Computing Power Infrastructure Monitoring Method, System and Computer Readable Storage Medium

Abstract

The present invention belongs to the technical field of intelligent operation and maintenance, and relates to a monitoring method, system and computer-readable storage medium for computing power infrastructure. The monitoring method is as follows: bottom-layer data is collected in real time and standardized to obtain standardized data; the standardized data is input into the spatio-temporal feature extraction layer, where the time features are captured by a bidirectional long short-term memory neural network, the spatial features are constructed by a graph attention network, and the time features and spatial features are fused based on a gated attention mechanism to obtain spatio-temporal features; the spatio-temporal features are input into the hybrid density memory layer, a memory matrix is designed, the density function of the Gaussian mixture model is calculated for the memory matrix to obtain the mixture density probability, and the memory matrix is dynamically allocated to the memory slots according to the probability; the memory matrix in the memory slots is input into the variational inference calculation layer, the latent variable distribution of the memory matrix in the memory slots is inferred by the encoder, and the abnormal sequence is reconstructed and output by the decoder; the present invention effectively improves the timeliness, accuracy and effectiveness of computing power monitoring.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent operation and maintenance, and particularly relates to a monitoring method, system and computer-readable storage medium for computing power infrastructure. Background Art

[0002] In the context of the current high-speed development of the digital economy, computing power security has become an important link to ensure the stable operation of enterprises and society. Computing power security monitoring not only involves the maintenance of data confidentiality and integrity, but also covers the effective management and use of computing resources. Based on this, how to conduct intelligent monitoring of computing power infrastructure has become a difficult problem that needs to be solved urgently. Summary of the Invention

[0003] Based on the above deficiencies in the prior art, the purpose of the present invention is to provide a monitoring method, system and computer-readable storage medium for computing power infrastructure.

[0004] In order to achieve the above invention purpose, the present invention adopts the following technical solutions:

[0005] A monitoring method for computing power infrastructure, comprising the following steps:

[0006] S1. Real-time collect underlying data; wherein, the underlying data includes physical infrastructure data, virtualized resource data and business application data;

[0007] S2. Perform standardization processing on the underlying data to obtain standardized data;

[0008] S3. Input the standardized data into a spatio-temporal memory network based on mixed density for abnormal fault detection;

[0009] Among them, the spatio-temporal memory network based on mixed density includes: a spatio-temporal feature extraction layer, a mixed density memory layer and a variational inference calculation layer;

[0010] The standardized data is input into the spatio-temporal feature extraction layer. The time features are captured through a bidirectional long short-term memory neural network, the spatial features are constructed through a graph attention network, and feature fusion of the time features and the spatial features is performed based on a gated attention mechanism to obtain spatio-temporal features; the spatio-temporal features are input into the mixed density memory layer, a memory matrix is designed, the density function of the Gaussian mixture model is calculated for the memory matrix to obtain the mixed density probability, and the memory matrix is dynamically allocated to several memory slots according to the mixed density probability. The memory slots carry time tags; the memory matrix in the memory slot is input into the variational inference calculation layer, the latent variable distribution of the memory matrix in the memory slot is inferred through an encoder, and an abnormal sequence is reconstructed and output through a decoder to achieve abnormal fault detection.

[0011] As a preferred solution, the monitoring method for computing power infrastructure further includes the following steps:

[0012] S4. Use a knowledge-enhanced graph attention network to trace the root cause of abnormal faults and locate the source of the faults.

[0013] As an optimal solution, the knowledge-enhanced graph attention network includes a knowledge fusion layer, a relational attention mechanism layer, and a causal reasoning layer;

[0014] The abnormal sequence is input into the knowledge fusion layer to construct a knowledge graph joint embedding for knowledge fusion, obtaining a knowledge graph; the knowledge graph is input into the relational attention mechanism layer, and heterogeneous feature aggregation is performed through relational attention coefficients to obtain a graph feature matrix; the graph feature matrix is input into the causal reasoning layer, and root cause determination is performed according to the influence propagation model, outputting abnormal graph nodes to locate the source of the faults.

[0015] As an optimal solution, the method for constructing a knowledge graph joint embedding for knowledge fusion is as follows:

[0016] ;

[0017] Among them, is the initial embedding vector of the i-th node, is the weight matrix, is the monitoring index obtained by hybrid cross-computation of the real-time underlying data of the i-th node, is the device type code of the i-th node, is the embedding vector of the knowledge graph of the i-th node.

[0018] As an optimal solution, the computing power infrastructure monitoring method further includes the following steps:

[0019] S5. Input the standardized data into a long short-term memory neural network based on a self-attention mechanism to predict the busy and idle state of resources, and realize the scheduling and optimization of resources.

[0020] As an optimal solution, the computing power infrastructure monitoring method further includes the following steps:

[0021] S6. Issue real-time alarms for faults that cannot be autonomously repaired or exceed the disaster tolerance threshold, and provide fault handling suggestions and fault risk scores.

[0022] As an optimal solution, the physical infrastructure data includes chip temperature, the overall power consumption of the server, fan speed, the current value of the cabinet, and sensor temperature;

[0023] The virtualized resource data includes the resource utilization rate related to containers, the allocation rate of resources of open-source platform K8S nodes, and the packet loss rate related to the network;

[0024] The business application data includes the transaction processing volume, the response delay of the application programming interface API, and the storage capacity.

[0025] As a preferred solution, in step S2, the standardization process includes unified timestamp, unit conversion and feature construction.

[0026] The present invention further provides a computing power infrastructure monitoring system, applying the computing power infrastructure monitoring method described in any of the above solutions, the computing power infrastructure monitoring system comprising:

[0027] Acquisition module, used to collect underlying data in real time;

[0028] The standardization module is used to standardize the underlying data to obtain standardized data;

[0029] The anomaly detection module is used to input standardized data into the spatiotemporal memory network based on mixed density for abnormal fault detection.

[0030] The present invention also provides a computer-readable storage medium, which stores instructions. When the instructions are executed on a computer, the computer executes the computing power infrastructure monitoring method as described in any of the above items.

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] (1) The computing power infrastructure monitoring method of the present invention performs abnormal fault detection based on a mixed density spatiotemporal memory network, comprehensively considers and analyzes from the two dimensions of time and space, and improves accuracy;

[0033] (2) This invention uses a knowledge-enhanced graph attention network to trace the root cause of the fault. Based on the graph structure, it integrates the internal association relationships and uses the form of a graph (node-edge-node) to perform fundamental positioning and tracing. The graph representation of association relationships is more effective.

[0034] (3) The present invention predicts the busy and idle status of resources in the algorithm network of the long short-term memory neural network based on the self-attention mechanism to achieve resource scheduling and optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 is a flow chart of a computing power infrastructure monitoring method according to an embodiment of the present invention;

[0036] Figure 2 is a processing flow chart of a hybrid density spatiotemporal memory network according to an embodiment of the present invention;

[0037] Figure 3 is a processing flow chart of a graph attention network based on knowledge enhancement according to an embodiment of the present invention;

[0038] Figure 4 This is a module architecture diagram of the computing power infrastructure monitoring system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0039] To more clearly illustrate the embodiments of the present invention, specific embodiments of the present invention will be described below with reference to the accompanying drawings. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings and other embodiments can be obtained based on these drawings without inventive efforts.

[0040] The computing power infrastructure monitoring architecture of the embodiment of the present invention is a four-layer structure, including a data acquisition layer, an intelligent analysis layer, a dynamic scheduling layer and a response scheduling layer. The data acquisition layer realizes real-time collection and standardized processing of multi-dimensional heterogeneous data; the intelligent analysis layer realizes abnormal fault detection through a spatiotemporal memory network based on mixed density, and realizes root cause analysis and tracing of faults through a graph attention network based on knowledge enhancement; the dynamic scheduling layer realizes resource busy and idle prediction through a long and short-term memory neural network based on a self-attention mechanism, so as to better schedule and optimize resource utilization; the response and disposal layer realizes automated alarms and intelligent fault handling.

[0041] Specifically, if Figure 1 As shown, the computing power infrastructure monitoring method according to an embodiment of the present invention includes the following steps:

[0042] (1) Real-time data collection;

[0043] The underlying data collected in real time by the embodiment of the present invention includes physical infrastructure data, virtualized resource data, and business application data;

[0044] The above-mentioned physical infrastructure data mainly includes chip-level related temperatures, server-level overall power consumption, fan speed, cabinet-level current value, sensor temperature, etc.

[0045] The above virtualization resource data collection mainly includes container-related resource utilization, open source platform K8S node resource allocation rate, network-related packet loss rate, etc.

[0046] The data collected by the above business application layer includes transaction processing volume, API response latency, storage performance, etc.

[0047] (2) Data standardization;

[0048] This embodiment of the present invention standardizes the collected underlying data, including standardizing timestamps, unit conversion, and feature construction. Unit conversion, such as converting minutes and seconds to seconds, is performed. Feature construction primarily generates important composite metrics, such as energy efficiency ratio (EER) = computing throughput / total machine power consumption. Alternatively, it constructs time series features, such as sliding window statistics and Fourier transforms to extract periodic features. After the underlying data undergoes this standardization, standardized data is obtained.

[0049] (3) Real-time abnormal fault detection;

[0050] In the embodiment of the present invention, the above standardized data is transmitted in real time to the spatio-temporal memory network based on mixed density in the intelligent analysis layer for abnormal fault detection.

[0051] Due to the deficiencies of the traditional long short-term memory neural network LSTM in modeling multi-modal time series data, the single density hypothesis cannot capture complex abnormal patterns, and the existing defect that the fixed threshold mechanism leads to a high false alarm rate, the embodiment of the present invention designs a spatio-temporal memory network based on mixed density, with the ability to jointly model from the spatio-temporal perspective, and realizes dynamic adjustment based on dynamic probability density estimation.

[0052] Specifically, the core components of the spatio-temporal memory network based on mixed density include: spatio-temporal feature extraction layer, mixed density memory layer, variational inference calculation layer, dynamic threshold adjustment layer, as Figure 2 shown, the specific training process includes:

[0053] (a)Collect the historical data set of the underlying data. After the above-mentioned standardized processing, the standard historical data set is obtained, and then input into the spatio-temporal feature extraction layer to obtain spatio-temporal features; specifically, the spatio-temporal feature extraction layer includes a two-stream encoding structure, namely a time stream and a space stream. Specifically, bidirectional LSTM is used to capture time features, and a graph attention network GAT is used to construct space features, and then based on the gated attention mechanism, the time features and space features are fused to obtain spatio-temporal features;

[0054] ;

[0055] Among them, is the spatio-temporal feature, is the set parameter, is the transposed matrix of the input data matrix, is the hidden state of the time stream, is the hidden state of the space stream, is the hidden state of the gated mixture;

[0056] (b)The mixed density memory layer designs a memory matrix based on the spatio-temporal features, such as splicing, and dynamically maintains M memory slots, where M is a positive integer, and each memory slot stores a time label; the density function of the Gaussian mixture model is calculated for the memory matrix to obtain the mixed density probability, and according to the preset mixed density probability rule (such as from large to small), the memory matrix is dynamically allocated to several memory slots, one memory matrix is stored in one memory slot, and the source corresponding to the memory matrix not allocated to the memory slot is determined as a normal sequence, while the source corresponding to the memory matrix allocated to the memory slot is determined as an abnormal sequence. Among them, the calculation of the density function of the Gaussian mixture model can refer to the existing technology and will not be elaborated here;

[0057] (c) The variational inference calculation layer infers the latent variable distribution of the memory matrix in the memory slots through the encoder, and reconstructs the output anomaly sequence and anomaly score through the decoder; among them, the KL loss is selected for loss calculation;

[0058] (d) Reconstruct and output the normal sequence;

[0059] The dynamic threshold adjustment layer statistically analyzes based on a sliding window and introduces an online learning mechanism for fusion to perform dynamic threshold adjustment until the dynamic threshold is adjusted to the preset target range, and a spatio-temporal memory network based on a mixture density for anomaly fault detection is obtained;

[0060] ;

[0061] Among them, is the mean of the error between the predicted result and the label, is three times the standard deviation of the error, is the entropy of a certain probability distribution p at time t t of, is a variable parameter;

[0062] In the embodiment of the present invention, a closed loop is formed through dynamic threshold adjustment and cyclic monitoring;

[0063] The specific anomaly fault detection process is as follows: The standardized data is input into the spatio-temporal feature extraction layer. The time features are captured through a bidirectional long short-term memory neural network, the spatial features are constructed through a graph attention network, and the time features and spatial features are feature-fused based on a gated attention mechanism to obtain spatio-temporal features; the spatio-temporal features are input into the mixture density memory layer, a memory matrix is designed, the density function of the Gaussian mixture model is calculated for the memory matrix to obtain the mixture density probability, and the memory matrix is dynamically allocated to several memory slots according to the mixture density probability. The memory slots are labeled with time tags; the memory matrix in the memory slots is input into the variational inference calculation layer. The latent variable distribution of the memory matrix in the memory slots is inferred through the encoder, and the anomaly sequence is reconstructed and output through the decoder to achieve anomaly fault detection.

[0064] (4) Anomaly fault root cause analysis;

[0065] The data of the monitored anomaly sequence is transmitted in real time to the root cause analysis algorithm in the intelligent analysis layer, that is, the knowledge-enhanced graph attention network, for root cause tracing of the anomaly fault and clarifying the fault source.

[0066] Due to the defects that the traditional Graph Attention Network (GAT) easily ignores important domain knowledge and its static topological structure cannot dynamically obtain real-time status, a method of jointly embedding a knowledge graph and real-time data is designed to dynamically achieve the ability of causal reasoning through a multi-modal relational attention mechanism.

[0067] Specifically, the core components of the knowledge-enhanced graph attention network include a knowledge fusion layer, a multi-relational graph attention layer, and a causal reasoning layer.

[0068] As Figure 3 shown, the knowledge fusion layer of the embodiment of the present invention performs knowledge fusion through constructing a joint embedding of the knowledge graph based on the input abnormal sequence to obtain a knowledge graph.

[0069] The above method of joint embedding is:

[0070] ;

[0071] Among them, is the initial embedding vector of the i-th node, is the weight matrix, is the monitoring index obtained by hybrid cross-computation of the real-time underlying data of the i-th node, is the device type code of the i-th node, is the embedding vector of the knowledge graph of the i-th node;

[0072] The multi-relational graph attention layer of the embodiment of the present invention aggregates heterogeneous features of the knowledge graph and the abnormal sequence through relational attention coefficients to obtain a graph feature matrix.

[0073] The causal reasoning layer of the embodiment of the present invention performs causal influence calculation on the graph feature matrix according to the influence propagation model, so as to perform root cause localization and output abnormal graph nodes. Among them, the influence propagation model can specifically refer to the prior art and will not be elaborated here.

[0074] (5) Dynamic scheduling and optimization of resources;

[0075] During fault tracing or automated repair, in order to achieve stable and efficient utilization of resources, the real-time monitored data is input into the long short-term memory neural network based on the self-attention mechanism in the dynamic scheduling layer to predict the busy and idle status of resources, thereby triggering the action of automatically and reasonably allocating resources, and realizing the reasonable scheduling and optimization of resources. Among them, the specific structure and processing process of the long short-term memory neural network based on the self-attention mechanism can refer to the prior art and will not be elaborated here.

[0076] (6) Automated alarm and fault handling;

[0077] For faults that cannot be autonomously repaired or exceed the disaster tolerance threshold, real-time alarm notifications are issued, providing handling suggestions and risk scores.

[0078] As Figure 4 shown, the computing power infrastructure monitoring system according to the embodiments of the present invention includes the following functional modules: a collection module, a standardization module, an anomaly detection module, a root cause tracing module, a prediction module, and an alarm handling module;

[0079] The above-mentioned collection module is used to collect underlying data in real time;

[0080] The above-mentioned standardization module is used to perform standardization processing on the underlying data to obtain standardized data;

[0081] The above-mentioned anomaly detection module is used to input the standardized data into a spatio-temporal memory network based on a mixture of densities for anomaly fault detection;

[0082] The above-mentioned root cause tracing module is used to trace the root cause of the anomaly fault based on a knowledge-enhanced graph attention network to locate the fault source;

[0083] The above-mentioned prediction module is used to input the standardized data into a long short-term memory neural network based on a self-attention mechanism to predict the busy and idle states of resources, realizing resource scheduling and optimization;

[0084] The above-mentioned alarm handling module is used to issue real-time alarm notifications for faults that cannot be autonomously repaired or exceed the disaster tolerance threshold, providing fault handling suggestions and fault risk scores;

[0085] The detailed processing procedures of the above functional modules can refer to the detailed description of the above monitoring method and will not be elaborated here.

[0086] The computer-readable storage medium according to the embodiments of the present invention stores instructions, and when the instructions run on a computer, the computer is caused to execute the above monitoring method to realize the intelligence of computing power infrastructure monitoring.

[0087] The above is only a detailed description of the preferred embodiments and principles of the present invention. For those of ordinary skill in the art, according to the idea provided by the present invention, there will be changes in the specific implementation manners, and these changes should also be regarded as the protection scope of the present invention.

Claims

1. A method for monitoring computing power infrastructure, characterized in that, Including the following steps: S1. Collect underlying data in real time; among them, the underlying data includes physical infrastructure data, virtualized resource data, and business application data; the physical infrastructure data includes chip temperature, overall power consumption of the server, fan speed, current value of the cabinet, and sensor temperature; the virtualized resource data includes resource utilization related to containers, allocation rate of resources of the open-source platform K8S nodes, and packet loss rate related to the network; the business application data includes transaction volume, application programming interface API response latency, and storage capacity. S2. Perform standardization processing on the underlying data to obtain standardized data. S3. Input the standardized data into the spatio-temporal memory network based on mixture density for anomaly and fault detection. Among them, the spatio-temporal memory network based on mixture density includes: a spatio-temporal feature extraction layer, a mixture density memory layer, and a variational inference calculation layer. The standardized data is input into the spatio-temporal feature extraction layer. The time features are captured through a bidirectional long short-term memory neural network, and the spatial features are constructed through a graph attention network. Feature fusion of the time features and the spatial features is performed based on a gated attention mechanism to obtain spatio-temporal features. The spatio-temporal features are input into the mixture density memory layer. A memory matrix is designed, and the density function of the Gaussian mixture model is calculated for the memory matrix to obtain the mixture density probability. According to the mixture density probability, the memory matrix is dynamically allocated to several memory slots, and the memory slots are labeled with time tags. Among them, one memory slot stores one memory matrix, and the source corresponding to the memory matrix not allocated to the memory slot is determined as a normal sequence, while the source corresponding to the memory matrix allocated to the memory slot is determined as an abnormal sequence. The memory matrix in the memory slot is input into the variational inference calculation layer. The latent variable distribution of the memory matrix in the memory slot is inferred through an encoder, and the abnormal sequence is reconstructed and output through a decoder to achieve anomaly and fault detection.

2. The computing power infrastructure monitoring method according to claim 1, wherein It also includes the following steps: S4. Perform root cause tracing of the anomaly and fault based on the knowledge-enhanced graph attention network to locate the fault source.

3. The computing power infrastructure monitoring method according to claim 2, wherein The knowledge-enhanced graph attention network includes a knowledge fusion layer, a relational attention mechanism layer, and a causal inference layer. The abnormal sequence is input into the knowledge fusion layer, and a knowledge graph joint embedding is constructed for knowledge fusion to obtain a knowledge graph. The knowledge graph is input into the relational attention mechanism layer, and heterogeneous feature aggregation is performed through relational attention coefficients to obtain a graph feature matrix. The graph feature matrix is input into the causal inference layer, and root cause determination is performed according to the influence propagation model, and abnormal graph nodes are output to locate the fault source.

4. The computing power infrastructure monitoring method according to claim 3, wherein, The method for constructing a knowledge graph joint embedding for knowledge fusion is: ； wherein, is the initial embedding vector of the i-th node, is the weight matrix, is the monitoring metric obtained by hybrid cross-computation of the real-time underlying data of the i-th node, is the device type code of the i-th node, is the embedding vector of the knowledge graph of the i-th node.

5. The computing power infrastructure monitoring method according to any one of claims 2-4, characterized in that, It also includes the following steps: S5. Input the standardized data into the long short-term memory neural network based on self-attention mechanism to predict the busy and idle state of resources, and realize the scheduling and optimization of resources.

6. The computing power infrastructure monitoring method according to claim 5, wherein It also includes the following steps: S6. Issue real-time alarms for faults that cannot be automatically repaired or exceed the disaster tolerance threshold, and provide fault handling suggestions and fault risk scores.

7. The method for monitoring computing power infrastructure according to any one of claims 1-4, characterized in that In step S2, the standardization processing includes unifying the timestamp, unit conversion, and feature construction.

8. A computing power infrastructure monitoring system, which applies the computing power infrastructure monitoring method according to any one of claims 1-7, is characterized in that The computing power infrastructure monitoring system includes: A collection module for collecting underlying data in real time. A normalization module for normalizing underlying data to obtain normalized data; An anomaly detection module for inputting the normalized data into a spatio-temporal memory network based on a mixture density for anomaly fault detection.

9. A computer-readable storage medium storing instructions, characterized in that, When the instruction runs on a computer, the computer is caused to execute the monitoring method according to any one of claims 1-7.

Images