306 results about "Cluster Node" patented technology

The invention relates to the technical field of data analysis, and discloses a GPU computing power resource scheduling method and system, and the method comprises the steps: collecting node hardware parameters and dynamic load indexes of a GPU cluster to construct a multi-dimensional resource feature vector of the GPU cluster, and constructing a resource portrait of the GPU cluster; establishing a node health degree scoring model of the GPU cluster, and generating a health degree score of a cluster node corresponding to the GPU cluster; analyzing a video memory demand of the GPU task request, and calculating an intensive identifier and a communication dependency relationship; determining the SLA weight of the GPU task request, calculating the resource shortage sensitivity of the GPU task request based on the video memory demand, and calculating the target task priority of the GPU task request in combination with the SLA weight; and determining a resource scheduling node group requested by the GPU task in the resource portrait, generating resource scheduling parameters of the resource scheduling node group, and executing scheduling of computing power resources of the GPU cluster based on the resource scheduling parameters. According to the method, the scheduling efficiency of the GPU computing power resources can be improved.

The invention discloses a cluster computing power energy efficiency perception scheduling and green computing system, which relates to the technical field of computers and comprises a multi-source energy efficiency perception and data acquisition module used for acquiring power consumption, utilization rate, temperature, cooling state, PUE index and environmental data of cluster nodes. According to the invention, through the multi-modal energy efficiency fusion sensing network and the multi-scale convolution and time sequence attention fusion network, multi-source heterogeneous energy efficiency data such as current, voltage, temperature, airflow and the like of a node level can be collected and fused in real time and with high precision, noise is effectively removed, abnormity self-correction is realized, the defect of energy efficiency sensing granularity in the prior art is made up, and the energy efficiency sensing precision is improved. And reliable input is provided for subsequent scheduling decisions. A cross-scale dynamic twinborn collaborative modeling mechanism is adopted, a physical information neural network and a computational fluid mechanics model are coupled, optimization is carried out through a generative adversarial network structure, and accurate prediction of a complex energy consumption evolution curve and a cooling flow field is achieved.

The invention discloses an inference service deployment method, electronic equipment and a storage medium, and the method comprises the steps: responding to the creation of an inference service, reading a model information configuration table and an engine information configuration table, and carrying out the matching processing of the model information configuration table and the engine information table according to a pre-configured model name during the creation of the inference service, processing resources, inference engine mirror image identifiers and inference engine starting parameters which can be used for loading inference services on the container scheduling platform are obtained, and workload metadata used for bearing the inference services are created based on the processing resources, the inference engine mirror image identifiers and the inference engine starting parameters; the deployment of the inference service on the container scheduling platform is realized by distributing the workload metadata to the target cluster node. Through the method, the technical problem of relatively low deployment efficiency caused by manually configuring the parameters in the workload metadata depending on manpower in related technologies is solved, and the technical effects of simplifying the deployment process of the model reasoning service and improving the deployment efficiency are achieved.

The invention discloses an MQTT message transmission optimization method and system. The method comprises the steps of analyzing a theme, extracting a device type, a data feature and a geographic position triple, calculating a hash value, and mapping a device to a specified Broker fragment cluster node to generate a fragment mapping table; processing the equipment data in the edge domain in the fragment mapping table through an edge calculation layer, removing invalid data according to a preset rule, merging the equipment data in the same fragment node, and embedding a fragment node ID for an aggregation message generated after merging; identifying a fragment node ID and routing to a target fragment node, positioning a corresponding shared memory pool, and writing the message into the shared memory pool; and responding to a direct access request of the client, so that the client directly accesses the data from the shared memory pool through the user mode network stack. According to the method, the problem of uneven load is effectively solved, multiple times of state switching in the data transmission process is avoided, and the MQTT message transmission efficiency is improved while the transmission cost is reduced.

The invention relates to the technical field of power loads, in particular to a power load dynamic optimization prediction method based on reinforcement learning, which comprises the following steps: acquiring real-time voltage frequency power of nodes, tracking frequency difference direction change of adjacent nodes, marking reverse disturbance to generate a distribution map, and identifying a frequent disturbance area according to the distribution map; clustering node groups with consistent fluctuation trends to determine a target area; monitoring power fluctuation inversion to lock energy steering nodes; constructing a prediction input set; comparing prediction and actual trends to extract a deviation interval, adjusting stride and rate of a reinforcement learning model, updating a decision and smoothing a curve, and outputting a power load dynamic prediction curve. According to the method, disturbance space-time dynamic identification is realized through multi-layer correlation analysis, judgment precision is improved through frequency power joint calibration, load path consistency and area extension are reflected through node clustering, power transmission tracking is enhanced through energy node identification, a feedback closed loop is constructed through offset monitoring, and curve continuity and response rate are improved. Stable convergence is predicted to be consistent with the trend under multiple disturbances.

The invention discloses a distributed database built-in monitoring system, belongs to the technical field of databases, and aims to solve the technical problem of how to overcome the defects of complex monitoring deployment, high data response delay, coarse information acquisition granularity and low system integration in the prior art. According to the technical scheme, a monitoring platform module is integrated with a database kernel, a front-end and rear-end separation framework is adopted, and visual display and centralized management and control of the database operation state are achieved; the data acquisition module is used for simultaneously starting with a database process through a built-in HTTP (Hyper Text Transport Protocol) server to realize native support of a monitoring function; the distributed state query module is used for discovering cluster node states through a Gossip protocol, supporting node health detection and remote calling state service, and supporting gRPC-based streaming communication and concurrent processing; and the time sequence data management module is used for storing, compressing and querying the collected time sequence monitoring data.

This application provides a radio access network architecture. The radio access network architecture includes a cluster node and a serving node. The serving node provides task scheduling and executing functions, and the cluster node provides a region-level centralized collaboration function for the serving node and a collaboration function between cross-region cluster nodes.

The invention discloses an execution time aware computing power network heterogeneous GPU cluster deep learning task node allocation method and system. The method comprises the following steps: firstly, based on a deep learning task, extracting and preprocessing task features and available node features; secondly, a sampler equally divides new tasks without historical data to available nodes, and each node performs mixed sampling on the tasks until all the tasks estimate execution time data; taking execution time data as a training set, taking the task features and the node features as a test set, and using a regression decision tree model to predict the execution time of the task on each node; performing task allocation on each node by using a cost search algorithm and a short job total JCT priority strategy; and finally, periodically monitoring node resources released in the cluster to obtain an optimal node allocation result. According to the method, task delay and total task JCT are remarkably reduced, cluster node resource changes are monitored in real time, and the resource utilization rate is increased.

The invention discloses a distributed system for dynamic perception and adaptation of an unmanned aerial vehicle cluster environment, and belongs to the technical field of computers. The distributed system comprises a sensing layer used for collecting state information of each node in an unmanned aerial vehicle cluster in real time, and a data processing layer used for preprocessing and analyzing the collected unmanned aerial vehicle node state information; the decision-making layer is used for generating an adjustment decision for an unmanned aerial vehicle cluster environment based on a preset unmanned aerial vehicle cluster adaptation strategy library and the dynamic feature data; the execution layer is used for performing corresponding adjustment operation on the unmanned aerial vehicle cluster nodes according to the adjustment decision; and the feedback layer is used for collecting new state information after the unmanned aerial vehicle node adjustment operation. Through innovative fusion and hierarchical collaboration of multiple technical modules and dynamic balance of a strategy library, cluster orderliness is maintained, and a self-growing closed loop of'perception-decision-execution-feedback 'is formed.

The invention relates to a hierarchical occlusion perception grid infinitesimal streaming transmission and adaptive rendering method, which comprises the following steps of: S1, decomposing a model into grid infinitesimal, recursively constructing a bounding volume hierarchical tree BVH from bottom to top by taking the grid infinitesimal as a leaf node, and setting a semantic error and a geometric error for each sub-node cluster; s2, acquiring camera parameters of a current frame, comparing a total screen space error with an error threshold value, if the error is smaller than the threshold value, stopping traversing, and adding the cluster node into a to-be-rendered list of the current frame; and S3, if the grid infinitesimal corresponding to the cluster node in the to-be-rendered list is not loaded to the GPU memory, sending a loading request to an asynchronous streaming manager, asynchronously loading the corresponding grid infinitesimal from the memory, and rendering a father node of the cluster node. Compared with the prior art, the method has the advantages that large-scale and high-precision model data can be rendered on the lightweight XR all-in-one machine in real time, and the like.

The invention discloses a control method and device of a container management system and a storage medium, and relates to the technical field of container management, and the control method of the container management system comprises the steps: responding to a trigger operation of a dynamic resource demand, and analyzing the operation type of the trigger operation and the operation content of the trigger operation; obtaining an initial resource quota allocated when the triggering operation corresponds to the tenant registration; determining a resource demand variation corresponding to the trigger operation according to the operation type and the operation content; substituting the used resource quantity, the initial resource quota and the resource demand change quantity into a preset resource matching formula, and calculating the real-time resource allocation quantity of the tenant; and adjusting the resource use quota of the tenant according to the real-time resource allocation amount, and synchronously updating the resource scheduling strategy of the cluster node. The technical effect of dynamic resource scheduling of the tenants is achieved.

The invention discloses a GPU resource scheduling method based on data awareness and computer equipment, and the method comprises the steps: carrying out the real-time monitoring of a cluster node, obtaining a current cluster load state, and predicting the cluster load state in a future time period; creating services according to an input service definition description file, adding the services into a to-be-scheduled service queue, periodically selecting the services from the to-be-scheduled queue, and allocating computing nodes for the services; distributing a target service for the user request according to the current cluster load state, establishing a session, and analyzing the intention of the user; dynamically adjusting and allocating static resources through deep reinforcement learning based on a hybrid expert model according to a user intention analysis result and a cluster load state; required resources are loaded to a GPU through a multi-level cache architecture, and after calculation is completed, a calculation result is returned to a user through a session. According to the method, the utilization rate and the performance of GPU resources in an interactive GPU acceleration service scene can be improved, and the energy consumption and the cost are reduced.

The invention relates to the technical field of unmanned aerial vehicle ad hoc network cluster management and control scheme design based on intelligent street lamps and satellite communication, in particular to an unmanned aerial vehicle ad hoc network cluster management and control system based on intelligent street lamps and satellite communication. The system comprises an intelligent street lamp subsystem, a satellite communication subsystem, an unmanned aerial vehicle cluster node and a dynamic cooperative control platform. The innovative scheme comprises the following steps: (1) the multimode communication gateway realizes seamless conversion between a satellite protocol and a ground Internet of Things protocol; (2) a street lamp sensor generates a dynamic risk map to guide cluster obstacle avoidance; (3) supporting signal interruption emergency networking by a rapid clustering algorithm based on geometric topology; (4) task priority drives time slot distribution to improve communication efficiency; and (5) the satellite-street lamp-relay unmanned aerial vehicle multistage redundancy guarantee system is continuously controllable. According to the scheme, the communication reliability, the environmental adaptability and the cooperative capability of the cluster in a complex urban area are remarkably enhanced.

The invention discloses a traffic flow prediction method for a double-layer multi-scale dynamic graph convolutional network, and relates to the technical field of traffic, and the method comprises the following steps: constructing a double-layer structure of a traffic network, which comprises a node layer composed of traffic sensor nodes and a region layer formed by clustering nodes in the node layer, node layer traffic flow data and area layer traffic flow data are extracted; respectively mapping the node layer traffic flow data and the region layer traffic flow data to a potential space to obtain a node layer initial hidden state and a region layer initial hidden state; inputting the initial hidden state of the node layer and the initial hidden state of the region layer into a sequence containing at least one space-time module for processing so as to capture a dynamic space-time dependency relationship; the method has the effect of providing more reliable decision support for traffic management and travel planning.

The invention discloses a MySQL database management platform, which relates to the field of database management optimization, and comprises the following steps: acquiring an operation feature vector, constructing a waiting time prediction model in combination with transaction feature parameters, obtaining an optimal waiting threshold, and forming a consistency time window baseline and a consistency time window default rate; collecting cluster node operation state information, calculating a node comprehensive health score and a diagnosis health score, obtaining a cluster comprehensive health score, and determining a new main node; shadow read-write testing is carried out on the new main node, and a batch revenue function is constructed to obtain a preliminary low-load time window; further converging a low-load time window to determine a lowest-load window, putting online candidate indexes in the lowest-load window in stages, dynamically monitoring and adjusting a batch scale and an execution interval in an execution process, and calculating a comprehensive improvement score according to a consistent time window baseline, a comprehensive health score, the lowest-load window and online monitoring data; and updating the consistent time window baseline and the comprehensive health score.

The invention discloses a task planning-based material information processing method and device. The method comprises the steps of obtaining a multi-modal material information file; dividing the material information file according to the target content of the material information file to generate a plurality of analysis sub-tasks; according to a processing cluster resource state, dynamically distributing the analysis subtasks to cluster nodes, and giving structured information corresponding to the analysis subtasks; substituting the structured information into the information processing model to obtain processing result data corresponding to the structured information, the processing result data including a processing score, a risk probability and a supplier order; and carrying out AB test weight adjustment voting fusion on the processing result data to generate a material information report comprising multi-modal evidence. By dividing the multi-modal material information file, synchronous processing of multi-type data of the bid invitation file is realized, the data processing efficiency and accuracy are improved, a reliable data basis is provided for bid evaluation, and the intelligent level and decision quality of a bid invitation process are comprehensively improved.

The invention relates to a method and device for improving IO reading performance based on object affinity. The method comprises the steps that a certain number of GC objects are configured in advance according to the cluster scale, and the affinity relation between the GC objects and all nodes is calculated; when a storage object is applied through aggregation writing or GC transfer writing, determining a GC object associated with the storage object according to the identification information of the storage object, and judging whether the GC object is compatible with a target node or not; if not, discarding the storage object and reselecting, and if yes, writing the data into the storage object; and performing a read service according to the local cache of the storage object in the target node. According to the method, by combining a read-write cache mechanism, the affinity relationship between the cluster node and the GC object is calculated in advance, and the data is written into the storage object compatible with the GC object, so that even if GC moving operation is triggered due to continuous coverage write, the foreground read IO still can obtain effective data from the cache of the original node, and the data storage efficiency is improved. The performance loss of data cross-network transmission and slow medium reading is greatly reduced, and the reading performance is remarkably improved.

The invention discloses an unmanned aerial vehicle cluster network node security isolation and re-access method, and the method comprises the steps: obtaining the behavior deviation attribute data of an unmanned aerial vehicle cluster node, and constructing an anomaly detection model in combination with a cluster topological structure attribute; calculating the comprehensive behavior deviation degree of each node, judging that the node is an abnormal node when the comprehensive behavior deviation degree exceeds a preset threshold value, and performing isolation processing on the abnormal node; behavior data and software and hardware state log data of the abnormal nodes are obtained, and diagnosis and restoration are carried out; obtaining state data of the repaired abnormal node, and dynamically verifying whether the node state of the repaired abnormal node is recovered to be normal or not; and the connection relation of the nodes passing the verification is adjusted according to the centrality of the nodes and the clustering coefficient, and the nodes are re-fused into a cluster topological structure. According to the invention, identification, isolation and repair of abnormal nodes of the unmanned aerial vehicle cluster and re-access to the network after security verification can be realized on the premise of ensuring network security, and the robustness, security and efficiency of the unmanned aerial vehicle cluster are improved.

The invention relates to the technical field of storage equipment, and particularly discloses an RAID card high-availability device and method based on a multi-path cluster, and the device comprises at least two RAID cards; the RAID engine is used for processing RAID read-write data, configuring RAID levels and processing hard disk hot plug; the front-end multi-path module is used for establishing a plurality of data channels with a host and exposing a virtual disk with the same NGUID to the host, so that the host multi-path module can access the virtual disk in an aggregated manner; the rear-end multi-path module is used for establishing a multi-path data channel with the dual-port hard disk, transmitting read-write data and commands of the dual-port hard disk, and realizing disk access link redundancy and fault automatic switching; the system management module is used for processing internal and external management control events, state conversion and internal processes of the RAID card; the cluster management module is used for establishing communication channels among the RAID cards, maintaining heartbeat information among the RAID cards and managing cluster node states; the problem of single-point failure of a traditional RAID card is solved, and the reliability and performance of a system are improved.

The invention provides a computing power resource allocation method and device, equipment and a storage medium, and the method comprises the steps: obtaining an opening application of a tenant, determining the grade type of the tenant according to the property of the tenant of the opening application, and determining the number range of target adaptive nodes according to the grade type, the target adaptive node number range is limited based on the tenant property, and resource waste caused by the fact that the tenants occupy too many nodes is avoided; according to the target adaptive node number range, configuring the scaling strategy of the computing power cluster nodes of the computing power cluster, monitoring the function indexes of the computing power cluster in the process that the computing power cluster is in the use state, and dynamically adjusting the number of the computing power cluster nodes according to the function indexes and the scaling strategy, so that the number of the computing power cluster nodes can be automatically and dynamically adjusted; and effective allocation of computing power resources of the computing power cluster nodes is facilitated.

The invention provides a Pod fault injection method and system based on kubemark in a kubernetes cluster, and the method comprises the steps: firstly providing a fault collection tool which is used for collecting Pod fault information in a real cluster; meanwhile, using modified Kubemark in a simulation cluster to create a virtual node with the same specification as a real cluster node, and deploying a CRI-Proxy (a container runtime interface agent) component to execute fault injection and container life cycle management; a fault scene is simulated through a'fault recording-fault broadcasting 'mode, and Pod and container life cycle abnormity are simulated through fault injection, so that the stability and the recovery capability of the system are tested, valuable performance test data and system optimization guidance are provided for a real cluster, resources are saved, the test efficiency is improved, and the test cost is reduced. And meanwhile, problems can be reproduced, potential defects can be found, and various risks can be avoided in advance.

The invention relates to an automatic cluster recombination method, a distributed system and computer equipment. Determining a plurality of independent network partitions after the distributed system is subjected to network partitioning; for each independent network partition, determining first existence information of each cluster node in each independent network partition; determining a target local leader node and a target follower node from the cluster nodes of the independent network partitions according to a preset consensus recombination strategy based on the first existence information; establishing a first local consensus cluster corresponding to the independent network partition according to a following relationship between the target local leader node and the target follower node; according to the invention, automatic reorganization of cluster nodes during network partitioning is realized, so that each independent network partition can still maintain local consensus and service availability under the condition that global intercommunication cannot be realized, service is provided to the greatest extent, the availability is ensured, and the service continuity and the overall availability of the system after network partitioning are remarkably improved.

The invention discloses a search engine cluster management method and device, electronic equipment and a program product, and belongs to the technical field of cloud computing. According to the method, deployment of a search engine cluster and a front gateway cluster is completed based on custom resource definition of a container arrangement platform; the front gateway cluster is responsible for receiving user requests, analyzing the requests, converting the requests into structured data units, storing the data units into the buffer queue subunits of the front gateway unit, and optimizing the user requests, so that the request processing cost in a high-concurrency scene is effectively reduced, and the user experience is improved. And the query and data writing efficiency is improved. Furthermore, the number of cluster nodes and the resource capacity are dynamically adjusted based on operation index monitoring and resource demand prediction results, and efficient response and stable operation of the search engine cluster are guaranteed.

This invention discloses a method and apparatus for implementing a high-availability cloud firewall cluster, relating to the field of cloud firewall technology. The method includes session-based traffic splitting of data packets using a Session Flow Table (SLB); simplified processing of data streams for ALG-related protocols; and real-time monitoring of the health status and concurrent session count of each firewall node via SLB to achieve dynamic scaling of firewall nodes. This invention supports automatic elastic scaling of cloud firewalls, as well as multi-active high-availability clusters for NAT and ALG scenarios, while avoiding hot session standby between cluster nodes.

The invention discloses a data migration method and device, a storage medium, electronic equipment and a program product, and relates to the field of distribution.The method comprises the steps that original transaction data of a host side is split into N data fragments according to service types and time windows; encrypting each data fragment into a target data fragment; leader nodes corresponding to the target data fragments are selected from the L cluster nodes, and each target data fragment and the corresponding data version vector are sent to the corresponding leader node; performing consensus confirmation on the target data fragment received by the leader node through the leader node and the M replica nodes to obtain a consensus result of the target data fragment; and migrating the target data fragment from the host end to a storage node in a cluster. The technical problem of low consistency judgment efficiency before and after data migration caused by lack of a data version control mechanism during data migration of a distributed storage system in the prior art is solved.

The embodiment of the invention provides a training cluster check point storage method, a training cluster check point recovery method and electronic equipment. The training cluster check point storage method comprises the steps that distributed cluster node memories cooperatively store check points, and the memories are used for storing check point data of each round of training. According to the method provided by the embodiment of the invention, cluster memory collaborative management is adopted, the node memory stores the whole check point, and the storage frequency of the check point is improved; in addition, a cluster memory control manages multi-node check points and supports grouped redundant backup; and moreover, the check points are directly stored in the memory, so that the time delay of storing the check points is greatly reduced.

The invention relates to the technical field of cloud computing, and particularly provides a cloud native database kubernetes dynamic scheduling method and equipment based on a tensor structure. The method comprises the following steps: acquiring a resource creation request; collecting multi-dimensional resource information of the cluster nodes in real time, and determining Pod types corresponding to all Pods in the cluster nodes based on the multi-dimensional resource information; the Pod type comprises a computing type, a storage type, a network type and an I / O type; based on the Pod type, the time dimension and the resource dimension, constructing a three-order tensor structure; dynamically calculating a node combination weight of the three-order tensor structure, and determining a target scheduling node; and scheduling the Pod corresponding to the resource creation request to the target scheduling node. According to the method, the problems of performance bottleneck and uneven load caused by resource view limitation are avoided, the technical problem of single measurement index in resource scheduling is solved through multi-target weight optimization, and the accuracy of target scheduling node selection is remarkably improved.

The invention relates to the technical field of intelligent management, in particular to an electronic city ICC management method and system based on a private cloud platform. The method comprises the following steps: carrying out physical resource and interface service real-time acquisition and standardization processing on an electronic city ICC cluster to obtain electronic city ICC standardized cluster resource data; constructing an electronic city ICC resource-service association model based on the electronic city ICC standardized cluster resource data, and performing mapping association and cluster node state analysis to obtain electronic city ICC cluster node state feature data; performing node load scheduling configuration and configuration optimization management and storage resource matching based on the electronic city ICC cluster node state feature data, and generating optimization deployment scheme data; and performing management iterative optimization on the optimized deployment scheme data to output a corresponding electronic city ICC management execution scheme. According to the invention, intelligent resource collaborative scheduling management of electronic city total factors can be realized.

The application provides a distributed cluster configuration method based on a SHELL script, comprising the following steps: architecture information and service distribution information of a cluster are intercepted by a script tool, and the intercepted information is stored in an associative array of the script tool; whether the cluster architecture information is a storage-computing integrated version or a storage-computing separated version is judged; a target configuration file is backed up, and the backup file is stored in a specified path; an instruction containing an operation type is received, the operation type comprises set, restart and rollback, and a parameter modification target is determined according to the instruction; before and after the parameter is modified, the parameter state is recorded and output. The application has the beneficial effects that the script tool is used to realize convenient modification of 8a cluster parameters, efficient restart of cluster services and quick recovery of configuration files, and through the script tool, configuration files of all cluster nodes corresponding services can be modified by one instruction.

The invention provides a dynamic disaster recovery method, system and device for cluster service shunting. The method comprises: obtaining load information of each server node in a service cluster, the load information comprising at least one of a CPU occupancy rate, a memory utilization rate, a disk occupancy rate and network port traffic; determining a load weight value of each server node according to the load information; service shunting information of the service cluster is acquired in real time, wherein the service shunting information comprises the number of server nodes for processing each service and a server node list; and updating the load weight value of each server node according to the service shunting information to obtain an updated load weight value of each server node, and generating alarm information and a service shunting scheme based on the updated load weight value of each server node. The problem that an existing intelligent flow distribution algorithm can dynamically adjust the flow distribution condition according to the load condition of cluster nodes and lacks disaster recovery processing means when the overall load of a cluster is high is solved.