; 12. Field attribute and timeliness labeling module; 13. Standardized encapsulation output module; 2. 5G / 4G IoT private network transmission unit; 21. Private network slicing scheduling and data offloading module; 22. Transmission link metadata mounting module; 23. Two-way encrypted forwarding module; 3. Remote operation and maintenance management and control unit; 31. State occurrence time envelope construction module; 32. State valid domain and event envelope generation module; 33. Simultaneous state reconstruction and event sequence determination module; 34. Fault evidence access and instruction generation module; 4. Out-of-band instruction execution unit; 41. Instruction permission verification module; 42. Device-side protocol mapping module; 43. Out-of-band control operation distribution module. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0019] like Figure 1 As shown, this embodiment provides an out-of-band remote operation and maintenance management system for data center equipment that integrates a 5G / 4G IoT private network. It is suitable for unattended data center scenarios where equipment is interconnected with a remote operation and maintenance platform via a 5G / 4G IoT private network. The system consists of an out-of-band status acquisition unit 1, a 5G / 4G IoT private network transmission unit 2, a remote operation and maintenance management processing unit 3, and an out-of-band instruction execution unit 4. The units interact using standardized out-of-band status datasets, link metadata, and categorized operation and maintenance management instruction sets, forming a closed loop from status acquisition, transmission, analysis to instruction issuance.

[0020] In this embodiment, out-of-band state acquisition unit 1 completes protocol adaptation and annotation encapsulation of multi-source heterogeneous data; 5G / 4G IoT private network transmission unit 2 completes link optimization and link metadata mounting; remote operation and maintenance management processing unit 3 constructs a state occurrence time envelope using acquisition timeliness and link metadata, and expands it into a state valid domain or event time envelope according to the field physical type. Then, it determines the state synchronicity and event sequence by the common intersection of valid domains and the overlapping relationship of event time envelopes, and executes fault evidence access and generates classified operation and maintenance management instructions accordingly. Finally, out-of-band instruction execution unit 4 completes instruction permission verification, protocol mapping, and issuance. The following describes the module composition, processing flow, and data transmission relationship of each unit in detail: Out-of-band status acquisition unit 1 receives out-of-band hardware status signals, operation logs, and acquired metadata from various devices under maintenance within the data center. It then performs protocol conversion and standardized encapsulation processing on the multi-source heterogeneous out-of-band raw data to obtain a standardized out-of-band status dataset carrying acquisition timeliness and field attribute identifiers. This includes a data access and protocol adaptation module 11, a field attribute and timeliness labeling module 12, and a standardized encapsulation output module 13, wherein: The data access and protocol adaptation module 11 receives out-of-band hardware status signals, operation logs and collected metadata of each device to be maintained in the data center, performs protocol conversion processing on the multi-source heterogeneous out-of-band raw data, and obtains an out-of-band status raw dataset with a unified data format. Specifically, the data access and protocol adaptation module 11 acquires input data through the out-of-band management interface of the device-side baseboard management controller. The input data includes three categories: the first category is out-of-band hardware status signals, covering hardware monitoring data such as temperature sensor readings, power supply output voltage values, real-time fan speed, and various module presence signals; the second category is operation logs, covering system event logs, the management controller's own operation logs, and hardware fault alarm records; the third category is acquired metadata, covering auxiliary information such as the acquisition time, device identifier, and sensor number corresponding to each status record.

[0021] In real-world environments, the communication protocols used by out-of-band management interfaces of devices from different manufacturers and models vary. Common protocols include IPMI, Redfish, SNMP, and some vendor-specific protocols. To address this, the data access and protocol adaptation module 11 is internally configured with multi-protocol parsing adapters corresponding to each of these protocols. Upon receiving data, it first identifies the protocol type of the current data based on the message format or port identifier. Then, it calls the corresponding parsing adapter to perform field parsing and format reshaping according to the protocol's preset field mapping rules, uniformly converting the raw data from various protocols into a consistent out-of-band status raw dataset. This unified data format uses a structured record with a fixed field template of "Device Identifier—Sensor Type Code or Log Type Code—Value or Status Description—Original Acquisition Time," eliminating differences at the field level for data from different protocol sources and providing a consistent data foundation for subsequent processing.

[0022] The field attribute and timeliness labeling module 12 receives the out-of-band status raw dataset in a unified data format, extracts the field physical attributes and collection time information corresponding to each status record, and adds field attribute identifiers and collection timeliness identifiers to each out-of-band status record. Regarding the extraction of field physical attributes, the field attribute and timeliness labeling module 12 pre-defines a mapping table between sensor type, log type, and physical state type. This embodiment classifies physical state types into three categories: continuous state, sustained state, and transient event. A continuous state refers to a state type where the physical quantity changes continuously over time, and a single sampled value is considered valid within the sampling interval before the next sampling time arrives. Typical examples include processor temperature, motherboard temperature, fan speed, and power supply output voltage. A sustained state refers to a state type where the state value remains valid once recorded until a reverse state record is received. Typical examples include power-on state, fan in-position state, management controller online state, and power module in-position state. A transient event refers to a log-type state that occurs only at a specific moment and does not possess a continuous attribute. Typical examples include fan failure events, temperature exceeding limits events, power module offline events, and management controller reset events. The field attribute and timeliness labeling module 12 assigns a corresponding field attribute identifier to each state record based on the above mapping relationship.

[0023] Regarding the generation of acquisition timeliness identifiers, the field attribute and timeliness labeling module 12 extracts the original acquisition time from each status record and converts the acquisition times from different devices and different protocol sources into a time expression based on the local clock of the acquisition unit. When there is a deviation between the device-side clock and the local clock of the acquisition unit, the field attribute and timeliness labeling module 12 uses the local clock of the acquisition unit as a unified reference and converts the original acquisition time into a millisecond-level offset relative to the start of the preset acquisition cycle, thereby generating an acquisition timeliness identifier with a unified time reference. After the above processing, each out-of-band status record carries both a field attribute identifier and an acquisition timeliness identifier.

[0024] The standardized encapsulation output module 13 receives the labeled out-of-band status records, performs standardized encapsulation processing on the labeled out-of-band status records, and generates a standardized out-of-band status dataset carrying the collection time and field attribute identifiers.

[0025] Specifically, standardized encapsulation processing refers to organizing each labeled out-of-band status record into a uniformly formatted data packet according to a preset data structure. This data structure includes device identifier, field attribute identifier, status value or status description, acquisition timeliness identifier, sensor type code or log type code, and reserved extended fields. The data packet is organized using the lightweight JSON structured data format, which facilitates cross-unit transmission and rapid parsing.

[0026] After encapsulation, the standardized encapsulation output module 13 outputs the generated standardized out-of-band state dataset to the 5G / 4G IoT private network transmission unit 2. This standardized dataset achieves uniformity at the format level and explicitly carries the physical type and acquisition time information of each state record at the information level. It provides structured input data for the subsequent 5G / 4G IoT private network transmission unit 2 to attach transmission link metadata and for the remote operation and maintenance management processing unit 3 to construct the state occurrence time envelope, generate the state valid field and event time envelope.

[0027] The 5G / 4G IoT private network transmission unit 2 receives a standardized out-of-band status dataset carrying collection timeliness and field attribute identifiers output by the out-of-band status acquisition unit 1. It performs link optimization processing using private network slicing and multi-link redundancy scheduling technology, adds transmission timeliness and link status metadata to the standardized out-of-band status dataset carrying collection timeliness and field attribute identifiers, and then performs encrypted transmission processing before transmitting the standardized out-of-band status dataset carrying collection timeliness and field attribute identifiers to the remote operation and maintenance management processing unit 3. The 5G / 4G IoT private network transmission unit 2 also receives a categorized operation and maintenance management instruction set output by the remote operation and maintenance management processing unit 3. It performs link optimization processing using private network slicing and multi-link redundancy scheduling technology, and then performs encrypted transmission processing before transmitting the categorized operation and maintenance management instruction set to the out-of-band instruction execution unit 4. The unit includes a private network slicing scheduling and data splitting module 21, a transmission link metadata mounting module 22, and a bidirectional encrypted forwarding module 23, wherein: The private network slice scheduling and data diversion module 21 receives the standardized out-of-band status dataset with collection time and field attribute identifiers output by the out-of-band status acquisition unit 1, and the classified operation and maintenance management instruction set output by the remote operation and maintenance management processing unit 3. It matches the corresponding private network slice resources according to the business attributes of the two types of data, and completes the transmission path allocation and redundant link reservation for the two types of data respectively. Specifically, the transmission processing of 5G / 4G IoT private network transmission unit 2 is divided into uplink and downlink directions. The processing logic of the two links is independent of each other, which will be explained below: In the uplink direction, the standardized out-of-band status dataset output by the out-of-band status acquisition unit 1, carrying acquisition timeliness and field attribute identifiers, first enters the private network slice scheduling and data offloading module 21. This dataset belongs to the status monitoring service, with moderate bandwidth requirements but stable transmission. In the 5G / 4G IoT private network environment, the private network slice scheduling and data offloading module 21 matches it to the enhanced mobile broadband slice or low latency high reliability slice. Afterwards, the private network slice scheduling and data offloading module 21 maintains an available uplink pool. The link pool contains real-time status information of multiple physical links or logical links. According to the preset link quality evaluation strategy, the private network slice scheduling and data offloading module 21 performs a comprehensive score based on indicators such as real-time signal strength, link load, available bandwidth, and bit error rate of each link. It selects the link with the highest comprehensive score from the available uplink pool as the current primary uplink transmission link, and selects one or more links with the second highest comprehensive score as backup uplink links. When the quality index of the primary uplink drops below the preset switching threshold or is interrupted, the private network slice scheduling and data diversion module 21 automatically switches the uplink data traffic to the backup uplink and updates the current primary uplink identifier. After completing the uplink transmission path allocation and redundant link reservation, the standardized out-of-band status dataset is delivered to the transmission link metadata mounting module 22 and enters the transmission metadata mounting stage.

[0028] After uplink metadata mounting and encryption processing (this part is completed by the transmission link metadata mounting module 22 and the bidirectional encryption forwarding module 23, and the specific processing process will be described in detail below), the standardized out-of-band status dataset is sent to the remote operation and maintenance management unit 3 through the selected uplink transmission link.

[0029] In the downlink direction, the categorized operation and maintenance management instruction set output by the remote operation and maintenance management processing unit 3 enters the private network slice scheduling and data diversion module 21. This instruction set belongs to the control category and has high requirements for latency and reliability. The private network slice scheduling and data diversion module 21 prioritizes matching it to low-latency, high-reliability slices. Similar to the uplink processing, the private network slice scheduling and data diversion module 21 maintains an available downlink pool. It performs a comprehensive evaluation based on indicators such as real-time signal strength, link load, available bandwidth, and bit error rate of each link. It selects the link with the highest comprehensive score from the available downlink pool as the current primary downlink transmission link, and selects one or more links with the second highest comprehensive score as backup downlink links. When the quality indicators of the primary downlink drop below the preset switching threshold or are interrupted, the private network slice scheduling and data diversion module 21 automatically switches the downlink data traffic to the backup downlink link and updates the current primary downlink identifier. After completing the downlink transmission path allocation and redundant link reservation, the categorized operation and maintenance management instruction set is delivered to the bidirectional encrypted forwarding module 23 and enters the encrypted forwarding stage. It should be noted that the classification operation and maintenance management instruction set does not need to go through the transmission link metadata mounting module 22 in the downlink direction, because the transmission metadata of the downlink instruction is not a necessary input for the fault evidence admission judgment of the remote operation and maintenance management processing unit 3. The transmission link metadata mounting module 22 only performs metadata mounting processing on the status dataset in the uplink direction.

[0030] The transmission link metadata mounting module 22 receives the standardized out-of-band status dataset carrying the collection time and field attribute identifiers after the transmission path allocation, records the transmission time, link identifier and transmission status information of the corresponding transmission link, generates transmission time and link status metadata and mounts it to the corresponding dataset. Specifically, in the uplink direction, when the normalized out-of-band state dataset is about to be transmitted through the selected primary uplink, the transmission link metadata mounting module 22 records the transmission time. This transmission time is provided by the local clock of the 5G / 4G IoT private network transmission unit 2, with millisecond precision. Simultaneously, the transmission link metadata mounting module 22 records the currently used uplink identifier, for example, the primary uplink identifier is UL_Link_A, and the backup uplink identifier is UL_Link_B. Transmission status information includes whether this uplink transmission was successfully transmitted on the first attempt, whether retransmissions occurred and the number of retransmissions, whether uplink switching was triggered, and the uplink identifiers before and after the switching.

[0031] The above information together constitutes the uplink transmission timeliness and link status metadata. The transmission link metadata mounting module 22 appends the above metadata to the normalized out-of-band status dataset in the form of structured fields. The appending method can be to embed the metadata fields into the header area of ​​the data packet, or to append them as independent data segments to the tail of the data packet.

[0032] After mounting, the dataset, in addition to the original collection time and field attribute identifiers, further carries the sending time, uplink identifier and transmission status information related to this uplink transmission process, and is then delivered to the bidirectional encrypted forwarding module 23.

[0033] The bidirectional encrypted forwarding module 23 receives a standardized out-of-band status dataset carrying collection time and field attribute identifiers after attaching transmission time and link status metadata, as well as a classified operation and maintenance management instruction set allocated by the transmission path. After performing encryption and encapsulation processing on the two types of data respectively, the standardized out-of-band status dataset carrying collection time and field attribute identifiers is forwarded to the remote operation and maintenance management processing unit 3, and the classified operation and maintenance management instruction set is forwarded to the out-of-band instruction execution unit 4.

[0034] Specifically, the bidirectional encrypted forwarding module 23 establishes independent encrypted channels for uplink and downlink service data. In the uplink direction, the bidirectional encrypted forwarding module 23 encrypts the payload portion of the standardized out-of-band status dataset containing transmission timeliness and link status metadata. The encryption method can be either the national cryptographic SM4 algorithm or a transport layer security protocol. After encryption, the necessary routing information in the packet header is retained, and the packet is sent to the remote operation and maintenance management unit 3 via the selected uplink transmission link. In the downlink direction, the bidirectional encrypted forwarding module 23 encrypts the payload portion of the classified operation and maintenance management instruction set. The encryption method is the same as for the uplink or an independent downlink encryption strategy is used. After encryption, the packet is sent to the out-of-band instruction execution unit 4 via the selected downlink transmission link.

[0035] After the bidirectional encryption and forwarding module 23 completes the encryption and forwarding processes in both directions, a complete bidirectional transmission process of the 5G / 4G IoT private network transmission unit 2 ends. The remote operation and maintenance management processing unit 3 receives a standardized out-of-band status dataset carrying triple tags: collection timeliness, field attribute identifier, and transmission timeliness and link status metadata. These three tags together provide the remote operation and maintenance management processing unit 3 with complete time and status basis on the collection side and transmission side for subsequent construction of status occurrence time envelope, calculation of transmission delay boundary, and execution of fault evidence admission judgment.

[0036] The remote operation and maintenance management processing unit 3 receives a standardized out-of-band status dataset carrying collection timeliness and field attribute identifiers, along with corresponding transmission timeliness and link status metadata. It combines the collection timeliness and field attribute identifiers with the transmission timeliness and link status metadata to complete the interval mapping from the source-end collection time base to the remote analysis time base, constructing the state occurrence time envelope of each out-of-band status record under the remote analysis time base. The state occurrence time envelope is a time interval with clearly defined upper and lower time boundaries, used to characterize the reliable time range of a single out-of-band status record after transmission delay correction. Based on the physical type of the status field, it performs differentiated processing on continuous status, sustained status, and transient event logs, generating corresponding state valid domains and event time envelopes. The state valid domain is a time interval obtained by extending the effective holding time based on the state occurrence time envelope and the update pattern of the corresponding state, used to limit the participation of a single status data in joint fault determination. The system defines the effective time range and uses pre-defined fault scenarios to extract necessary state conditions. The intersection of the valid state domains of each state is used as the admission criterion for joint fault evidence. When the intersection is not empty, a synchronous observable state of the target device is constructed; when the intersection is empty, a non-synchronous state identifier is generated, and the corresponding state combination is prohibited from forming joint fault evidence. Simultaneously, a three-state event sequence constraint is generated based on the relative position and overlap of the event time envelopes of multiple log events. This constraint represents three types of timing judgment results based on the event time envelope: determined forward sequence, determined reverse sequence, and uncertain sequence. These are used to verify whether the causal timing of events required by the pre-defined fault scenario is supported by temporal evidence. Finally, combining the judgment rules of the pre-defined fault scenario, the system executes fault evidence admission judgment based on the synchronous observable state, non-synchronous state identifier, and three-state event sequence constraint, generating a classification and maintenance management instruction set for the target device.

[0037] Specifically, the remote operation and maintenance management and control processing unit 3 is the decision-making center of the entire system. It is deployed on the remote operation and maintenance platform side and is responsible for the fusion processing of the standardized out-of-band status dataset and corresponding transmission link metadata sent by the 5G / 4G IoT private network transmission unit 2. It completes the entire decision-making link from the time metadata of the collection side and the transmission side to the judgment of the credibility of fault evidence and the generation of classification and control instructions.

[0038] It should be noted that the remote operation and maintenance management processing unit 3 maintains an internal fault rule base, which stores rule definitions for various preset fault scenarios pre-configured by operation and maintenance personnel. Each fault rule includes at least a fault scenario identifier, a list of necessary states corresponding to the fault scenario, and event timing conditions corresponding to the fault scenario. Upon receiving a new valid state domain and event time envelope, the remote operation and maintenance management processing unit 3 iterates through the preset fault scenarios in the fault rule base, matching the currently arriving valid state domain and event time envelope with the list of necessary states for each fault scenario. If a match is successful, the judgment process for that fault scenario is triggered. The following explanation is based on four stages: construction of the state occurrence time envelope, generation of the valid state domain and event envelope, simultaneous state reconstruction and event sequence determination, and fault evidence admission and instruction generation. The state occurrence time envelope construction module 31 receives the standardized out-of-band state dataset carrying the acquisition time and field attribute identifiers, as well as the corresponding transmission time and link status metadata, generates the state occurrence time envelope of each out-of-band state record, and outputs it to the state valid field and event envelope generation module 32. The state occurrence time envelope construction process of the state occurrence time envelope construction module 31 includes the following steps: S31.1. Based on the transmission timeliness and link status metadata, the lower and upper bounds of the delay of the corresponding transmission link are statistically determined. The upper bound of the delay is extended by combining the retransmission status and link switching status of the current status record to obtain the extended upper bound of the delay of the corresponding status record. The original lower bound of the delay and the extended upper bound of the delay together constitute the upper and lower bounds of the delay of the status record. In this step, the state occurrence time envelope construction module 31 has a built-in transmission link delay statistics table, which records the lower and upper bounds of delay measured during the operation of each transmission link. The lower bound is the minimum delay achievable by the link under normal transmission conditions, and the upper bound is the maximum transmission delay of the link without retransmissions or link switching. This statistics table is updated periodically based on the delay distribution of historical transmission data. In an optional implementation, for the initial deployment of the system or the stage where there is insufficient historical statistical data for newly activated links, the lower and upper bounds of delay can be initialized using default empirical values ​​corresponding to the link type. For example, for 5G uplinks, the lower bound of delay is preset to 3ms and the upper bound to 100ms. After sufficient historical measurement data is accumulated, the default values ​​are replaced with actual statistical values.

[0039] Next, assuming the current state record is delivered via a certain transmission link, the state occurrence time envelope construction module 31 obtains the basic delay parameters of the link from the transmission link delay statistics table, where the lower and upper bounds of the delay are denoted as follows: and . This indicates the minimum latency that the link can achieve under normal transmission conditions. This indicates the maximum transmission delay of the link under conditions where no retransmissions or link switching occur.

[0040] After receiving the transmission time and link status metadata corresponding to the current status record, the status occurrence time envelope construction module 31 extracts the transmission status information of this transmission from the metadata, specifically including whether a retransmission occurred, the number of retransmissions, whether a link switch was triggered, and the link identifiers before and after the switch.

[0041] Based on this, if the transmission status information indicates that no retransmission occurred and no link switching was triggered in this transmission, then the upper limit of the delay remains unchanged from its initial value, and the upper limit of the delay is extended. Determine using the following formula: ; This indicates the upper bound of the extended delay after correction based on the current transmission conditions.

[0042] If the transmission status information indicates that a retransmission has occurred, the status occurrence time envelope construction module 31 constructs the envelope based on the number of retransmissions. and average additional delay per retransmission Extend the upper bound of the delay: ; This indicates the number of retransmissions for this transmission. The retransmission delay is set based on the historical average of the link. If there is no historical data, the default empirical value can be used.

[0043] If the transmission status information indicates that a link switch was triggered during this transmission, the status occurrence time envelope construction module 31 will include the switching interruption duration introduced by the link switch. Superimposed to the upper limit of latency: ; Indicates the duration of the link switching interruption. The handover protection duration is determined based on the system's preset duration. In a 5G private network, the typical handover interruption duration is 20ms to 50ms, and the specific value can be adjusted according to the actual network configuration.

[0044] If retransmission and link switching occur simultaneously, the upper bound of the delay is increased by the additional delay corresponding to both: ; After the above processing, the transmission delay of the state record is... The following interval constraints must be satisfied: ; This represents the transmission latency, which is the time it takes for data to travel from the 5G / 4G IoT private network transmission unit 2 to the remote operation and maintenance management processing unit 3. Its actual value is unknown, but it must fall within the lower bound of the latency. With extended upper delay Within the closed interval formed.

[0045] S31.2 Based on the upper and lower bounds of the delay recorded in this state, calculate the offset range of the source acquisition time base relative to the remote analysis time base; In this step, let's assume that the transmission time recorded by the 5G / 4G IoT private network transmission unit 2 when sending this status record is... The remote operation and maintenance management processing unit 3 receives this status record at the time when The time difference between the two The definition is as follows.

[0046] ; It represents the end-to-end time difference between the time of transmission and the time of reception.

[0047] Furthermore, transmission delay The true value is unknown; according to S31.1, its range is... Additionally, there is an inherent deviation between the local clock of the out-of-band status acquisition unit 1 and the local clock of the 5G / 4G IoT private network transmission unit 2. This deviation is calibrated using a lightweight clock synchronization mechanism, and its value range is as follows: ; This indicates the inherent clock offset between the acquisition unit and the transmission unit. This indicates the upper limit of clock deviation, in milliseconds. Additionally, the cross-node clock deviation between the 5G / 4G IoT private network transmission unit 2 and the remote operation and maintenance management processing unit 3 is already included in the upper and lower bounds of the latency statistics in the transmission link latency statistics table. This is covered by the measured distribution of historical end-to-end latency, eliminating the need to add an additional deviation term to the time base mapping formula.

[0048] Collection time Mapping value under remote analysis time base The following relationship is satisfied.

[0049] ; This indicates the time at which the status was recorded on the acquisition side. This indicates the time value mapped to the remote analysis time base at the time of acquisition.

[0050] Based on the above relationship, combined with and Their respective value ranges can be derived The value range of is as follows: ; This range is the offset range of the source acquisition time base relative to the remote analysis time base. Its lower bound corresponds to the case where the transmission delay is at its maximum value and the clock deviation is at its most negative value, and its upper bound corresponds to the case where the transmission delay is at its minimum value and the clock deviation is at its most positive value.

[0051] S31.3 Combining the collection time identifier and offset interval in the standardized out-of-band state dataset that carries the collection time and field attribute identifier, the collection time of each out-of-band state record is mapped to the state occurrence time envelope with upper and lower time boundaries.

[0052] In this step, the acquisition time indicator is the acquisition time based on the local clock of the acquisition unit. ,Will Substituting into the offset interval expression obtained in S31.2, the lower bound of the time for this state record under the remote analysis time base can be obtained. and the upper bound of time .

[0053] ; It represents the lower bound of the state occurrence time envelope, that is, the earliest possible occurrence time of the state record under the remote analysis time base.

[0054] ; This represents the upper bound of the state occurrence time envelope, i.e., the latest possible occurrence time of the state record under the remote analysis time base.

[0055] Depend on and The state occurrence time envelope is denoted as The time envelope indicates that, under the remote analysis time base, the actual occurrence time of the state record must fall within this time interval.

[0056] Furthermore, the state occurrence time envelope construction module 31 performs the processing steps S31.1 to S31.3 described above for each state record in the normalized out-of-band state dataset. Since each state record may be delivered via different transmission links during transmission, the latency statistics of each link are different, and the retransmission and link switching situations of each record may also differ. Therefore, the upper and lower bounds of latency and the offset intervals corresponding to each state record may differ, resulting in differentiated widths for the generated state occurrence time envelopes. This differentiated construction method ensures that each state record has a personalized time envelope that matches its actual transmission conditions, rather than using a uniform time tolerance range. After completing the state occurrence time envelope construction for all records, the state occurrence time envelope construction module 31 outputs each state occurrence time envelope along with the corresponding field attribute identifiers to the state valid field and event envelope generation module 32.

[0057] The state valid domain and event envelope generation module 32 receives the state occurrence time envelope and the field attribute identifiers from the normalized out-of-band state dataset carrying the acquisition time and field attribute identifiers. It generates the corresponding state valid domain and event time envelope and outputs them to the synchronous state reconstruction and event sequence determination module 33. The state valid domain and event envelope generation module 32 includes a continuous state valid domain generation submodule, a sustained state valid domain generation submodule, and an instantaneous event envelope generation submodule. These three submodules process the input data in parallel based on the field attribute identifiers. The continuous state valid domain generation submodule receives the state occurrence time envelope and field attribute identifiers corresponding to continuous states, and extends the valid retention time based on the update rules of continuous states to generate the state valid domain corresponding to the continuous states. Here, continuous states refer to state types where physical quantities change continuously over time and a single sampled value can be considered valid within the sampling interval. Typical examples include processor temperature, motherboard temperature, fan speed, and power supply output voltage. For this type of state, the value obtained from a single sample is considered valid until the next sampling time. Therefore, the continuous state valid domain generation submodule uses... As a lower bound of the valid domain of the state, Add a sampling period As the upper bound of the effective domain of states, the effective domain of continuous states is determined by the following formula: ; This indicates the sampling interval of out-of-band state acquisition unit 1, which is 1000ms in this embodiment. After this expansion, the effective domain of the continuous state covers the entire time period from the earliest possible occurrence of the state to the next sampling update time, ensuring that the state value can be reasonably referenced within this time period.

[0058] The hold-state valid field generation submodule receives the state occurrence time envelope and field attribute identifiers corresponding to the hold-state class, and extends the valid duration by combining the hold-state's state duration rules to generate the state valid field corresponding to the hold-state. The hold-state in this step refers to a state type that remains valid once its state value is recorded, until a contrary state record is received. Typical examples include power-on state, fan-in-place state, management controller online state, and power module-in-place state. The hold-state valid field generation submodule uses... The state valid domain extends in the future from this lower bound. Regarding the method for determining the endpoint of the state valid domain, the state valid domain generation submodule checks whether the opposite state record corresponding to the current state has been received before the current remote analysis time. If no opposite state record has been received, the endpoint of the state valid domain is the current remote analysis time plus the preset maximum acknowledgment time. The endpoint is calculated as follows: ; This indicates the current system time of remote operation and maintenance management unit 3. This indicates the maximum confirmation time. This is a configurable parameter used to limit the maximum range of the valid state domain extending in the future direction when there is a lack of opposite state records. In this embodiment, it is preset to 10s.

[0059] If a record of the opposite state obtained by matching based on a preset field attribute mapping table has been received, when multiple records of the opposite state exist, the earliest record of the opposite state in chronological order is taken as the basis for determining the valid endpoint. The preset field attribute mapping table is pre-configured with the mapping relationship between each hold-alive state and its corresponding opposite state based on the field attribute identifier and the state value, which is a conventional configurable mapping rule in this field. Let the lower bound of the state occurrence time envelope of the opposite state record be... ,like Then the endpoint of the valid domain of the state is taken as This indicates the effective duration of the current state up to the earliest possible occurrence of the opposite state record; if This indicates that the temporal envelope of the opposite state record overlaps with or is earlier than the starting point of the current state's valid domain. In this case, the valid domain of the state degenerates into a state with... Starting from, with A zero-length interval ending at a certain point indicates that, under the current evidence, it cannot be confirmed whether the state remains valid after the acquisition time. The zero-length state valid domain cannot participate in the common intersection solution of simultaneous state reconstruction, and the corresponding fault evidence is classified as a non-combinable state. The state valid domain is kept uniformly represented as follows: ;in This indicates the end point of the valid state field. Pick and The earlier of the two, when hour .

[0060] The transient event envelope generation submodule receives the state occurrence time envelope corresponding to the transient event log and directly determines this state occurrence time envelope as the event time envelope. Transient events refer to log-type states that occur only at a specific moment and do not possess a continuous validity attribute. Typical examples include fan failure events, temperature exceeding limits events, power module offline events, and management controller reset events. Since the occurrence of such events does not have a continuous characteristic, there is no need to extend the validity period; therefore, the transient event envelope generation submodule directly uses the state occurrence time envelope. This outputs the occurrence time interval of the event under the remote analysis time base.

[0061] After being processed separately by the state valid domain and event envelope generation module 32, different types of out-of-band state records are uniformly transformed into state valid domains or event time envelopes with clear time boundaries and physical semantics. Continuous states and held states express their valid time range that can be referenced in the form of state valid domains, while instantaneous events express their time range of occurrence in the form of event time envelopes. These time intervals together constitute the unified input for subsequent synchronicity determination and event sequence determination.

[0062] The synchronous state reconstruction and event sequence determination module 33 receives multiple sets of valid state domains and event time envelopes corresponding to a preset fault scenario, generates synchronous observable state or non-synchronous state identifiers, and three-state event sequence constraints; the synchronous state reconstruction and event sequence determination module 33 includes a synchronous state reconstruction submodule and a three-state event sequence determination submodule, the two submodules process in parallel and independently and jointly output the determination result, wherein: The synchronous state reconstruction submodule receives the state effective domains of multiple necessary states corresponding to a preset fault scenario and solves the common intersection of the state effective domains. When the intersection is not empty, it constructs synchronous observable states. When the intersection is empty, it generates non-synchronous state identifiers, which correspond to the uncombinable states of the fault evidence and prohibit the combination of the corresponding states into joint fault evidence. Specifically, the failure scenario requirements are set. The necessary states must be true simultaneously, and the valid domains of each necessary state are as follows: ; The synchronous state reconstruction submodule is for this Find the common intersection of the given time intervals, and the lower bound of the intersection interval. and the Upper Realm Calculate according to the following formulas respectively.

[0063] ; ; This represents the lower bound of the intersection interval, taking the maximum value of the lower bound of the effective domain of each state; It represents the upper bound of the intersection interval, taking the minimum value of the upper bound of the effective domain of each state.

[0064] like If the intersection is non-empty, it indicates that there exists at least one time window in which all necessary states are simultaneously valid. The synchronous state reconstruction submodule constructs synchronous observable states in this case and sets the intersection interval... The effective time range of this synchronously observable state is also output. If If the intersection is empty, it indicates that the necessary states do not have any overlapping intervals on the time axis, and these states cannot be established at the same time. In this case, the synchronous state reconstruction submodule generates a non-synchronous state identifier, which corresponds to the non-combinable state of the fault evidence. This group of states is prohibited from being used as joint fault evidence for subsequent fault determination.

[0065] The three-state event sequence determination submodule receives event time envelopes corresponding to multiple logs and generates three-state event sequence constraints based on the relative positions and overlapping relationships of the event time envelopes, which determine the forward sequence, the reverse sequence, and the uncertain sequence.

[0066] Specifically, let the event time envelopes of the two events be respectively and The lower and upper bounds of each are defined as follows: ; ; This represents the event time envelope of the first event. express The lower bound, express The upper boundary.

[0067] This represents the event time envelope of the second event. express The lower bound, express The upper boundary.

[0068] The three-state event sequence determination submodule makes a determination by comparing the relative positions of two intervals on the time axis, and the determination rules are as follows.

[0069] like ,Right now The upper boundary is earlier than If the lower bound of the interval is zero and the two intervals do not overlap, then the order of events is determined, indicating that the events are in a positive sequence. In the event This happened before.

[0070] like ,Right now The upper world early If the lower bound of the interval is zero and the two intervals do not overlap, then it is determined that the reverse order is determined, indicating that the events are... In the event This happened before.

[0071] If neither of the above two conditions is met, that is, if the two time intervals overlap or contain each other, then the order is determined to be uncertain, meaning that under the existing transmission delay and collection time information conditions, the order between the two events cannot be determined.

[0072] The three-state event sequence constraint is a general term for the three types of timing determination results mentioned above. It is used to determine whether the causal timing relationship between events is supported by temporal evidence when verifying a preset fault scenario. For example, a preset fault scenario requires event A to occur before event B. If the three-state event sequence determination submodule determines that events A and B are in a definite forward order, then the event sequence relationship meets the requirements of the fault scenario; if the determination result is a definite reverse order, then the event sequence relationship is opposite to the causal relationship required by the fault scenario; if the determination result is an uncertain order, then the current temporal evidence is insufficient to confirm whether the event sequence relationship meets the requirements of the fault scenario.

[0073] The outputs of the synchronous state reconstruction submodule and the three-state event sequence determination submodule are used together as input conditions for the fault evidence admission determination and are passed to the fault evidence admission and instruction generation module 34.

[0074] The fault evidence admission and instruction generation module 34 receives simultaneous observable states, non-simultaneous state identifiers, and three-state event sequence constraints, performs fault evidence admission judgment, and generates a set of classified operation and maintenance management instructions for the target equipment. The fault evidence admission and classified operation and maintenance management instruction generation process of the fault evidence admission and instruction generation module 34 includes the following steps: S34.1. Using simultaneous observable states, non-simultaneous state identifiers, and three-state event sequence constraints as pre-conditions, the system verifies whether the common intersection of the state valid domains corresponding to the necessary states of the preset fault scenario exists, and whether the event sequence relationship matches the requirements of the preset fault scenario. Specifically, the fault evidence admission and instruction generation module 34 reads the judgment rules for the current preset fault scenario from the fault rule base. The judgment rules include two parts: state synchronicity requirements and event sequence requirements, wherein: The state synchronicity requirement stipulates that all necessary states in the fault scenario must constitute a synchronously observable state, meaning the common intersection of the valid domains of all necessary states must be non-empty. The event sequence requirement stipulates that the sequential relationship between relevant log events in the fault scenario must be consistent with the pre-defined causal logic of the fault scenario. If the pre-defined fault scenario only configures the state synchronicity requirement and not the event sequence requirement, then there is no need to perform the three-state event sequence constraint verification; the synchronicity determination result is used only as the basis for fault evidence admission.

[0075] The fault evidence admission and instruction generation module 34 compares the synchronous observable state or non-synchronous state identifiers output by the synchronous state reconstruction submodule and the three-state event sequence constraints output by the three-state event sequence determination submodule with the above-mentioned determination rules item by item.

[0076] S34.2. Based on the verification results, the fault evidence is divided into three categories: confirmable, unverifiable, and uncombinable. If the verification results show that the event sequence is reversed from the preset fault scenario requirements, the fault scenario is directly determined to be invalid and is not included in the above three categories of fault evidence status. Furthermore, the rules for classifying fault evidence are as follows: If the output of the synchronous state reconstruction submodule is a synchronous observable state, and the judgment result of the three-state event sequence determination submodule on the relevant event sequence is consistent with the event sequence relationship required by the preset fault scenario, then the fault evidence is classified as verifiable, indicating that there is sufficient time evidence to support the establishment of the fault scenario under the current acquisition and transmission conditions. If the output of the synchronous state reconstruction submodule is a synchronous observable state, but the result of the three-state event sequence determination submodule on the timing of related events is a determination of reverse order, that is, the order of events is completely opposite to the causal timing required by the preset fault scenario, then the fault scenario is determined to be invalid and no corresponding fault handling instruction is generated. If the output of the synchronous state reconstruction submodule is a synchronous observable state, but the result of the three-state event sequence determination submodule for determining the timing of related events is uncertain, that is, it is impossible to confirm whether the sequential relationship between events meets the requirements of the fault scenario, then the fault evidence is classified as pending verification, indicating that the synchronous state condition has been met but the timing relationship of events still lacks definite time evidence, and further data needs to be supplemented before making a determination. If the output of the synchronous state reconstruction submodule is a non-synchronous state identifier, then regardless of the event timing determination result, the fault evidence will be classified as non-combinable, indicating that each necessary state cannot be established simultaneously in time, and this group of states does not have the basic time conditions to serve as joint fault evidence.

[0077] S34.3. For the three types of fault evidence status, generate fault handling instructions, status verification instructions, log replenishment instructions or operation suppression instructions respectively, to form a set of classified operation and maintenance management instructions for the target device.

[0078] Specifically, for verifiable fault evidence, the fault evidence admission and instruction generation module 34 generates fault handling instructions based on the handling strategies configured in the preset fault scenario. The fault handling instructions are expressed in the form of abstract operation type codes and parameters, and include specific handling operations for the fault scenario, such as adjusting the fan speed of the target device, reducing the processor operating frequency, triggering alarm notifications, or executing device restarts. The specific device-side protocol mapping is completed by the out-of-band instruction execution unit 4.

[0079] For fault evidence to be verified, the fault evidence admission and instruction generation module 34 generates a status verification instruction or a log supplementation instruction. The status verification instruction is used to request the out-of-band status acquisition unit 1 to perform supplementary acquisition of the relevant status in order to obtain more status records to narrow down the scope of the valid status domain or confirm the time node of the status change; the log supplementation instruction is used to request supplementary acquisition of relevant historical log records from the target device or log server to eliminate the uncertainty of the event timing relationship.

[0080] For incombinable fault evidence, the fault evidence admission and instruction generation module 34 generates an operation suppression instruction (this instruction is the local interception control logic of the remote operation and maintenance management processing unit and does not need to be transmitted downlink through the 5G / 4G IoT private network). The operation suppression instruction is used to prevent the issuance of high-risk control operations based on the incombinable evidence, and to prevent misoperation caused by incorrect state splicing.

[0081] The fault evidence access and instruction generation module 34 aggregates the various instructions generated above to form a categorized operation and maintenance management instruction set for the target device. Each instruction in this instruction set carries the target device identifier, instruction type code, and instruction parameters, so that the 5G / 4G IoT private network transmission unit 2 and out-of-band instruction execution unit 4 can subsequently execute the corresponding transmission scheduling and device-side operations according to the instruction type. After the categorized operation and maintenance management instruction set is generated, the fault evidence access and instruction generation module 34 outputs it to the 5G / 4G IoT private network transmission unit 2 to enter the downlink transmission process.

[0082] It should be further explained that, through the sequential processing of the above-mentioned state occurrence time envelope construction module 31, state valid domain and event envelope generation module 32, synchronous state reconstruction and event sequence determination module 33, and fault evidence admission and instruction generation module 34, the remote operation and maintenance management processing unit 3 completes the complete decision-making closed loop from the time metadata of the collection side and the transmission side to the fault evidence credibility classification determination, and then to the generation of classified management instructions. This processing link revolves around the core issue of determining the credibility of out-of-band maintenance evidence under asynchronous transmission conditions, and integrates three heterogeneous information types—collection timeliness, transmission link metadata, and state physical type—layer by layer. First, the state occurrence time envelope construction module 31 uses the latency statistics of the transmission link and the retransmission and switching states of this transmission to map the collection time of each state record into a personalized time envelope, so that the time uncertainty of each state record under the remote analysis time base can be expressed in a differentiated and quantifiable way. Then, the state effective domain and event envelope generation module 32 extends the effective duration of the time envelope differently according to the physical duration characteristics of continuous states, held states, and instantaneous events, forming a state effective domain or event time envelope that conforms to their respective physical laws. Based on this, the synchronous state reconstruction and event sequence determination module 33 uses the solution of the common intersection of the effective domains and the comparison of the relative positions of the event time envelopes to output multi-state synchronous observability conclusions and three-state event sequence constraints, respectively, to explicitly distinguish between the two cases of determining the order and the uncertain order; finally, the fault evidence admission and instruction generation module 34 uses the synchronous conditions and event sequence constraints as the pre-admission basis to divide the fault evidence into three levels: confirmable, unverifiable, and uncombinable, and generates corresponding fault handling type, state verification type, log replenishment type, and operation suppression type instructions; Therefore, without the need for global clock synchronization and without relying on the accuracy of the acquisition time, the system can make a distinguishable judgment on the correlation between heterogeneous state data and log events in the time dimension, and match differentiated control strategies to fault evidence of different trust levels.

[0083] Out-of-band instruction execution unit 4 receives the categorized operation and maintenance management instruction set output by 5G / 4G IoT private network transmission unit 2, performs device-side protocol mapping and operation permission verification on the categorized operation and maintenance management instruction set, and issues the corresponding out-of-band control operation to the target device. In this embodiment, the execution order of operation permission verification first and then device-side protocol mapping is adopted, which can intercept unauthorized and invalid instructions in advance and improve processing efficiency. There is no strict order constraint between the two. Specifically, the out-of-band instruction execution unit 4 is located at the execution end of the system. It is responsible for converting the categorized operation and maintenance management instruction set generated by the remote operation and maintenance management processing unit 3 into specific operation commands that the target device can recognize, and actually issues them through the out-of-band channel after completing security verification. This unit can be deployed on the equipment room side and can be shared with the out-of-band status acquisition unit 1 or set up independently on the same hardware platform. It establishes a communication connection with the baseboard management controller of each managed device through the out-of-band management interface. The following describes the three modules in sequence: instruction permission verification module 41, device-side protocol mapping module 42, and out-of-band control operation issuance module 43. The instruction permission verification module 41 receives the classified operation and maintenance management instruction set output by the 5G / 4G IoT private network transmission unit 2, performs operation permission verification on the classified operation and maintenance management instruction set, and outputs the classified operation and maintenance management instruction set that has passed the permission verification. Specifically, the categorized operation and maintenance management command set is transmitted via downlink encrypted transmission unit 2 of the 5G / 4G IoT private network to the command permission verification module 41. The command permission verification module 41 has a built-in operation permission policy table, which pre-configures the operation permission rules for various commands on their corresponding target devices. The operation permission rules define at least the following: the allowed command type code, the allowed target device range, the allowed parameter value range, and the command's effective time window.

[0084] The instruction permission verification module 41 performs the following verification process for each instruction in the categorized operation and maintenance management instruction set. First, it extracts the target device identifier and instruction type code from the instruction, queries the operation permission policy table, and determines whether the instruction type belongs to the permitted operation types for the current target device. If the instruction type is not in the permitted list, the instruction is directly rejected and a permission verification failure record is generated. Second, it extracts the instruction parameters from the instruction and verifies whether the parameter values ​​fall within the permitted value range specified in the operation permission policy table. Parameters exceeding the value range also trigger rejection. Finally, it verifies whether the current system time is within the time window during which the instruction is allowed to take effect. If not, execution is either postponed or rejected. Specific handling methods can be preset and configured through the operation permission policy table; high-priority instructions can be postponed until execution within the time window, while non-urgent instructions are directly rejected.

[0085] For instructions that pass all the above verification items, the instruction permission verification module 41 retains the original field content unchanged, marks it as permission verification passed, and outputs it to the device-side protocol mapping module 42. For instructions that fail verification, the instruction permission verification module 41 can generate an alarm notification, indicating that there is an unauthorized operation or abnormal instruction request, and terminate the subsequent processing flow of the instruction to prevent unauthorized operations from being sent to the target device.

[0086] The device-side protocol mapping module 42 receives the classified operation and maintenance control instruction set that has passed the permission verification, performs device-side protocol mapping processing on the classified operation and maintenance control instruction set according to the out-of-band management protocol specification of the target device, and generates control instructions that are adapted to the protocol format of the target device. Specifically, the commands in the categorized operation and maintenance management command set are expressed in the form of abstract operation type codes and parameters during generation, independent of the protocol implementation of specific devices. The device-side protocol mapping module 42 maintains a device protocol mapping table, which records the out-of-band management protocol types supported by each target device and their corresponding command mapping relationships. Common out-of-band management protocols include IPMI, Redfish, SNMP, and some vendor-specific protocols. For each command that passes the authorization verification, the device-side protocol mapping module 42 queries the device protocol mapping table based on the target device identifier to find the out-of-band management protocol type used by the device, and then looks up the corresponding protocol mapping rule based on the command type code and command parameters.

[0087] Taking the IPMI protocol as an example, a fault-handling command with the abstract operation type "set fan speed level" is mapped by the device-side protocol mapping module 42 to the corresponding byte sequence of the original equipment manufacturer extended command or standard fan control command in the IPMI specification, including network function code, command code, data fields, etc. Taking the Redfish protocol as an example, it is mapped to an HTTP PATCH or POST request to the corresponding URI on the target device management controller, with the request body containing the target value of the fan speed attribute. After the protocol mapping is completed, the generated control command adapted to the target device protocol format is ready to be sent directly through the out-of-band management interface and is output to the out-of-band control operation distribution module 43.

[0088] The out-of-band control operation sending module 43 receives control commands adapted to the target device's protocol format and sends corresponding out-of-band control operations to the target device.

[0089] Specifically, the out-of-band control operation sending module 43 establishes a secure communication session with the baseboard management controller of the target device through the out-of-band management interface. In an optional embodiment, the communication session can use an IPMI session or a TLS-based Redfish secure connection. The out-of-band control operation sending module 43 packages the control commands adapted to the target device's protocol format according to the transmission specifications of the corresponding protocol and sends them to the baseboard management controller of the target device through the communication session.

[0090] After transmission is completed, the out-of-band control operation sending module 43 receives the execution response returned by the baseboard management controller and determines whether the instruction was successfully sent based on the response status code. If a normal response is returned, the instruction is marked as executed successfully; if an abnormal response is returned or no response is received after a timeout, a limited number of retries can be performed according to the preset retry strategy. If the retries still fail, an execution failure record is generated and an alarm is triggered.

[0091] like Figure 2As shown, this embodiment also provides an out-of-band remote operation and maintenance management method for data center equipment integrating 5G / 4G IoT private networks. Based on the aforementioned out-of-band remote operation and maintenance management system for data center equipment integrating 5G / 4G IoT private networks, it includes: S1. Receive out-of-band hardware status signals, operation logs and collected metadata of each device to be maintained in the data center, perform protocol conversion and standardized encapsulation processing on the multi-source heterogeneous out-of-band raw data, and obtain a standardized out-of-band status dataset carrying collection time and field attribute identifiers. S2. Using private network slicing and multi-link redundancy scheduling technology, link optimization processing is performed on the standardized out-of-band status dataset carrying collection timeliness and field attribute identifiers. After adding transmission timeliness and link status metadata, encrypted transmission processing is performed. S3. Combine the collection time and field attribute identifier, transmission time and link status metadata to construct the status occurrence time envelope of each out-of-band status record; S4. Based on the physical type of the status field, generate the valid status field and event time envelope for each out-of-band status record. S5. Match the preset fault scenario, solve the common intersection of the state effective domains of the corresponding necessary states, generate simultaneous observable state or non-simultaneous state identifiers, and determine the three-state event order constraints of forward sequence, reverse sequence, and uncertain order. S6. Using simultaneous observable state, non-simultaneous state identifiers and three-state event sequence constraints as prerequisite admission conditions, perform fault evidence admission judgment and generate a set of classified operation and maintenance control instructions for the target equipment. S7. After performing link optimization processing on the classified operation and maintenance management instruction set using private network slicing and multi-link redundancy scheduling technology, encrypted transmission processing is performed. S8. Perform device-side protocol mapping and operation permission verification on the classified operation and maintenance management instruction set, and issue corresponding out-of-band control operations to the target device.

[0092] Those skilled in the art will understand that the process of implementing all or part of the steps of the above embodiments can be carried out by hardware or by a program instructing the relevant hardware.

[0093] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention.

Claims

1. An out-of-band remote operation and maintenance management system for data center equipment integrating 5G / 4G IoT private networks, characterized in that, include: Out-of-band status acquisition unit (1) receives the out-of-band status, operation log and collected metadata of the data center equipment to be maintained, and outputs a standardized out-of-band status dataset with collection time and field attribute identifiers after protocol conversion and standardized encapsulation. The 5G / 4G IoT private network transmission unit (2) receives a standardized dataset uplink, and after private network slicing and multi-link redundancy scheduling, it loads transmission timeliness and link status metadata and encrypts it before transmitting it to the remote operation and maintenance management processing unit (3); and receives a classified operation and maintenance management instruction set downlink, and encrypts it before transmitting it to the out-of-band instruction execution unit (4). The remote operation and maintenance management processing unit (3) receives a standardized dataset with attached transmission metadata, constructs a state occurrence time envelope by combining the collection and transmission timeliness, generates a state validity domain and event time envelope according to field attributes, completes fault access determination based on the intersection of state validity domains and three-state event sequence constraints, and generates a classified operation and maintenance management instruction set; the remote operation and maintenance management processing unit (3) includes a state occurrence time envelope construction module (31), a state validity domain and event envelope generation module (32), a synchronous state reconstruction and event sequence determination module (33), and a fault evidence access and instruction generation module (34), wherein: The state occurrence time envelope construction module (31) receives the standardized out-of-band state dataset carrying the acquisition time and field attribute identifiers, as well as the corresponding transmission time and link status metadata, generates the state occurrence time envelope of each out-of-band state record, and outputs it to the state valid field and event envelope generation module (32); including the following steps: S31.

1. Based on the transmission timeliness and link status metadata, the lower and upper bounds of the delay of the corresponding transmission link are statistically determined. The upper bound of the delay is extended by combining the retransmission status and link switching status of the current status record to obtain the extended upper bound of the delay of the corresponding status record. The original lower bound of the delay and the extended upper bound of the delay together constitute the upper and lower bounds of the delay of the status record. S31.2 Based on the upper and lower bounds of the delay recorded in this state, calculate the offset range of the source acquisition time base relative to the remote analysis time base; S31.3 Combining the collection time identifier in the standardized out-of-band state dataset carrying collection time and field attribute identifiers with the offset interval, the collection time of each out-of-band state record is mapped to a state occurrence time envelope with upper and lower time boundaries; The state valid domain and event envelope generation module (32) receives the state occurrence time envelope and the field attribute identifiers in the normalized out-of-band state dataset carrying the acquisition time and field attribute identifiers, generates the corresponding state valid domain and event time envelope, and outputs it to the synchronous state reconstruction and event sequence determination module (33); the state valid domain and event envelope generation module (32) includes a continuous state valid domain generation submodule, a sustained state valid domain generation submodule, and an instantaneous event envelope generation submodule. The three submodules process the input data in parallel according to the field attribute identifiers, wherein: The continuous state valid domain generation submodule receives the state occurrence time envelope and field attribute identifier corresponding to the continuous class state, and extends the valid retention time in combination with the update rule of the continuous state to generate the state valid domain corresponding to the continuous state. The state validity field generation submodule receives the state occurrence time envelope and field attribute identifier corresponding to the state of the maintained class, and extends the validity duration by combining the state duration rule of the maintained state to generate the state validity field corresponding to the maintained state. The instantaneous event envelope generation submodule receives the state occurrence time envelope corresponding to the instantaneous event class log and directly determines the state occurrence time envelope as the event time envelope; The synchronous state reconstruction and event sequence determination module (33) receives multiple sets of valid state domains and event time envelopes corresponding to the preset fault scenario, generates synchronous observable state or non-synchronous state identifiers, and three-state event sequence constraints; wherein, the three-state event sequence constraints refer to three types of event sequence determination results obtained by dividing the event time envelopes corresponding to multiple log events according to the relative position and overlap relationship between the event time envelopes, namely, determining the forward sequence, determining the reverse sequence, and the sequence is uncertain, which are used to verify whether the event sequence relationship required by the preset fault scenario has time evidence support; The fault evidence admission and instruction generation module (34) receives the synchronous observable state, the non-synchronous state identifier and the three-state event sequence constraint, performs fault evidence admission judgment and generates a set of classified operation and maintenance control instructions for the target equipment. Out-of-band instruction execution unit (4) receives a set of classified operation and maintenance control instructions, and after device-side protocol mapping and permission verification, issues out-of-band operations to the target device.

2. The out-of-band remote operation and maintenance management system for data center equipment integrating 5G / 4G IoT private network as described in claim 1, characterized in that, The out-of-band status acquisition unit (1) includes a data access and protocol adaptation module (11), a field attribute and timeliness labeling module (12), and a standardized encapsulation output module (13), wherein: The data access and protocol adaptation module (11) receives out-of-band hardware status signals, operation logs and collected metadata of each device to be maintained in the data center, performs protocol conversion processing on the multi-source heterogeneous out-of-band raw data, and obtains out-of-band status raw dataset with unified data format. The field attribute and timeliness labeling module (12) receives the out-of-band state raw dataset in a unified data format, extracts the field physical attributes and collection time information corresponding to each state record, and adds field attribute identifiers and collection timeliness identifiers to each out-of-band state record. The standardized encapsulation output module (13) receives the labeled out-of-band status records, performs standardized encapsulation processing on the labeled out-of-band status records, and generates a standardized out-of-band status dataset carrying the collection time and field attribute identifiers.

3. The out-of-band remote operation and maintenance management system for data center equipment integrating 5G / 4G IoT private network as described in claim 1, characterized in that, The 5G / 4G IoT private network transmission unit (2) includes a private network slicing scheduling and data diversion module (21), a transmission link metadata mounting module (22), and a two-way encrypted forwarding module (23), wherein: The private network slice scheduling and data splitting module (21) receives the standardized out-of-band status dataset with collection time and field attribute identifiers output by the out-of-band status acquisition unit (1), and the classified operation and maintenance management instruction set output by the remote operation and maintenance management processing unit (3). It matches the corresponding private network slice resources according to the business attributes of the two types of data, and completes the transmission path allocation and redundant link reservation of the two types of data respectively. The transmission link metadata mounting module (22) receives the standardized out-of-band status dataset carrying the collection time and field attribute identifier after the transmission path allocation, records the sending time, link identifier and transmission status information of the corresponding transmission link, generates transmission time and link status metadata and mounts it to the corresponding dataset. The bidirectional encrypted forwarding module (23) receives a standardized out-of-band status dataset carrying collection time and field attribute identifiers after mounting transmission time and link status metadata, and a classified operation and maintenance management instruction set after transmission path allocation. After performing encryption and encapsulation processing on the two types of data respectively, the standardized out-of-band status dataset carrying collection time and field attribute identifiers is forwarded to the remote operation and maintenance management processing unit (3), and the classified operation and maintenance management instruction set is forwarded to the out-of-band instruction execution unit (4).

4. The out-of-band remote operation and maintenance management system for data center equipment integrating 5G / 4G IoT private network as described in claim 1, characterized in that, The synchronous state reconstruction and event sequence determination module (33) includes a synchronous state reconstruction submodule and a three-state event sequence determination submodule. The two submodules process in parallel and independently and jointly output the determination result, wherein: The synchronous state reconstruction submodule receives the state effective domains of multiple states corresponding to a preset fault scenario and solves the common intersection of the state effective domains; when the intersection is not empty, it constructs a synchronous observable state; when the intersection is empty, it generates a non-synchronous state identifier, which corresponds to the uncombinable state of the fault evidence, and prohibits the combination of the corresponding states into joint fault evidence. The three-state event sequence determination submodule receives event time envelopes corresponding to multiple logs and generates three-state event sequence constraints based on the relative positions and overlapping relationships of the event time envelopes, which determine the forward sequence, the reverse sequence, and the uncertain sequence.

5. The out-of-band remote operation and maintenance management system for data center equipment integrating 5G / 4G IoT private network as described in claim 1, characterized in that, The fault evidence admission and classification operation and maintenance management instruction generation process of the fault evidence admission and instruction generation module (34) includes the following steps: S34.1 Using the simultaneous observable state, non-simultaneous state identifiers and three-state event sequence constraints as pre-entry conditions, verify whether the common intersection of the state effective domains of the corresponding states in the preset fault scenario exists, and whether the event sequence relationship matches the preset fault scenario requirements. S34.2 Based on the verification results, the fault evidence is divided into three categories: confirmable, pending verification, and uncombinable. If the verification shows that the event sequence is reversed from the preset fault scenario requirements, then the fault scenario is directly determined to be invalid and is not included in the classification of the three types of fault evidence states. S34.

3. For the three types of fault evidence status, generate fault handling instructions, status verification instructions, log replenishment instructions or operation suppression instructions respectively, to form a set of classified operation and maintenance management instructions for the target device.

6. The out-of-band remote operation and maintenance management system for data center equipment integrating 5G / 4G IoT private network as described in claim 1, characterized in that, The out-of-band instruction execution unit (4) includes an instruction permission verification module (41), a device-side protocol mapping module (42), and an out-of-band control operation distribution module (43), wherein: The instruction permission verification module (41) receives the classified operation and maintenance management instruction set output by the 5G / 4G IoT private network transmission unit (2), performs operation permission verification on the classified operation and maintenance management instruction set, and outputs the classified operation and maintenance management instruction set that has passed the permission verification. The device-side protocol mapping module (42) receives the classified operation and maintenance control instruction set that has passed the permission verification, performs device-side protocol mapping processing on the classified operation and maintenance control instruction set according to the out-of-band management protocol specification of the target device, and generates control instructions that are adapted to the protocol format of the target device. The out-of-band control operation sending module (43) receives control instructions adapted to the target device protocol format and sends corresponding out-of-band control operations to the target device.

7. A method for out-of-band remote operation and maintenance management of data center equipment integrating 5G / 4G IoT private networks, based on the out-of-band remote operation and maintenance management system for data center equipment integrating 5G / 4G IoT private networks as described in any one of claims 1-6, characterized in that, Includes the following steps: S1. Receive out-of-band hardware status signals, operation logs and collected metadata of each device to be maintained in the data center, perform protocol conversion and standardized encapsulation processing on the multi-source heterogeneous out-of-band raw data, and obtain a standardized out-of-band status dataset carrying collection time and field attribute identifiers. S2. Using private network slicing and multi-link redundancy scheduling technology, link filtering processing is performed on the standardized out-of-band status dataset carrying collection timeliness and field attribute identifiers. After adding transmission timeliness and link status metadata, encrypted transmission processing is performed. S3. Combine the collection time and field attribute identifier, transmission time and link status metadata to construct the status occurrence time envelope of each out-of-band status record; S4. Based on the physical type of the status field, generate the valid state field and event time envelope for each out-of-band status record. S5. Match the preset fault scenario, solve the common intersection of the state effective domains of the corresponding states, generate simultaneous observable state or non-simultaneous state identifiers, and determine the three-state event order constraints of forward sequence, reverse sequence, and uncertain order. S6. Using simultaneous observable state, non-simultaneous state identifiers and three-state event sequence constraints as prerequisite admission conditions, perform fault evidence admission judgment and generate a set of classified operation and maintenance control instructions for the target equipment. S7. After performing link filtering processing on the classified operation and maintenance management instruction set using private network slicing and multi-link redundancy scheduling technology, encrypted transmission processing is performed. S8. Perform device-side protocol mapping and operation permission verification on the classified operation and maintenance management instruction set, and issue corresponding out-of-band control operations to the target device.

Images