Fine-grained multi-level management system for audio frequency integrated signal processing

Abstract

The invention discloses a fine-grained multi-level management system for audio frequency integrated signal processing, which belongs to the technical field of audio frequency signal processors and comprises a computing board card management module, a node priority dynamic computing module, a case management control module and a fault self-healing resource scheduling module. Resource allocation is optimized in real time through a dynamic priority scoring mechanism, an abnormal source is accurately positioned by combining a graph neural network analysis node dependency graph, and millisecond-level fault recovery is realized by adopting data migration, so that the static property and hysteresis of a traditional technology are broken through, and the self-adaptive capability and reliability of a system are remarkably improved; through a multi-level and fine-grained management architecture, technical breakthroughs are realized in the aspects of real-time performance, reliability, self-healing capability and the like, and an efficient solution is provided for complex application of an audio signal processing system in the fields of 5G communication, intelligent Internet of Things and the like.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of audio signal processing machines, and particularly relates to an audio integrated signal processing fine-grained multi-level management system. BACKGROUND

[0002] With the rapid popularization of 5G communication, Internet of Things and intelligent terminals, the importance of audio signal processing systems in the fields of multimedia transmission, real-time communication, intelligent voice interaction and the like is increasingly prominent. Efficient and stable audio processing capability has become a core requirement to support the operation of modern digital society, especially in the scenarios of autonomous driving, remote medical treatment, virtual reality and the like, the real-time performance, reliability and resource management efficiency of the system directly affect the user experience and safety. However, the traditional audio signal management system has gradually exposed its technical limitations when facing high-concurrency tasks, dynamic load fluctuations and complex fault scenarios, and it is urgent to break through the existing bottlenecks through technical innovation, and the traditional audio signal management system cannot comprehensively and real-timely monitor the state of each module in the signal processing machine.

[0003] The current mainstream audio signal management system mostly adopts a static resource allocation strategy and a threshold-based abnormality monitoring mechanism. For example, resource allocation relies on preset rules such as fixed priority or round-robin scheduling, which cannot dynamically adjust the priority, resulting in high-load node task backlog and low-priority task response delay. In terms of abnormality detection, the existing technology usually triggers an alarm based on a single parameter threshold, lacks comprehensive analysis of multi-dimensional indicators (temperature, power, error rate), and is prone to misjudgment or omission. In addition, abnormality tracing relies on human experience, and it is difficult to construct a dynamic node dependency graph, resulting in a fuzzy fault propagation path and low source localization efficiency. In the fault recovery link, service interruption or data loss may occur during the migration process, and there is a lack of quantitative evaluation of node health, the recovery strategy is single, and fast self-healing cannot be achieved.

[0004] In view of the above problems, the industry urgently needs an audio signal management system that can dynamically optimize resource allocation, accurately locate abnormal sources and achieve intelligent fault self-healing. SUMMARY

[0005] Therefore, the purpose of the present application is to provide an audio integrated signal processing fine-grained multi-level management system, which integrates a dynamic priority scoring mechanism of real-time load and task urgency, an abnormality propagation path analysis technology based on a graph neural network, and a fault-tolerant protocol supporting data state freezing and rapid migration, to break through the static and lagging nature of traditional technology and significantly improve the adaptability and reliability of the system.

[0006] To achieve the above purpose, the present application provides the following technical solutions: An audio integrated signal processing fine-grained multi-level management system for managing an audio signal processing machine, comprising: The computing board management module is configured to manage sensor information of each computing board and overall information of the computing board; real-time collection of relevant data of each node is performed by deploying corresponding sensors on each node to be monitored, and information is uploaded to the case management control module through an IPMB protocol; The node priority dynamic calculation module is configured to dynamically calculate a priority score of each node according to a real-time load rate and a task urgency of each node; The case management control module is configured to process and analyze information uploaded by the computing board management module, construct a dependency graph between nodes after an abnormal node is found, analyze an abnormal propagation path of the abnormal node through a graph neural network, find a source abnormal node, and report abnormal information of the source abnormal node and corresponding alarm information to a user upper operation and maintenance platform through a network. The fault self-healing resource scheduling module is configured to perform fault state analysis on the source abnormal node, perform data migration according to a node priority when multiple source abnormal nodes are determined as hard faults, freeze a current data state and compress storage, and distribute a task copy to a node with a lowest priority score.

[0007] As a further preferred embodiment of the present application, the node priority dynamic calculation module performs the following steps: A load rate of each node is calculated based on a CPU usage rate and a memory occupancy rate, and standardized processing is performed; A task urgency of each node is calculated based on a remaining time of each task participated by the node, and standardized processing is performed; A priority score of the node is dynamically calculated by weighted summation of the real-time load rate and the task urgency of the node.

[0008] As a further preferred embodiment of the present application, the case management control module performs the following steps: A parameter threshold value of the corresponding node at this time is predicted based on historical data of each node, and if it is found that a parameter of a node exceeds the corresponding predicted threshold value, the node is determined as an abnormal node; A dependency graph between nodes is constructed based on a dependency relationship between the abnormal node and other nodes, wherein the dependency relationship is determined through task scheduling, a communication relationship or a data flow; An abnormal propagation path of the abnormal node is analyzed through a graph neural network, a source abnormal node is identified according to the abnormal propagation path, abnormal information of the source abnormal node is recorded, including an abnormal time, an abnormal type and a severity, corresponding alarm information is generated, and the abnormal information and the alarm information of the source abnormal node are reported to the user upper operation and maintenance platform through a network.

[0009] As a further preferred embodiment of the present application, the fault self-healing resource scheduling module performs the following steps: The monitoring parameters of each source abnormal node are collected by corresponding sensors, and the monitoring parameters include temperature, power and error rate, wherein the error rate is the number of CRC check failures in a unit time; Based on the monitoring parameters of each source abnormal node collected by the corresponding sensors, the health degree of each source abnormal node is calculated, and the node health degree is defined as The calculation formula is as follows: Among them, is the temperature weight, is the power weight, is the error rate weight, and ; is the temperature of the source abnormal node, is the maximum temperature threshold of the node, is the minimum temperature threshold of the node, is the power of the source abnormal node, is the maximum power threshold of the node, is the number of CRC check failures in a unit time of the source abnormal node, is the total number of CRC checks in a unit time; Based on the health degree of each source abnormal node, the fault type is determined, and the fault type includes a recoverable soft fault and an unrecoverable hard fault, wherein when and lasts less than 100ms, it is determined as a soft fault, and the corresponding early warning information is triggered, and the early warning information is sent to the user upper operation and maintenance platform through the chassis management control module; when or the node is unresponsive, it is determined as a hard fault, data migration is performed according to the priority order of each source abnormal node, the current data state is frozen and compressed storage, and the task copy is distributed to the node with the lowest priority score for transmission.

[0010] As a further preferred embodiment of the present application, the health degree trend in the future time is predicted based on the LSTM model, and if the prediction , the task migration is started in advance.

[0011] As a further preferred embodiment of the present application, the computing board management module adopts an in-band-out-of-band software and hardware integrated management technology based on BMC-CPU, including in-band management and out-of-band management, wherein the in-band management is that the computing board management module communicates with the chassis management control module through the IPMB protocol, and reports sensor information and overall information of the computing board; the out-of-band management is that the computing board management module communicates with the CPU through the LPC bus, obtains system information and BIOS information, and reports to the chassis management control module.

[0012] As a further preferred embodiment of the present application, based on the system state, the chassis management control module dynamically adjusts the parameters of each board card, including gain and bandwidth, and dynamically adjusts and allocates resources according to the load of each board card, as follows: wherein, is the amount of resources allocated to the th board card, is the load of the th board card, is the total number of board cards, is the total amount of resources.

[0013] As a further preferred embodiment of the present application, the computing board card management module uses a master-slave dual module management technology based on the IPMB protocol to manage the entire computing board card, a redundant communication channel is designed between the chassis management control module and the computing board card management module, and bidirectional dynamic switching is supported.

[0014] The present application has the following advantages: The audio integrated signal processing fine-grained multi-level management system provided by the present application solves the core defects of traditional audio signal management systems in dynamic resource allocation, abnormal source tracing and fault self-healing through innovative technical means. Through the node priority dynamic calculation module, based on the weighted scoring mechanism of real-time load rate (CPU, memory) and task urgency (remaining time), dynamic adjustment of resource allocation is realized. The graph neural network analyzes the node dependency graph, and combined with the comprehensive health degree of multi-dimensional parameters (temperature, power, error rate), the abnormal propagation path and source node can be accurately identified. At the same time, based on the health trend prediction technology of the LSTM model, task migration can be started in advance to reduce the risk of system downtime. Through data migration, the fault node is quickly recovered. In the hard fault scenario, the data state is frozen and compressed and then migrated to a low-priority node, and the task interruption time is reduced from seconds to milliseconds in traditional cold migration. The present application realizes technical breakthroughs in real-time, reliability and self-healing capability through a multi-level, fine-grained management architecture, providing an efficient solution for complex applications of audio signal processing systems in 5G communication, intelligent Internet of Things and other fields.

[0015] Other advantages, objects, and features of the present application will become apparent to those skilled in the art from the following specification, which is to be taken in conjunction with the accompanying drawings, wherein: BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to make the objects, technical solutions and beneficial effects of the present application clearer, the present application provides the following drawings for illustration: Figure 1 This is a schematic diagram of the overall modules of a fine-grained multi-level management system for audio integrated signal processing according to the present invention; Figure 2 Schematic diagram of the execution of the node priority dynamic calculation module of the present invention; Figure 3 Schematic diagram of the execution of the chassis management control module of the present invention; Figure 4 Schematic diagram of the execution of the fault self-healing resource scheduling module of the present invention. DETAILED DESCRIPTION

[0017] like Figures 1-4 As shown, the present invention discloses a fine-grained multi-level management system for audio integrated signal processing, which is used to manage audio signal processors, including: The chassis management module (BMC) manages the sensor information of each computing board and its overall information. By deploying sensors on each node that needs to be monitored, it collects node data in real time and uploads this information to the chassis management module via the IPMB protocol, enabling comprehensive monitoring of the entire system and ensuring stable system operation.

[0018] The computing board management module adopts in-band and out-of-band integrated hardware and software management technology based on BMC-CPU, including in-band management and out-of-band management. In-band management is: the computing board management module communicates with the chassis management control module through the IPMB protocol to report sensor information and overall information of the computing board; out-of-band management is: the computing board management module communicates with the CPU through the LPC bus to obtain system information and BIOS information, and reports it to the chassis management control module.

[0019] The use of in-band and out-of-band integrated hardware and software management technology based on BMC-CPU enables the computing board management module to obtain various information of the computing board in a more comprehensive and in-depth manner. It not only realizes stable communication and information reporting with the chassis management control module through the IPMB protocol, but also communicates with the CPU through the LPC bus to obtain system and BIOS information, providing richer data support for the chassis management control module, ensuring refined management of the computing board, improving management reliability and accuracy, and enhancing the overall performance and stability of the audio signal processor management system.

[0020] In order to ensure that the user upper layer operation and maintenance platform can timely and accurately monitor the information and running state of the computing board card, the computing board card management system adopts a master-slave dual module management technology based on IPMB protocol to manage the entire computing board card. The CPU computing board card management system can actively report the alarm information generated in the board card to the chassis management control module, or the ChMC management system can actively obtain various monitoring information from the computing board card management system.

[0021] The computing board card management system provides a dual-channel IPMB protocol for the chassis management control module. The computing board card management module designs a redundant communication channel between the chassis management control module and the computing board card management module, and supports bidirectional dynamic switching. The chassis management control module can obtain the required information through the above protocol by actively or passively. Through the above, the sensor information of the computing board card and the board card and module information are obtained, forming a redundant design. It can prevent the entire management system from being unable to communicate normally with the computing board card management system when one IPMB channel fails to communicate.

[0022] The master-slave dual module management technology based on IPMB protocol is adopted, and a redundant communication channel is designed and bidirectional dynamic switching is supported, which greatly enhances the reliability and stability of the computing board card management module. Even in the case of failure of a communication channel or management module, the system can automatically switch to the backup channel or module, ensuring uninterrupted management, improving the availability and reliability of the audio signal processor, and providing a strong guarantee for the stable operation of the system.

[0023] Node priority dynamic calculation module: used for dynamically calculating the priority score of each node according to the real-time load rate and task urgency of each node. This module dynamically calculates the priority score of each node according to the real-time load rate and task urgency of each node, which provides an important basis for subsequent task allocation and resource scheduling.

[0024] The node priority dynamic calculation module performs the following steps: Calculate the load rate of each node based on CPU usage and memory occupancy, and perform standardization processing. First, collect these two parameters, and then perform standardization processing to eliminate the influence of dimension difference and data magnitude, so that different indicators have comparability.

[0025] Calculate the task urgency of each node based on the remaining time of each task participated by the node, and perform standardization processing.

[0026] Weighted sum the real-time load rate and task urgency of each node to dynamically calculate the priority score of the node. The weight coefficient can be flexibly set according to the actual situation to reflect the relative importance of different factors in priority evaluation.

[0027] The node priority dynamic calculation module comprehensively considers the real-time load rate and task urgency of the node, dynamically calculates the node priority score through scientific standardized processing and weighted summation method, so that the task allocation and resource scheduling are more reasonable and accurate, the task demand of different nodes in different running states can be fully met, the overall performance and efficiency of the audio signal processor are improved, and the key task is ensured to be processed and resource guaranteed in priority.

[0028] The chassis management control module ChMC is used for processing and analyzing the information uploaded by the computing board management module, constructing a dependency graph between nodes after discovering abnormal nodes, analyzing the abnormal propagation path of the abnormal nodes through a graph neural network, finding the source abnormal node, and uploading the abnormal information of the source abnormal node and the corresponding alarm information to the user's upper operation and maintenance platform through the network.

[0029] The chassis management control module can be further divided into multiple sub-function modules, which are independent and interconnected, and maintain and control the normal operation of the entire system. The chassis management control module can independently display the reported information of each board card, and can also report to the user's upper operation and maintenance platform through the network. When the chassis management control module processes the reported information of each board card management subsystem and finds alarm information, it is reported to the user's upper operation and maintenance platform through the network.

[0030] The chassis management control module executes the following steps: Based on the historical data of each node, the parameter threshold of the corresponding node at this time is predicted, and if it is found that the parameter of the node exceeds the corresponding predicted threshold, the node is determined to be an abnormal node. Specifically, first, based on the historical data of each node, the parameter threshold of the corresponding node at this time is predicted using data mining and machine learning algorithms, and if it is found that the parameter of the node exceeds the corresponding predicted threshold, the node is determined to be an abnormal node.

[0031] Based on the dependency relationship between the abnormal node and other nodes, a dependency graph between nodes is constructed, wherein the dependency relationship is determined through task scheduling, communication relationship or data flow. For example, if there is frequent data interaction or joint participation in a task between two nodes, it is considered that there is a dependency relationship between them, and the corresponding representation is made in the dependency graph. The dependency graph between nodes can also be constructed according to the task scheduling log (such as the output of node A as the input of node B), and the dependency relationship is represented by a directed edge.

[0032] The abnormal propagation path of the abnormal node is analyzed through the graph neural network, and the source abnormal node is identified based on the abnormal propagation path. The abnormal information of the source abnormal node is recorded, including the time, type and severity of the abnormality, and the corresponding alarm information is generated. The abnormal information and alarm information of the source abnormal node are reported to the user's upper-level operation and maintenance platform through the network. Specifically: The dependency graph is fed into a graph neural network (GNN), and the anomaly propagation path is analyzed through a message-passing mechanism. For example, if node B is abnormal and the parameters of its dependent node A fluctuate abnormally, node A is determined to be the source of the anomaly. The traceability results include the anomaly type (such as temperature overload), timestamp, and propagation path. Alarm information is generated and pushed to the operation and maintenance platform, allowing operators to promptly understand the equipment's operating status and take appropriate measures.

[0033] The chassis management and control module predicts parameter thresholds based on historical data, enabling timely detection of abnormal nodes. By constructing a dependency graph between nodes and analyzing the abnormal propagation path using a graph neural network, it achieves rapid and accurate identification of the source abnormal node. This provides a key basis for subsequent fault handling and system optimization, helping to take proactive measures to prevent the further spread and expansion of the fault, thereby enhancing the robustness and maintainability of the system. Detailed information about abnormal source nodes is recorded, including the time, type, and severity of the abnormality. Corresponding alarm information is generated and reported to the user's upper-level operation and maintenance platform, enabling operation and maintenance personnel to immediately understand the equipment's fault status and respond and handle it promptly. This improves the timeliness and effectiveness of operation and maintenance, and ensures the reliable operation of the audio signal processing system.

[0034] Fault self-healing resource scheduling module: used to analyze the fault status of abnormal source nodes. When multiple abnormal source nodes are determined to be hard faults, data migration is performed according to node priority, the current data status is frozen and compressed for storage, and task copies are distributed to the node with the lowest current priority score.

[0035] The fault self-healing resource scheduling module performs the following steps: The corresponding sensors collect monitoring parameters of each source abnormal node, including temperature, power, and error rate, where the error rate is the number of CRC check failures per unit time; Based on the monitoring parameters of each source abnormal node collected by the corresponding sensor, the health of each source abnormal node is calculated and the node health is defined , the calculation formula is as follows: in, is the temperature weight, is the power weight, is the error rate weight, and ; temperature of the source abnormal node, maximum temperature threshold of the node, minimum temperature threshold of the node, power of the source abnormal node, maximum power threshold of the node, number of CRC check failures per unit time of the source abnormal node, total number of CRC checks per unit time; based on the health degree of each source abnormal node, the fault type is determined, and the fault type includes a recoverable soft fault and an unrecoverable hard fault, wherein when and lasts less than 100 ms, the soft fault is determined, and the corresponding early warning information is triggered, and the early warning information is sent to the user upper operation and maintenance platform through the case management control module; when or the node is unresponsive, the hard fault is determined, data migration is performed according to the priority order of each source abnormal node, in the hard fault scenario, the data calculation state of the current node is frozen, compressed into an encrypted data packet, and the task copy is distributed to the available node with the lowest priority score through the redundant channel, to realize rapid recovery of the task and reasonable allocation of resources.

[0036] The fault self-healing resource scheduling module can accurately determine the fault type as a recoverable soft fault or an unrecoverable hard fault by collecting various monitoring parameters of the source abnormal node and calculating the health degree, combining the change of the health degree and the node response state, providing an accurate basis for adopting different fault handling strategies, avoiding unnecessary resource waste and business impact, and improving the fault handling efficiency and accuracy of the system. For soft faults, early warning information is triggered in time to remind the operation and maintenance personnel to pay attention; for hard faults, data migration is performed according to the priority order of the node, rapid recovery of the task and reasonable allocation of resources are realized, the continuity of the audio signal processing business is ensured, the impact of the fault on the system performance and business quality is reduced, and the self-healing ability and reliability of the system are enhanced.

[0037] The health degree trend in the future time can also be predicted based on the LSTM model, and if the predicted health degree trend is abnormal, task migration is started in advance. Predicting the health degree trend in the future time based on the LSTM model can predict the possible fault risk of the node in advance, start task migration in advance before the fault occurs, avoid task interruption and data loss caused by the fault, further improve the reliability and stability of the system, provide more continuous and stable service guarantee for the audio signal processing business, and optimize the overall operation efficiency and quality of the system.

[0038] Based on the system state, the chassis management control module dynamically adjusts the parameters of each board card, including gain and bandwidth, and dynamically adjusts and allocates resources according to the load of each board card, as follows: wherein, is the resource amount allocated to the th board card, is the load of the th board card, is the total number of board cards, is the total resource amount.

[0039] According to the system state, the parameters of each board card, such as gain and bandwidth, are dynamically adjusted, and resources are dynamically allocated according to the load of each board card according to a certain formula, realizing reasonable allocation and efficient use of resources, ensuring that each board card can obtain optimal resource support under the current load condition, improving the overall performance and efficiency of the audio signal processor, avoiding performance bottlenecks caused by waste and unreasonable allocation of resources, and ensuring high-quality operation of audio signal processing business.

[0040] The present application constructs a complete fine-grained multi-level management system for audio integrated signal processing, covering multiple key modules from computing board card management, node priority dynamic calculation, abnormal analysis to fault self-healing resource scheduling, realizing all-round and fine-grained management of the audio signal processor, effectively improving the management efficiency and reliability of the equipment, and ensuring the stability and continuity of the audio signal processing process. Through the in-depth analysis of the abnormal node by the chassis management control module, the fault source can be quickly and accurately found, and the task can be quickly migrated and restored with the help of the fault self-healing resource scheduling module, thereby minimizing the business interruption time caused by faults, improving the availability and stability of the system, and reducing the operation and maintenance cost and business risk.

[0041] Finally, it should be pointed out that the above preferred embodiments are only used to illustrate the technical solutions of the present application and not to limit, although the present application has been described in detail through the above preferred embodiments, those skilled in the art should understand that various changes can be made in form and details without departing from the scope defined by the claims of the present application.

Claims

1. A fine-grained multi-level management system for audio integrated signal processing, for managing audio signal processors, characterized by: include: Computing board management module: used to manage the sensor information of each computing board and the overall information of the computing board; By deploying corresponding sensors on each node that needs to be monitored, relevant data of the node is collected in real time and uploaded to the chassis management control module through the IPMB protocol; Node priority dynamic calculation module: used to dynamically calculate the priority score of each node based on the real-time load rate and task urgency of each node; Chassis management control module: This module processes and analyzes the information uploaded by the computing board management module. After discovering abnormal nodes, it constructs a dependency graph between the nodes. It uses a graph neural network to analyze the abnormal propagation path of the abnormal nodes, finds the source abnormal node, and reports the abnormal information and corresponding alarm information of the source abnormal node to the user's upper-level operation and maintenance platform through the network. Fault self-healing resource scheduling module: used to analyze the fault status of abnormal source nodes. When multiple abnormal source nodes are determined to be hard faults, data migration is performed according to node priority, the current data status is frozen and compressed for storage, and task copies are distributed to the node with the lowest current priority score.

2. The fine-grained multi-level management system for audio integrated signal processing according to claim 1, characterized in that: The node priority dynamic calculation module performs the following steps: Calculate the current load rate of each node based on CPU usage and memory occupancy, and perform normalization; Based on the remaining time of each task in which the node participates, the task urgency of each node is calculated and standardized; The real-time load rate and task urgency of each node are weighted and summed to dynamically calculate the priority score of the node.

3. The fine-grained multi-level management system for audio integrated signal processing according to claim 1, characterized in that: The chassis management control module performs the following steps: Based on the historical data of each node, the parameter threshold of the corresponding node is predicted. If the parameter of a node exceeds the corresponding prediction threshold, the node is determined to be an abnormal node. Based on the dependency relationship between the abnormal node and other nodes, a dependency graph between the nodes is constructed, wherein the dependency relationship is determined by task scheduling, communication relationship or data flow; The abnormal propagation path of the abnormal node is analyzed through the graph neural network, and the source abnormal node is identified based on the abnormal propagation path. The abnormal information of the source abnormal node is recorded, including the time, type and severity of the abnormality, and the corresponding alarm information is generated. The abnormal information and alarm information of the source abnormal node are reported to the user's upper-level operation and maintenance platform through the network.

4. The fine-grained multi-level management system for audio integrated signal processing according to claim 1, characterized in that: The fault self-healing resource scheduling module performs the following steps: The corresponding sensors collect monitoring parameters of each source abnormal node, including temperature, power, and error rate, where the error rate is the number of CRC check failures per unit time; Based on the monitoring parameters of each source abnormal node collected by the corresponding sensor, the health of each source abnormal node is calculated and the node health is defined , the calculation formula is as follows: in, is the temperature weight, is the power weight, is the error rate weight, and ; is the temperature of the source abnormal node, is the maximum temperature threshold of the node, is the minimum temperature threshold of the node, is the power of the source abnormal node, is the maximum power threshold of the node, is the number of CRC check failures per unit time of the source abnormal node, The total number of CRC checks per unit time; Based on the health of each source abnormal node, the fault type is determined. The fault type includes recoverable soft faults and unrecoverable hard faults. If the delay is less than 100ms, it is determined to be a soft fault and the corresponding warning information is triggered. The warning information is sent to the user's upper-level operation and maintenance platform through the chassis management control module. Or when a node is unresponsive, it is determined to be a hard failure. Data migration is performed according to the priority order of each source abnormal node, the current data state is frozen and compressed for storage, and the task copy is distributed to the node with the lowest current priority score for transmission.

5. The fine-grained multi-level management system for audio integrated signal processing according to claim 4, characterized in that: Predicting the future based on LSTM model Health trend over time, if predicted , then start task migration in advance.

6. The fine-grained multi-level management system for audio integrated signal processing according to claim 1, characterized in that: The computing board management module adopts in-band and out-of-band integrated hardware and software management technology based on BMC-CPU, including in-band management and out-of-band management. In-band management is: the computing board management module communicates with the chassis management control module through the IPMB protocol to report sensor information and overall information of the computing board; out-of-band management is: the computing board management module communicates with the CPU through the LPC bus to obtain system information and BIOS information, and reports it to the chassis management control module.

7. The fine-grained multi-level management system for audio integrated signal processing according to claim 1, characterized in that: Based on the system status, the chassis management control module dynamically adjusts the parameters of each board, including gain and bandwidth, and dynamically adjusts the allocation of resources according to the load of each board. The formula is as follows: in, is assigned to The amount of resources per board, It is The load of the board, is the total number of boards, is the total resource amount.

8. The fine-grained multi-level management system for audio integrated signal processing according to claim 1, characterized in that: The computing board management module uses active and passive dual-module management technology based on the IPMB protocol to manage the entire computing board. A redundant communication channel is designed between the chassis management control module and the computing board management module, and supports two-way dynamic switching.

Images