422 results about "Server load" patented technology

The invention discloses a control method and a control system of a data center liquid cooling heat dissipation system, and relates to the technical field of liquid cooling heat dissipation, the system realizes real-time state monitoring of server operation through a data acquisition and monitoring module, integrates multiple data such as temperature, flow velocity and pressure difference, forms a detailed comprehensive heat dissipation data set, and improves the heat dissipation efficiency. And a database is constructed for storage. And meanwhile, a cooling liquid flow velocity analysis control module of the system generates a server heat dissipation coefficient CN through the server operation environment parameter C1 and the server operation parameter C2, the server heat dissipation coefficient CN is compared with a preset heat dissipation threshold C for evaluation, and the cooling liquid flow velocity is adjusted in real time. Therefore, when the load of the server is increased, the cooling intensity can be automatically increased, the chip temperature is reduced, and stable operation of the server is kept; and when the load is reduced, the cooling intensity is reduced, the energy consumption is reduced, and the cooling efficiency and the energy utilization rate of the system are remarkably improved.

The invention relates to the technical field of server load management and dynamic resource scheduling, and discloses a server load state evaluation method based on a dynamic evaluation algorithm, which comprises the following steps: data acquisition: acquiring performance index data, log data and network flow data from a multi-modal data source of a server; data preprocessing: performing time alignment, data cleaning, feature extraction and normalization processing on the acquired data; and self-supervised learning modeling: extracting feature representation from the multi-modal data through a self-supervised pre-training task, wherein the self-supervised pre-training task comprises a time sequence covering prediction task. Through multi-modal data fusion and self-supervised learning, the dynamic characteristics of the server load are accurately extracted, real-time load classification, anomaly detection and trend prediction are realized, the evaluation precision and real-time performance are significantly improved, and at the same time, a resource scheduling strategy of rule driving and reinforcement learning is combined, the server resource allocation efficiency is optimized, and the overload risk is reduced.

The invention relates to the technical field of liquid cooling, and discloses a server cabinet liquid cooling control method and system based on intelligent liquid cooling distribution, and the method comprises the steps: collecting the temperature, cooling liquid flow, pressure and environmental parameters of each node in a server cabinet in real time through a liquid cooling distribution unit; the load change of the server is predicted based on a dynamic distribution algorithm, and the pump set rotating speed and the valve opening degree of the liquid cooling distribution unit are dynamically adjusted so as to dynamically distribute the cooling liquid flow according to needs; a multi-parameter coupling optimization algorithm is adopted, and the energy efficiency ratio of the liquid cooling system is adjusted according to the temperature, the cooling liquid flow and the pressure; cooling liquid leakage is detected through a multi-mode sensor, a liquid leakage protection mechanism is triggered in a grading mode according to the severity degree of leakage, and the liquid leakage protection mechanism comprises local pipeline isolation and whole system shutdown. The system corresponds to the method. By adopting the method, the energy efficiency improvement and the reliability enhancement of the heat dissipation system of the server cabinet are realized.

The invention relates to the field of green computing and intelligent computing center management, and discloses a green intelligent computing center computing power dynamic regulation and control method based on carbon emission demand analysis, which comprises the following steps: data acquisition and analysis: acquiring multi-dimensional data of a green intelligent computing center in real time, including carbon emission intensity, server load, energy consumption, environmental conditions and cooling system energy efficiency data, preprocessing and analyzing the data; energy efficiency management and control standards are dynamically generated, the energy efficiency management and control standards of the intelligent calculation center are generated based on the data analysis result, and the energy efficiency management and control standards comprise a server load threshold value, carbon emission intensity limitation and a cooling system energy efficiency standard; dynamic regulation and control of computing power are implemented based on an energy efficiency management and control standard; and real-time load conditions, energy consumption characteristics and carbon emission intensity of the server are combined. According to the invention, by dynamically generating the energy efficiency management and control standard, feeding back the regulation and control strategy in real time and optimizing the computing power distribution, the efficient distribution of the computing power of the intelligent computing center is realized, the resource utilization rate is improved, and the carbon emission and the energy consumption are obviously reduced.

The invention discloses a railway intelligent construction site safety penetration type management messenger platform which comprises a multi-modal data fusion processing module, an edge computing node cluster module, a three-dimensional visual penetration type management interface module, an intelligent early warning and emergency response module, a self-adaptive network transmission module and the like. Real-time cleaning, alignment and correlation analysis are realized by using a dynamic data calibration algorithm, and a data island is broken; the edge computing node cluster carries out localization preprocessing and the like on data in a key area, so that the load of a central server is reduced; the three-dimensional visual interface is based on a digital twinborn construction model, supports drilling type viewing and realizes three-dimensional monitoring; the intelligent early warning system adopts a reinforcement learning model to automatically trigger multi-channel early warning; and the adaptive network transmission module dynamically switches communication modes to ensure low-delay transmission of key data. The platform realizes real-time acquisition and integration of construction site data and reduces manual intervention.

The invention discloses a resource scheduling method of a server and electronic equipment, and relates to the technical field of network optimization. Time features based on historical load data and spatial features based on load correlation of adjacent servers are introduced in a feature extraction stage; the dynamic change rule of the server load and the collaborative influence of the cluster environment are comprehensively captured, and the input information richness and context sensing ability of the prediction model are improved. The method comprises the following steps: predicting load data at a target moment through a load prediction model, carrying out weighted calculation on a server load state in combination with a task type weight, quantitatively evaluating actual occupation demands of different task types on server resources, and dynamically adjusting a task allocation strategy based on an inverse relation between the load state and an allocation priority. According to the method and the device, the problem that a load balancing strategy cannot be optimized in a short time according to fluctuation of task loads and changes of resource consumption can be solved, and refined scheduling of server cluster resources and collaborative optimization of task allocation efficiency are realized.

The invention relates to the technical field of data transmission and updating, in particular to a data transmission and updating system and method of a cloud mobile phone. According to the technical scheme, the system comprises an intelligent network monitoring module, a dynamic bandwidth allocation module, a server load balancing module, an intelligent data synchronization module and a cache management module, and the intelligent network monitoring module monitors network conditions, including network bandwidth, network congestion degree and network stability, between a user terminal and a cloud server in real time. Through the synergistic effect of the intelligent network monitoring module, the dynamic bandwidth allocation module, the server load balancing module, the intelligent data synchronization module, the cache management module and the like, the network condition is accurately monitored, the bandwidth is reasonably allocated, the server load is optimized, the data synchronization and completeness are ensured, and the cache efficiency is improved; stable and reliable network connection, preferential key data transmission and rapid data synchronization and loading are brought to users, and the use experience of the cloud mobile phone and the overall performance of the system are greatly improved.

The invention discloses a storage, calculation and training AI server load fluctuation self-balancing method based on a recursive game network. The method comprises the following steps: S1, recording real-time load data of each task; s2, constructing a macroscopic hierarchical recursive game network model based on the overall resource pool of the AI server; s3, generating an overall resource allocation strategy through multi-agent competition calculation of the macroscopic hierarchical recursive game network model; s4, constructing a micro-hierarchy recursive game network model based on the resource allocation strategy generated by the macro-hierarchy; s5, sub-task resources are optimally allocated; s6, issuing the adjusted resource quota to the microscopic hierarchical recursive game network model in real time; s7, carrying out resource pre-allocation on the upcoming high-load task; and S8, enabling resource allocation to gradually tend to an optimal state. According to the method, the resource utilization efficiency, the system response capability and the task completion stability of the AI server in a multi-task load fluctuation scene are remarkably improved.

The present invention discloses a server load prediction method based on deep learning, collecting the trend change of server load long series, and utilizing server load periodic information to establish a deep neural network prediction model to optimize peak load prediction. The present invention provides a method for improving the accuracy of neural network prediction by combining periodic information, long-term trend information, and short-term time series information, and demonstrates superiority over traditional methods in the peak load section. The method of the present invention can effectively improve prediction accuracy, provide more accurate scheduling and evacuation decision-making basis for cloud service providers, thereby reducing the redundant equipment required to ensure high reliability, reducing the operating costs of cloud service providers, and reducing the rental expenses of cloud service tenants.

The invention relates to the technical field of server heat dissipation, in particular to a cloud-based server heat dissipation method and system, and the method comprises the following steps: monitoring the key operation parameters of a server in real time through a cloud platform, updating the state of a heat dissipation unit according to the heat dissipation demands monitored in real time, optimizing the heat dissipation efficiency, and obtaining the optimal configuration of the heat dissipation parameters. Through real-time monitoring and data-driven heat dissipation control, heat dissipation efficiency and energy consumption management are remarkably optimized, dynamic heat dissipation unit state updating and optimization are introduced, so that heat dissipation configuration can respond to real-time load and temperature changes of the server in real time, and by continuously adjusting feedback control logic including the feedback period and strength, the heat dissipation efficiency of the server is improved. The influence of the state of the heat dissipation unit can be simulated and predicted, the optimal configuration of the heat dissipation parameters is realized, the energy utilization of the server is optimized, the optimal matching between the heat dissipation control and the server load is ensured, the energy consumption is reduced, and meanwhile, the heat dissipation control efficiency is improved.

The invention discloses a federated learning GPU acceleration ciphertext aggregation method based on homomorphic encryption, and relates to the technical field of information security, and the method comprises the following steps: a server initializes model parameters of a global model and sends the model parameters to a client; the client generates a private key and a calculation key for homomorphic encryption, trains the local model according to the received model parameters, updates the local model, encrypts the model parameters of the local model by using the private key and a mode component homomorphic encryption algorithm to obtain a ciphertext, and sends the ciphertext to the server; the server loads the ciphertexts of all the clients to the GPU, parallel addition calculation is carried out according to the calculation key to obtain an addition value, the addition value is transmitted back to the CPU for weighted average aggregation, and an aggregation ciphertext is obtained and sent to the clients; and the client decrypts the aggregated ciphertext by using the private key, the decrypted model parameters update the local model, and the training and encryption transmission process is repeated. The federal learning efficiency and precision can be effectively improved, and the learning safety is ensured.

The invention discloses a dynamic business data updating and querying method and system, and particularly relates to the technical field of information processing, and the method comprises the following steps: S1, real-time data acquisition and preprocessing; s2, updating the dynamic data; s3, intelligent query optimization; s4, result feedback and visualization are carried out; and S5, performing anomaly detection and self-repairing. According to the method, the real-time processing capability and query efficiency of massive heterogeneous data are remarkably improved, the distributed sensor network is combined with lightweight encryption and a differential privacy mechanism to realize localized cleaning and feature extraction of multi-source data, the load of a central server is effectively reduced, and the risk of sensitive information leakage is avoided; an incremental synchronization mechanism depends on a message queue and Hash chain log technology to ensure the atomicity and traceability of data updating in a high-concurrency scene, and by combining a dynamically generated space-time multi-dimensional index and an adaptive weight distribution algorithm, the system can intelligently analyze query semantics and optimize path selection, and the response time of complex query is greatly shortened.

The invention belongs to the technical field of data centers, discloses an energy efficiency evaluation method of an artificial intelligence data center, and solves the problem that a static evaluation method is disjointed with the real-time operation state of the data center in the prior art by constructing a dynamic feature library and an AI adaptive weight model. The dynamic feature library collects and associates service scene labels and time-space coupling data in real time, and provides a scene basis for an AI adaptive weight model, so that weight adjustment is converted from passive response index mutation in the prior art to active adaptive service scene change. Nonlinear influences of dynamic factors such as server load fluctuation and environment adjustment on energy efficiency are effectively captured, so that an energy efficiency evaluation result better fits the actual operation state of the data center, energy efficiency factor relevance changes caused by service switching of the AI data center are adapted, scene-based self-adaptive adjustment of weights is achieved through a deep learning model, and energy efficiency evaluation accuracy is improved. The limitation that the fixed weight distribution ignores the difference of different service scenes in the prior art is solved.

Example methods and systems for dynamic site selection in a global server load balancer (GSLB) environment are described. In one example, a computer system may obtain first health information from a first entity and second health information from a second entity. The first health information may be generated based on multiple first traffic flows between (a) multiple first client devices and (b) a first pool of backend servers. The second health information may be generated based on multiple second traffic flows between (a) multiple second client devices and (b) a second pool of backend servers. In response to receiving a request to access the service, the computer may select a selected site based on the first health information and / or second health information. A response may be generated and sent to cause a third client device to access the service by directing a third traffic flow towards the selected site.

The invention discloses an energy consumption optimization method and system for a computing power server, and relates to the technical field of server optimization, and the method comprises the steps: monitoring the operation data of the computing power server in real time, and recognizing the task and load conditions; constructing a load prediction model to predict the future load of the computing power server, and carrying out the energy consumption optimization of the computing power server according to a predicted load scheduling task; continuously optimizing the load prediction model based on a reinforcement learning algorithm; and displaying the load condition of the computing power server in real time, and storing the operation data of the computing power server to a database. The load prediction model is constructed by collecting the processing tasks and the load data of the computing power server to predict the future load, the accuracy of load prediction is improved, task processing scheduling is performed by predicting the future load to achieve load optimization of the computing power server, and load energy consumption adjustment of the processing tasks of the computing power server is effectively achieved.

The invention discloses a computing power server load prediction method and system, and belongs to the field of data processing, the computing power server load prediction method comprises the following steps: collecting historical load data of a computing power server, and preprocessing the historical load data, the preprocessing comprising data cleaning, time window construction, and data set division; constructing an LSTM (Long Short Term Memory) model, and training the LSTM model; predicting a future computing power server load through an LSTM model; compared with the prior art, the method has the beneficial effects that the overall computing resource utilization rate is effectively improved by collecting the historical load data of the computing power server, training the LSTM model, finally predicting the future computing power server load through the LSTM model and allocating the computing resources according to the predicted data; meanwhile, abnormal tasks are recognized, and the abnormal tasks are processed in time when appearing, so that other tasks are prevented from being affected.

The invention discloses a distributed server load balancing system based on multi-modal data fusion. The distributed server load balancing system comprises a multi-modal sensing layer, a fusion analysis layer, an intelligent decision-making layer and an elastic execution layer, the multi-modal sensing layer collects multi-source heterogeneous data; the fusion analysis layer is used for mining an association relationship between data of different dimensions; meanwhile, converting multi-source data into a unified feature vector, and outputting a dynamic load balancing strategy; the intelligent decision-making layer generates an executable load balancing strategy according to the fusion analysis result; and the elastic execution layer converts the strategy generated by the intelligent decision-making layer into an actual action, and deals with load change through an elastic mechanism. According to the method, server performance, network topology and request semantic data are fused, a unified feature tensor is generated by using a cross-modal attention mechanism, and load association is accurately quantified; on the basis of cooperation of a global mixed integer programming model and a local lightweight DQN agent, load migration is completed in a very short time, and the response speed is increased.

Example methods and systems for connection establishment in a global server load balancer (GSLB) environment are described. In one example, a computer system may establish a first connection with a first entity and a second connection with the second entity. The first connection may be established based on first parameter information that includes a shared certificate and a first identifier (ID). The second connection may be established based on second parameter information that includes the shared certificate and a second ID. The shared certificate may be shared by multiple entities that include the first entity and a second entity. In response to receiving a first request, a first response may be generated and sent towards the first entity via the first connection. In response to receiving a second request, a second response may be generated and sent towards the second entity via the second connection.

The invention, which relates to the technical field of information and communication, discloses an intelligent operation and maintenance monitoring system for a machine room, comprising a state monitoring module, an abnormity early warning module, a data acquisition module, a load evaluation module and a control module. The state monitoring module acquires temperature distribution state data information and vibration state data information in real time; the abnormity early warning module performs correlation analysis on the data, judges whether the server is normal or not and sends out an early warning signal; the data acquisition module acquires memory state data information and network transmission state data information of the server after receiving the early warning signal; the load evaluation module analyzes the load condition of the current server in combination with the trained load degree evaluation model; and the control module judges whether the server is in overload operation or not according to the current server load condition, generates a corresponding control instruction and executes the control instruction, so that stable operation and operation and maintenance efficiency of the server are guaranteed.

The invention provides a component-level collaborative editing control method, which is applied to a multi-user collaborative editing control system, and the system comprises a plurality of front-end engines and a back-end server. A front-end engine obtains user editing operation data and a rear-end real-time version snapshot, component structure difference comparison is carried out, and a component operation list is generated; and performing component conflict judgment based on the operation list, merging and editing operation data according to a conflict judgment result and a preset conflict resolution strategy, generating processed version snapshots, sending the processed version snapshots to a rear end, storing the processed version snapshots by a rear-end server, and sending the processed version snapshots to each front-end engine. Compared with the prior art, the method has the advantages that the data transmission quantity and the load of the rear-end server are reduced by a lightweight architecture through the front-end component level difference comparison and the preset conflict resolution strategy, the rear-end server does not need to be frequently requested for conflict resolution and data merging, the response speed and the collaborative editing efficiency are remarkably improved, and the user experience is improved. And better user experience is brought to the user. The system has the same beneficial effects.

The invention relates to the technical field of computers, and discloses a server load balancing method based on a comprehensive score, computer equipment and a computer readable storage medium in order to solve the problem that in an existing server load balancing method, parameters for determining a server operation state score are not comprehensive enough. The method comprises the following steps: acquiring hardware performance parameters, environmental conditions, a current load state, the number of connection requests successfully processed in unit time and historical fault data of each server; determining a comprehensive score of each server according to the hardware performance parameters, the environmental conditions, the current load state, the number of connection requests successfully processed in unit time and historical fault data; and sending the current task to a target server for processing based on the comprehensive score. By adopting the method, the performance of the server is improved, and the method can cope with complex and changeable operating environments.

The invention belongs to the technical field of data processing, particularly relates to a server data processing method and system based on deep learning, and aims to solve the problems of low modal fusion efficiency, poor model adaptability and insufficient resource utilization rate in multi-modal data processing. The method comprises the following steps: firstly, preprocessing mixed to-be-processed data and dividing the mixed to-be-processed data into structured data, image data, text data and time series data subsets; a time-varying neural network topological structure is constructed, network edge weights are dynamically adjusted based on server loads and data features, and preliminary extraction and correlation capture of multi-modal features are achieved; generating a fusion feature vector by using a cross-modal feature extractor of the attention mechanism; a dynamic adaptive data processing module is used for matching the deep learning model for parallel processing; and finally, constructing a reinforcement learning reward function based on the processing efficiency, the resource occupancy rate and the accuracy rate, and carrying out joint iterative optimization on the model.

The invention discloses a server temperature monitoring and regulating method and electronic equipment, and relates to the technical field of servers. According to the scheme, the sensor group comprising a plurality of different types of temperature sensors is used for monitoring the temperature of each component in the server, so that the measurement error is obviously reduced; meanwhile, the working temperature interval of the component compared with the measured temperature value is not a fixed interval, but is dynamically adjusted according to the change of the load in the operation process of the server, so that the optimal heat dissipation effect is realized during temperature regulation and control; when it is confirmed that temperature regulation and control need to be carried out, the target cooling liquid flow is determined through a pre-constructed cooling liquid flow prediction model, the model is specifically constructed based on the average temperature of all parts of the server, the heat dissipation power and the server load, the optimal cooling liquid flow can be predicted to achieve energy consumption minimization of a pump set, heat dissipation energy consumption is effectively reduced, and the service life of the pump set is prolonged. And comprehensive optimization of the heat dissipation effect is realized.

The invention discloses an intelligent water meter recognition system based on a three-layer edge computing architecture, which has the following beneficial effects: (1) based on the three-layer edge computing architecture, the intelligent water meter recognition system gives full play to the advantages of edge computing, the load of a central server is reduced, and the service life of the central server is prolonged; and the service life of the water meter image data acquisition equipment is prolonged. (2) YOLO-METER is improved on the basis of YOLO11 and is used for water meter reading recognition, a FastC3k2 module and an MRFBlock are introduced into the YOLO-METER, and the FastC3k2 module can improve the image feature extraction performance and enhance the low-contrast feature extraction capability of a water meter image by reducing the number of parameters of a model and FLOPs; and the MRFBlock obviously expands the receptive field of the model, enhances the capability of autonomously selecting features, improves multi-scale fusion, and can optimize the feature selection of a digital roller region and improve the positioning capability of a water meter reading region. According to the invention, the calculation cost is reduced while the identification precision is improved.

The invention discloses a data center low-carbon control method and system, and relates to the technical field of data centers, and the method comprises the steps: collecting the environment parameters and equipment operation data of a data center in a hot-summer and warm-winter region in real time based on a multi-mode sensor network; constructing a climate-adaptive space-time diagram convolutional network, taking a server rack, a cooling tower and a power distribution cabinet as nodes, taking a cold and hot flow path as a dynamic weighted edge, embedding climate feature vectors, and generating a thermal load prediction map; a dynamic refrigeration strategy and a server load migration strategy are generated by taking a thermodynamic equation and a fluid dynamic model as strategy optimization boundary conditions through a physical constraint reinforcement learning framework and combining a thermal load prediction map; and based on the dynamic refrigeration strategy, the air conditioning system is controlled to start a hybrid refrigeration mode according to the real-time climate type, the water cooling main cycle and indirect evaporative cooling standby loop are synchronously operated, and the cold charging and discharging logic of the chilled water storage system is switched according to the electricity price time period.

The invention discloses a server cluster load balancing cooperative processing system and method and electronic equipment, and relates to the technical field of servers, and the method comprises the steps: receiving node operation data sent by each edge node in a region; carrying out load prediction in the region according to the node operation data and a pre-trained neural network model to obtain a prediction result; determining a dynamic scheduling weight; solving calculation is carried out according to the multiple pre-stored node mapping schemes, preset solving parameters and a pre-established target function, and a multi-target Pareto solution set is obtained; performing preferential processing according to the Pareto solution set and the dynamic scheduling weight to obtain a target node mapping scheme; and according to the target node mapping scheme, controlling each edge node in the region to execute load balancing cooperative operation. The problem that the load balancing performance of the server cluster is poor is solved, and the server load balancing performance of the server cluster is improved by selecting a proper target node mapping scheme based on the node operation data collected by each edge node of the edge layer.

The invention discloses a state grid cloud-based cloud service deployment method and system, and relates to the field of cloud service deployment, and the method comprises the steps: collecting service types needing to be deployed in a power grid, analyzing the use demands of tenants, and evaluating the server performance demands of the service types; installing a Ceph distributed storage framework on a plurality of storage nodes of the state grid cloud platform; the method comprises the following steps of: disassembling a service category into a plurality of independent micro-services, packaging each micro-service in an independent container through Docker, and performing automatic management on the container by using a Kubernetes container arrangement tool; docker is used for achieving multi-tenant isolation, a dynamic resource scheduling mechanism is introduced, and the number of container instances is adjusted according to changes of server loads and use requirements of the tenants; and testing the multi-tenant micro-service container environment, and performing online deployment on a state grid cloud platform after the test is passed. According to the method, the Docker is adopted to realize multi-tenant isolation, and a dynamic resource scheduling mechanism is introduced, so that efficient, safe and extensible cloud service deployment is realized.

The invention relates to a load balancing algorithm based on deep learning. The load balancing algorithm comprises the following steps: step 1, system initialization and model preparation; 2, receiving and preprocessing a user request; step 3, collecting and analyzing server load data; 4, intelligent load distribution decision making; 5, processing and forwarding the request; step 6, an error processing and retry mechanism; 7, evaluating and optimizing the model and the strategy; 8, the user requests are completed, and the system is in a standby state. According to the invention, through the technical means of deep learning modeling, reinforcement learning mechanism, adaptive queue management and global load optimization, intelligent prediction, dynamic adjustment and global optimization of the server load are realized; and the positive effects of improving the load balancing efficiency, shortening the response time, improving the resource utilization rate, improving the user experience and the like are achieved.

The invention discloses a micro-service deployment and task unloading optimization method for tasks with complex dependency relationships. The method comprises the following steps: a system architecture covers a plurality of base stations, edge servers and user equipment; an application program generated by the user equipment comprises a plurality of micro-service tasks with a complex dependency relationship, and the tasks are unloaded to an edge server to be processed; in a task unloading process, optimizing a task unloading decision in each time slot, and managing a micro-service layer cache by adopting a long-time scale updating strategy; through joint optimization of a micro-service layer caching strategy and a task unloading decision, a system performance problem is constructed into an optimization model; firstly, an unloading path of a task is planned by applying a belief propagation algorithm, and then a caching strategy is trained by means of a deep reinforcement learning model; the problem of complicated task unloading and cache updating in different time scale scenes is solved by utilizing an alternative optimization algorithm, and finally, a trained model is deployed in an edge server, so that the long-term average time delay of the system is minimized, and meanwhile, the load balancing of the server is realized. According to the method, the utilization efficiency of system resources can be remarkably improved, the time delay is reduced, and various application requirements sensitive to the delay are fully met.

The invention discloses a secure cloud workstation and a resource scheduling method suitable for a high-altitude weak network environment, and the secure cloud workstation comprises a virtualization management platform which is used for carrying out the virtualization processing of core computing components of a plurality of servers contained in a back-end server cluster, and forms a plurality of virtual hosts; the front-end thin client only reserves display and input functions, establishes connection with the virtualization management platform, receives a virtual host human-computer interface pushed by any virtual host in the virtualization computing resource pool, and feeds back a user operation instruction to the virtualization management platform; the transmission protocol optimization module is used for integrating graph, input and audio data transmission channels, monitoring the network bandwidth and the delay state of a high-altitude area in real time by adopting a self-adaptive compression algorithm, and dynamically adjusting the transmission image quality and the frame rate of a virtual host human-computer interface; and the resource scheduling module is used for automatically allocating computing resources to the plurality of virtual hosts based on the load data of the virtualization management platform so as to realize cross-server load balancing.