Cloud laboratory architecture design method and system

Abstract

The invention provides a cloud laboratory architecture design method and system, and relates to the technical field of cloud computing and network security interdisciplinary, and the method comprises the steps: building a hybrid cloud architecture laboratory resource pool, and dividing a physical server into computing, storage and network node groups; constructing a dynamic sensing matrix to generate a multi-dimensional resource state map; based on this, adaptive load balancing scheduling is executed to optimize a resource allocation weight and a data transmission path; deploying a multi-level security protection tunnel, generating a dynamic authentication factor by using equipment identification characteristics, and establishing an encryption link rule; constructing an intelligent experiment resource composer, analyzing the resource state atlas to generate a resource association atlas, and optimizing a cross-platform experiment process; realizing a heterogeneous device co-verification mechanism, generating a trusted execution verification code by capturing device interaction characteristics, and synchronizing the trusted execution verification code with a dynamic authentication factor; and deploying a visual monitoring interface to display the resource topological relation and the encrypted link state. The resource utilization rate and the task execution efficiency are improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of the intersection of cloud computing and network security, and in particular, to a method and system for designing a cloud laboratory architecture. Background Art

[0002] In modern scientific research and enterprise environments, with the rapid growth and increasing complexity of data volume, the demand for high-performance computing, large-scale data analysis, and complex experimental tasks is increasing day by day. The method for designing a cloud laboratory architecture has become one of the key means to meet these demands. This architecture needs to be able to efficiently manage heterogeneous hardware resources, including computing node groups, storage node groups, and network node groups, and provide dynamic resource scheduling to optimize resource utilization efficiency. At the same time, in order to support cross-platform experimental processes, the system must have a strong security protection mechanism to ensure the security and integrity of data transmission. In addition, users also need an intuitive way to monitor the running state of the system so as to quickly respond to any potential problems and ensure the continuity and reliability of the experimental process. Therefore, an ideal cloud laboratory architecture should have functions such as real-time monitoring, intelligent scheduling, multi-level security protection, and visual monitoring to cope with complex scientific research and business needs.

[0003] Currently, some advanced cloud laboratory architecture designs adopt an artificial intelligence-based resource scheduling system, which predicts resource requirements through machine learning algorithms and makes dynamic adjustments to improve resource utilization. At the same time, software-defined network technology is used to achieve intelligent management and path optimization of network traffic. In addition, a distributed security authentication system is constructed using blockchain technology to enhance the trust between devices and the security of data transmission. The application of these technologies has significantly improved the automation level and security of the system.

[0004] Although the existing advanced cloud laboratory architecture solutions have certain intelligent and automated capabilities, there are still some key defects. First, in terms of resource scheduling accuracy, although the existing adaptive load balancing scheduling strategy can be adjusted according to real-time data, it fails to fully consider the relevance and dependency relationships between resources, resulting in inaccurate resource allocation and affecting the overall efficiency. Second, although the existing security protection measures adopt encryption tunnel technology, due to the lack of an effective mechanism for capturing and verifying the interaction characteristics between heterogeneous devices, it is difficult to completely prevent potential security threats, especially more obvious during cross-platform operations. Finally, most of the existing visualization tools can only display basic system status information, and the dynamic changes of complex resource topology relationships and encryption link rules are not displayed intuitively enough, limiting the user's ability to quickly locate and solve problems. Summary of the Invention

[0005] The embodiments of the present application provide a method and system for cloud laboratory architecture design to solve the problems of low resource utilization rate and low task execution efficiency in the prior art.

[0006] In a first aspect, the embodiments of the present application provide a method for cloud laboratory architecture design, including: Establish a laboratory resource pool with a hybrid cloud architecture, divide the physical server cluster into a computing node group, a storage node group, and a network node group, and construct a dynamic perception matrix. The dynamic perception matrix generates a multi-dimensional resource status map by collecting the real-time operation load rate of the computing node group, the data access heat of the storage node group, and the link delay coefficient of the network node group; Execute adaptive load balancing scheduling based on the multi-dimensional resource status map. The adaptive load balancing scheduling generates resource allocation weights by fusing resource correlation parameters and dynamically adjusts the data transmission path across node groups; Deploy a multi-level security protection tunnel. The multi-level security protection tunnel generates dynamic authentication factors based on device identification features and establishes an encryption link generation rule based on the data transmission path; Construct an intelligent experimental resource orchestrator. The intelligent experimental resource orchestrator generates a resource association map by parsing the multi-dimensional resource status map and constructs an optimized path for cross-platform experimental processes; Implement a heterogeneous device collaborative verification mechanism. The heterogeneous device collaborative verification mechanism generates a trusted execution verification code by capturing device interaction features and synchronizes the trusted execution verification code to the dynamic authentication factors of the multi-level security protection tunnel; Deploy a visual experimental environment monitoring interface. The visual experimental environment monitoring interface dynamically maps the topological relationship of the resource association map and displays the real-time effective status of the encryption link generation rule.

[0007] Optionally, the executing adaptive load balancing scheduling based on the multi-dimensional resource status map, which generates resource allocation weights by fusing resource correlation parameters and dynamically adjusts the data transmission path across node groups, includes: According to the multi-dimensional resource status map, generate a set of resource correlation parameters through weighted fusion. The set of resource correlation parameters includes dynamically adjusted weight coefficients, and the weight coefficients are dynamically corrected based on historical task execution records and the current experimental task type to generate resource allocation weights; Dynamically optimize the data transmission path across node groups based on the resource allocation weights. By real-time monitoring of the link delay coefficient and data access heat, construct a path selection probability matrix. The path selection probability matrix feeds back and updates the state values of the nodes in the data transmission path according to the task completion time and the number of path hops, and outputs an optimal path set and the corresponding link performance coefficient. The link performance coefficient is associated with the ratio of the remaining bandwidth to the total bandwidth; When it is detected that the real-time operation load rate of the computing node group exceeds the first preset threshold, trigger an elastic contraction strategy, reallocate the resource allocation weights according to the ratio of the number of idle nodes to the total number of nodes, and adjust the concurrency threshold of the optimal path set based on the updated resource allocation weights and the link performance coefficient. When the link performance coefficient is lower than the second preset threshold, trigger a network bypass strategy, switch the data transmission path to a preset backup logical channel and update the data transmission path across the node group.

[0008] Optionally, the dynamically optimizing the data transmission path across node groups based on the resource allocation weights, by real-time monitoring of the link delay coefficient and data access heat, constructing a path selection probability matrix, the path selection probability matrix feeding back and updating the state values of path nodes according to the task completion time and the number of path hops, and outputting an optimal path set and the corresponding link performance coefficient, the link performance coefficient being associated with the ratio of the remaining bandwidth to the total bandwidth, includes: Based on the real-time resource status data of the multi-dimensional resource status map, initialize the initial state value of the data transmission path across node groups. The initial state value is directly proportional to the weighted reciprocal of the link delay coefficient and the data access heat, and real-time monitor the instantaneous delay fluctuation and storage access request distribution of the data transmission path; According to the resource requirement characteristics of the experimental tasks in the task queue, dynamically calculate the path heat weight using the initial state value. The path heat weight is jointly determined by the difference between the task calculation intensity and the real-time operation load rate of the path calculation node group, the data access heat matching degree of the path storage node group, and the adaptation degree of the task network bandwidth requirement to the path remaining bandwidth; Combine the initial state value, the path heat weight, and the historical task execution feedback data to generate a dynamic path selection probability. Based on the dynamic path selection probability, when the task completion time is lower than the preset threshold, positively reinforce the initial state value, otherwise negatively decay it, to generate an updated set of path state values. The historical task execution feedback data includes the normalized ratio of the task completion time to the number of paths; Based on the updated set of path status values, filter the paths that meet the collaborative constraints of the real-time operation load rate, the data access heat, and the link delay coefficient, generate an optimal path set, and calculate a link performance coefficient based on the ratio of the remaining bandwidth of the path to the total bandwidth, the number of paths, and the instantaneous delay fluctuation. The link performance coefficient is dynamically bound to the path status values in the optimal path set.

[0009] Optionally, combine the initial state value, the path heat weight, and the historical task execution feedback data to generate a dynamic path selection probability. Based on the dynamic path selection probability, when the task completion time is lower than a preset threshold, positively reinforce the initial state value; otherwise, perform negative attenuation to generate an updated set of path status values. The historical task execution feedback data includes the normalized ratio of the task completion time to the number of paths, and includes: Generate a path status dynamic baseline parameter based on the initial state value and a preset real-time operation load rate tolerance interval. The path status dynamic baseline parameter dynamically expands and contracts according to the distribution ratio of compute-intensive tasks and storage-intensive tasks in the current experimental task queue, and is inversely correlated with the fluctuation range of the link delay coefficient; Normalize the task completion time and the number of paths in the historical task execution feedback data to generate a path feedback reinforcement factor and a decay factor. Based on the path feedback reinforcement factor and the decay factor, when the task completion time is lower than a preset threshold, calculate a positive reinforcement step size based on the fitness between the task computing intensity and the remaining bandwidth of the path; otherwise, calculate a negative attenuation step size according to the product of the number of paths and the instantaneous delay fluctuation. Perform a bimodal update on the initial state value according to the positive reinforcement step size or the negative attenuation step size. If the path feedback reinforcement factor is higher than the dynamic baseline parameter, use the exponential weighted superposition method to increase the initial state value; if the decay factor triggers the threshold condition, use the piecewise linear decay method to decrease the initial state value to generate an updated set of path status values.

[0010] Optionally, performing a bimodal update on the initial state value according to the positive reinforcement step size or the negative attenuation step size, if the path feedback reinforcement factor is higher than the dynamic baseline parameter, use the exponential weighted superposition method to increase the initial state value, if the decay factor triggers the threshold condition, use the piecewise linear decay method to decrease the initial state value, to generate an updated set of path status values, including: Divide the positive reinforcement interval according to the difference between the path feedback reinforcement factor and the dynamic baseline parameter of the path state. The positive reinforcement interval includes a rapid improvement area, a smooth transition area, and a saturation convergence area. The rapid improvement area corresponds to the scenario where the difference is higher than the first preset threshold, and the exponential weighted superposition method is used to increase the initial state value. The smooth transition area corresponds to the scenario where the difference is between the first preset threshold and the second preset threshold, and the linear weighted superposition method is used to adjust the initial state value of the path. The saturation convergence area corresponds to the scenario where the difference is lower than the second preset threshold, and the current level of the initial state value is maintained; Divide the negative attenuation interval according to the ratio of the attenuation factor to the dynamic baseline parameter of the path state. The negative attenuation interval includes a rapid attenuation area, a smooth attenuation area, and a protection stagnation area. The rapid attenuation area corresponds to the scenario where the ratio is higher than the third preset threshold, and the piecewise linear attenuation method is used to decrease the initial state value. The smooth attenuation area corresponds to the scenario where the ratio is between the third preset threshold and the fourth preset threshold, and the linear attenuation method is used to adjust the initial state value of the path. The protection stagnation area corresponds to the scenario where the ratio is lower than the fourth preset threshold, and the current level of the initial state value is maintained; Based on the positive reinforcement interval and the negative attenuation interval, perform dual-mode update on the initial state value. When the path feedback reinforcement factor is higher than the dynamic baseline parameter of the path state, select the exponential weighted superposition method according to the positive reinforcement interval to increase the initial state value. When the attenuation factor triggers the threshold condition, select the piecewise linear attenuation method according to the negative attenuation interval to decrease the initial state value, and generate an updated set of path state values.

[0011] Optionally, deploy a multi-level security protection tunnel. The multi-level security protection tunnel generates a dynamic authentication factor through device identification features, and establishes an encryption link generation rule based on the data transmission path, including: Generate a set of device identification features based on the device hardware fingerprint and software configuration information of the multi-dimensional resource status map, and encode the set of device identification features through an asymmetric encryption algorithm to generate a device identification ciphertext; Generate a dynamic authentication factor based on the device identification ciphertext and the resource allocation weight, and bind the dynamic authentication factor to the data transmission path through a key derivation function to generate a path-bound authentication token; Construct an encryption link generation rule according to the optimal path set of the data transmission path and the link performance coefficient.

[0012] Optionally, construct an intelligent experimental resource orchestrator. The intelligent experimental resource orchestrator generates a resource association map by parsing the multi-dimensional resource status map, and constructs an optimized path for cross-platform experimental processes, including: Based on the multi-dimensional resource status graph, extract the real-time operation load rate of the computing node group, the data access heat of the storage node group, and the link delay coefficient of the network node group, generate a resource status feature vector, and compress the resource status feature vector through a dimensionality reduction algorithm to generate a resource status feature set; Based on the resource status feature set, construct a resource association graph, and optimize the resource association graph through a graph neural network to generate an optimized resource association graph; According to the task type and resource requirement characteristics of the experimental task queue, construct a cross-platform experimental process model, where the cross-platform experimental process model includes a task sharding strategy, a resource allocation strategy, and a task scheduling strategy; Based on the cross-platform experimental process model and the optimized resource association graph, generate a cross-platform experimental process optimization path.

[0013] In a second aspect, an embodiment of the present application provides a cloud laboratory architecture design system, including: A building module, configured to build a laboratory resource pool with a hybrid cloud architecture, divide a physical server cluster into a computing node group, a storage node group, and a network node group, and construct a dynamic perception matrix, where the dynamic perception matrix generates a multi-dimensional resource status graph by collecting the real-time operation load rate of the computing node group, the data access heat of the storage node group, and the link delay coefficient of the network node group; An execution module, configured to perform adaptive load balancing scheduling based on the multi-dimensional resource status graph, where the adaptive load balancing scheduling generates a resource allocation weight by fusing resource correlation parameters and dynamically adjusts the data transmission path across node groups; A deployment module, configured to deploy a multi-level security protection tunnel, where the multi-level security protection tunnel generates a dynamic authentication factor through device identification features and establishes an encryption link generation rule based on the data transmission path; A generation module, configured to construct an intelligent experimental resource orchestrator, where the intelligent experimental resource orchestrator generates a resource association graph by parsing the multi-dimensional resource status graph and constructs a cross-platform experimental process optimization path; An implementation module, configured to implement a heterogeneous device collaborative verification mechanism, where the heterogeneous device collaborative verification mechanism generates a trusted execution verification code by capturing device interaction features and synchronizes the trusted execution verification code to the dynamic authentication factor of the multi-level security protection tunnel; A mapping module, configured to deploy a visual experimental environment monitoring interface, where the visual experimental environment monitoring interface dynamically maps the topological relationship of the resource association graph and displays the real-time effective status of the encryption link generation rule.

[0014] In a third aspect, an embodiment of the present application provides a computing device, including a processor and a memory. A computer program is stored in the memory, and the processor is configured to run the computer program to execute the method for designing a cloud laboratory architecture according to any one of the first aspect.

[0015] In a fourth aspect, an embodiment of the present application provides a computer storage medium, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the method for designing a cloud laboratory architecture according to any one of the first aspect is implemented.

[0016] In an embodiment of the present application, a laboratory resource pool with a hybrid cloud architecture is established. The physical server cluster is divided into a computing node group, a storage node group, and a network node group, and a dynamic perception matrix is constructed. The dynamic perception matrix generates a multi-dimensional resource status map by collecting the real-time operation load rate of the computing node group, the data access heat of the storage node group, and the link delay coefficient of the network node group; based on the multi-dimensional resource status map, adaptive load balancing scheduling is performed. The adaptive load balancing scheduling generates resource allocation weights by fusing resource correlation parameters and dynamically adjusts the data transmission path across node groups; a multi-level security protection tunnel is deployed. The multi-level security protection tunnel generates a dynamic authentication factor through device identification features and establishes an encryption link generation rule based on the data transmission path; an intelligent experimental resource orchestrator is constructed. The intelligent experimental resource orchestrator generates a resource association map by parsing the multi-dimensional resource status map and constructs an optimized path for cross-platform experimental processes; a heterogeneous device collaborative verification mechanism is implemented. The heterogeneous device collaborative verification mechanism generates a trusted execution verification code by capturing device interaction features and synchronizes the trusted execution verification code to the dynamic authentication factor of the multi-level security protection tunnel; a visual experimental environment monitoring interface is deployed. The visual experimental environment monitoring interface dynamically maps the topological relationship of the resource association map and displays the real-time effective status of the encryption link generation rule.

[0017] The technical solution of the present application has the following beneficial effects: The present application improves resource utilization and task execution efficiency, enhances the security and stability of the system, and at the same time provides an intuitive operation experience through the visual monitoring interface, greatly simplifying the resource management and task scheduling processes in complex environments, promoting the efficient development of scientific research work and the rapid transformation of technical achievements. In addition, the collaborative verification mechanism between heterogeneous devices further ensures the reliability and consistency of the entire system, contributing to the development of cross-platform applications.

[0018] Furthermore, the embodiments of the present application also perform adaptive load balancing scheduling based on the multi-dimensional resource status map, generate a set of resource correlation parameters including dynamically adjusted weight coefficients through weighted fusion, and these weight coefficients are dynamically corrected according to the historical task execution records and the current experimental task type to generate resource allocation weights. Based on this weight, the data transmission paths across node groups are dynamically optimized. By real-time monitoring the link delay coefficient and data access heat, a path selection probability matrix is constructed, and the node state values are updated according to the task completion time and the number of path hops, and the optimal path set and its corresponding link performance coefficient (related to the ratio of the remaining bandwidth to the total bandwidth) are output. When the real-time operation load rate of the computing node group exceeds the first preset threshold, an elastic contraction strategy is triggered, the resource allocation weights are reallocated according to the ratio of the number of idle nodes to the total number of nodes, and the concurrency threshold of the optimal path set is adjusted based on the updated resource allocation weights and link performance coefficients; when the link performance coefficient is lower than the second preset threshold, a network bypass strategy is triggered, and the data transmission path is switched to the backup logical channel and the data transmission path across the node group is updated.

[0019] This method significantly improves the resource utilization efficiency and the optimization effect of the data transmission path. By dynamically adjusting the resource allocation weights and the path selection probability matrix, the system can more accurately match the resource requirements, reduce unnecessary resource waste, and at the same time ensure that high-priority tasks are processed in a timely manner. The elastic contraction strategy and the network bypass strategy further enhance the flexibility and robustness of the system, enabling it to still operate efficiently under high load or low link performance conditions, and avoiding task interruptions caused by resource overload or link failures. Overall, this method effectively improves the response speed, stability, and security of the system, providing a solid technical guarantee for complex experimental tasks.

[0020] These aspects or other aspects of the present application will be more clearly understood in the following description of the embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0022] Figure 1 It is a flowchart of a cloud laboratory architecture design method provided by an embodiment of the present application; Figure 2 It is a schematic structural diagram of a cloud laboratory architecture design system provided by an embodiment of the present application; Figure 3A structural schematic diagram of a computing device provided by an embodiment of the present application. Detailed implementation manners

[0023] In order to enable those skilled in the art to better understand the solutions of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application.

[0024] In some processes described in the specification, claims and the above-mentioned drawings of the present application, a plurality of operations that appear in a specific order are included. However, it should be clearly understood that these operations may not be executed in the order in which they appear herein or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions such as "first" and "second" in this article are used to distinguish different messages, devices, modules, etc., and do not represent a sequence, nor do they limit that "first" and "second" are of different types.

[0025] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present application.

[0026] Figure 1 A flowchart of a cloud laboratory architecture design method provided by an embodiment of the present application is as Figure 1 shown, and the method includes: Step 101: Establish a laboratory resource pool for a hybrid cloud architecture, divide a physical server cluster into a computing node group, a storage node group, and a network node group, and construct a dynamic perception matrix. The dynamic perception matrix generates a multi-dimensional resource status map by collecting the real-time operation load rate of the computing node group, the data access heat of the storage node group, and the link delay coefficient of the network node group; In this step, the hybrid cloud architecture is a cloud computing model that combines the advantages of private clouds and public clouds, allowing data and applications to migrate flexibly between the two. The physical server cluster is divided into a computing node group (responsible for processing computing tasks), a storage node group (for data storage), and a network node group (managing data transmission). The dynamic perception matrix is a real-time monitoring system that generates a multi-dimensional resource status map by collecting key indicators of each node group (such as operation load rate, data access heat, and link delay coefficient).

[0027] In actual operation, first, physical servers are divided into different node groups according to their functions. Then, a dynamic perception matrix is used to continuously monitor the status information of these nodes, and this information is integrated into a multi-dimensional resource status map. This map provides basic data support for subsequent resource scheduling.

[0028] For example, in a bioinformatics research project, researchers need to process a large amount of gene sequencing data. Through a hybrid cloud architecture, they can use the private cloud to process sensitive data and borrow public cloud resources during peak hours. The dynamic perception matrix monitors the status of each node in real time to ensure the effective allocation and efficient utilization of resources.

[0029] Step 102: Perform adaptive load balancing scheduling based on the multi-dimensional resource status map. The adaptive load balancing scheduling generates resource allocation weights by integrating resource correlation parameters and dynamically adjusts the data transmission path across node groups. In this step, the adaptive load balancing scheduling is a method that automatically adjusts resource allocation according to the current state of the system. It integrates resource correlation parameters (such as dependencies between nodes and weight coefficients) to generate resource allocation weights and dynamically adjusts the data transmission path across node groups accordingly. This method can optimize resource utilization, reduce bottlenecks, and improve overall performance.

[0030] In actual operation, using the resource status map generated in step 101, the system dynamically adjusts resource allocation weights according to resource correlation parameters and optimizes the data transmission path. This step ensures that resources can be flexibly allocated according to actual needs, reduces bottlenecks, and improves overall performance.

[0031] For example, in the above bioinformatics project, as the data analysis tasks increase, the system automatically adjusts the resource allocation weights of computing nodes to prioritize high-priority tasks. When a computing node is overloaded, the system will automatically reallocate tasks to other idle nodes to ensure the stable operation of the entire system.

[0032] Step 103: Deploy a multi-level security protection tunnel. The multi-level security protection tunnel generates dynamic authentication factors through device identification features and establishes an encryption link generation rule based on the data transmission path. In this step, the multi-level security protection tunnel is a security mechanism that generates dynamic authentication factors through device identification features and establishes an encryption link generation rule based on the data transmission path. The dynamic authentication factors include timestamps, task identifiers, and hash values of device identification ciphertexts to ensure the security and integrity of data transmission.

[0033] In actual operation, the system generates a dynamic authentication factor using device identification features and establishes an encryption link generation rule in combination with the data transmission path. This step ensures that data will not be tampered with or stolen during transmission, thus guaranteeing the security and integrity of the data.

[0034] For example, in a bioinformatics project, to protect sensitive gene data, the system encrypts all data transmissions using a multi-level security protection tunnel. When transmitting gene sequencing data across multiple data centers, the dynamic authentication factor ensures that only legitimate devices can participate in the data transmission, while the encryption link generation rule prevents illegal access to the data during transmission.

[0035] Step 104: Construct an intelligent experimental resource orchestrator that generates a resource association graph by parsing the multi-dimensional resource status graph and constructs an optimized cross-platform experimental process path; In this step, the intelligent experimental resource orchestrator is a tool that parses the multi-dimensional resource status graph to generate a resource association graph and constructs an optimized cross-platform experimental process path. The resource association graph shows the topological connection relationships and resource dependency weights between nodes, helping to implement a more efficient resource scheduling strategy.

[0036] In actual operation, the intelligent experimental resource orchestrator parses the resource status graph, generates a resource association graph, and constructs an optimized path according to the requirements of the experimental task queue. This step not only considers the physical connections between nodes but also the dependency relationships between resources, thus achieving more efficient resource scheduling and task sharding.

[0037] For example, in a bioinformatics project, the intelligent experimental resource orchestrator decomposes a large task into multiple subtasks according to the requirements of gene data analysis tasks and assigns them to the most suitable nodes according to the resource association graph. Some computationally intensive tasks will be assigned to high-performance computing nodes, while data-intensive tasks will be assigned to nodes with large-capacity storage, thus maximizing resource utilization.

[0038] Step 105: Implement a heterogeneous device collaborative verification mechanism that generates a trusted execution verification code by capturing device interaction features and synchronizes the trusted execution verification code to the dynamic authentication factor of the multi-level security protection tunnel; In this step, the heterogeneous device collaborative verification mechanism generates a trusted execution verification code by capturing device interaction features and synchronizes it to the dynamic authentication factor of the multi-level security protection tunnel. This mechanism ensures the security and consistency of device interactions, especially important during cross-platform operations.

[0039] In actual operation, the system captures the interaction characteristics between devices, generates a trusted execution verification code, and synchronizes it to the dynamic authentication factors of the multi-level security protection tunnel. This step ensures that even during cross-platform operations, potential security threats can be effectively prevented.

[0040] For example, in a bioinformatics project, when multiple heterogeneous devices (such as high-performance computing nodes and storage nodes) jointly process gene data, the heterogeneous device collaborative verification mechanism ensures that all interactions between devices are secure. When data is transmitted from a computing node to a storage node, the trusted execution verification code verifies the identity of the device, preventing unauthorized devices from participating in data transmission.

[0041] Step 106: Deploy a visual experimental environment monitoring interface, which dynamically maps the topological relationship of the resource association graph and displays the real-time effective status of the encrypted link generation rule.

[0042] In this step, the visual experimental environment monitoring interface is a tool that dynamically maps the topological relationship of the resource association graph and displays the real-time effective status of the encrypted link generation rule. It provides an intuitive operation experience and helps users quickly locate and solve problems.

[0043] In actual operation, the visual experimental environment monitoring interface dynamically displays the topological relationship of the resource association graph and the real-time status of the encrypted link, enabling users to clearly understand the operation of the system at a glance. Any anomalies can be quickly discovered and resolved, improving the maintainability and stability of the system.

[0044] For example, in a bioinformatics project, an administrator can use the visual monitoring interface to view the progress and resource usage of gene data analysis tasks in real time. If a node fails or is overloaded, the administrator can immediately take measures, such as reassigning tasks or adjusting resource configuration, to ensure the continuous operation of the system.

[0045] Through the implementation of the above six steps, the system realizes all-round optimization from the establishment of the resource pool to dynamic perception, adaptive load balancing, multi-level security protection, intelligent resource orchestration, heterogeneous device collaborative verification to visual monitoring. This method significantly improves the resource utilization efficiency and the security of data transmission, ensuring the efficient execution of complex experimental tasks and the stability of the system. Especially when dealing with large-scale data analysis and high-performance computing tasks, this method can effectively address various challenges, provide reliable technical support, and ensure the smooth progress of scientific research work and the rapid transformation of technological achievements.

[0046] To address the issues of inaccurate resource allocation and insufficient optimization of data transmission paths, in some embodiments, in step 102, the adaptive load balancing scheduling based on the multi-dimensional resource status map is performed. The adaptive load balancing scheduling generates resource allocation weights by fusing resource correlation parameters and dynamically adjusts the data transmission paths across node groups, including: According to the multi-dimensional resource status map, a set of resource correlation parameters is generated through weighted fusion. The set of resource correlation parameters includes dynamically adjusted weight coefficients, and the weight coefficients are dynamically corrected based on historical task execution records and the current experimental task type to generate resource allocation weights. Based on the resource allocation weights, dynamic optimization is performed on the data transmission paths across node groups. By real-time monitoring of the link delay coefficient and data access heat, a path selection probability matrix is constructed. The path selection probability matrix feeds back and updates the state values of the nodes in the data transmission path according to the task completion time and the number of path hops, and outputs an optimal path set and the corresponding link performance coefficient. The link performance coefficient is associated with the ratio of the remaining bandwidth to the total bandwidth. When it is detected that the real-time operation load rate of the computing node group exceeds a first preset threshold, an elastic contraction strategy is triggered. The resource allocation weights are reallocated according to the ratio of the number of idle nodes to the total number of nodes, and the concurrency threshold of the optimal path set is adjusted based on the updated resource allocation weights and the link performance coefficient. When the link performance coefficient is lower than a second preset threshold, a network bypass strategy is triggered, and the data transmission path is switched to a preset backup logical channel and the data transmission path across the node group is updated.

[0047] In this embodiment, the set of resource correlation parameters is generated by weighted fusion of key metrics in the multi-dimensional resource status map (such as real-time operation load rate, data access heat, and link delay coefficient). These parameters not only reflect the current state of each node but also combine historical task execution records and the current task type to generate dynamically adjusted weight coefficients. The resource allocation weights are used to guide how to allocate computing and storage resources among different node groups to maximize system efficiency. The path selection probability matrix is constructed based on the real-time monitored data, which helps the system select the optimal data transmission path and dynamically update the node state values according to the task completion time and the number of path hops to ensure the efficiency and flexibility of path selection.

[0048] In the embodiments of the present application, first, the system generates a set of resource correlation parameters using a multi-dimensional resource status map, which includes dynamically adjusted weight coefficients. Then, based on these weight coefficients, resource allocation weights are generated and applied to the optimization of data transmission paths across node groups. By real-time monitoring of the link delay coefficient and data access popularity, the system constructs a path selection probability matrix, dynamically updates the node status values, and outputs the optimal path set and its link performance coefficients. When the real-time operation load rate of the computing node group is too high, the system triggers an elastic contraction strategy, reallocates the resource allocation weights, and adjusts the concurrency threshold of the optimal path set. If the link performance coefficient is too low, the network bypass strategy is triggered to switch to the backup logical channel to ensure the stable operation of the system.

[0049] The following is a specific example: In a bioinformatics research project, researchers need to process a large amount of gene sequencing data. To improve resource utilization efficiency and data transmission security, the system adopts the above optional solution. First, the dynamic perception matrix real-time monitors the operation load rate of the computing node group, the data access popularity of the storage node group, and the link delay coefficient of the network node group, and generates a multi-dimensional resource status map. Based on this map, the system generates a set of resource correlation parameters through weighted fusion. These parameters not only consider the current task requirements but also combine historical task execution records to generate dynamically adjusted weight coefficients. Next, the system optimizes the data transmission paths across node groups according to the generated resource allocation weights. For example, when processing large-scale data analysis tasks, the system real-time monitors the link delay coefficient and data access popularity, constructs a path selection probability matrix, and dynamically updates the node status values according to the task completion time and the number of path hops to select the optimal data transmission path. When the load rate of a certain computing node exceeds the preset threshold, the system automatically triggers an elastic contraction strategy, reallocates the resource allocation weights, and adjusts the concurrency threshold of the optimal path set to ensure that the overall performance of the system is not affected by the high load of a single node. In addition, when the link performance coefficient is lower than the preset threshold, the system triggers the network bypass strategy to switch the data transmission path to the preset backup logical channel to ensure the continuity and security of data transmission. In this way, the system not only improves resource utilization efficiency but also enhances the security and stability of data transmission, ensuring the efficient execution of complex experimental tasks.

[0050] In order to further improve the optimization accuracy and dynamic adjustment ability of the cross-node group data transmission path, in some embodiments, in step 202, the dynamic optimization of the cross-node group data transmission path based on the resource allocation weight is performed by monitoring the link delay coefficient and data access heat in real time, constructing a path selection probability matrix, and the path selection probability matrix feeds back and updates the state value of the path node according to the task completion time and the number of path hops, and outputs an optimal path set and the corresponding link performance coefficient, and the link performance coefficient is associated with the ratio of the remaining bandwidth to the total bandwidth, including: Based on the real-time resource status data of the multi-dimensional resource status map, initialize the initial state value of the data transmission path of the cross-node group. The initial state value is directly proportional to the weighted reciprocal of the link delay coefficient and the data access heat, and the instantaneous delay fluctuation and storage access request distribution of the data transmission path are monitored in real time; according to the resource demand characteristics of the experimental tasks in the task queue, use the initial state value to dynamically calculate the path heat weight, and the path heat weight is jointly determined by the difference between the task calculation intensity and the real-time operation load rate of the path calculation node group, the data access heat matching degree of the path storage node group, and the adaptation degree of the task network bandwidth demand and the path remaining bandwidth; combine the initial state value, the path heat weight and the historical task execution feedback data to generate a dynamic path selection probability. Based on the dynamic path selection probability, when the task completion time is lower than the preset threshold, perform positive reinforcement on the initial state value, otherwise perform negative attenuation to generate an updated path state value set, and the historical task execution feedback data includes the normalized ratio of the task completion time to the number of paths; according to the updated path state value set, screen the paths that meet the collaborative constraints of the real-time operation load rate, the data access heat and the link delay coefficient to generate an optimal path set, and calculate the link performance coefficient based on the ratio of the path remaining bandwidth to the total bandwidth, the number of paths and the instantaneous delay fluctuation, and the link performance coefficient is dynamically bound to the path state value in the optimal path set.

[0051] In this embodiment, the initial state value is initialized based on the real-time resource status data of the multi-dimensional resource status map, and is directly proportional to the weighted reciprocal of the link delay coefficient and the data access heat, and is used to evaluate the initial health status of each path. The path heat weight comprehensively considers the task calculation intensity, the real-time operation load rate of the path calculation node group, the data access heat of the path storage node group, and the adaptation degree of the task network bandwidth demand and the path remaining bandwidth, and reflects the applicability of the path in the current task environment. The dynamic path selection probability is generated by combining the initial state value, the path heat weight and the historical task execution feedback data, and guides the system to select the optimal data transmission path. The link performance coefficient is calculated based on the ratio of the path remaining bandwidth to the total bandwidth, the number of paths and the instantaneous delay fluctuation, and is used to evaluate the actual performance of the path.

[0052] In the embodiments of the present application, first, the system initializes the initial state values of the data transmission paths across node groups using a multi-dimensional resource status map, and monitors the instantaneous delay fluctuations and storage access request distributions in real time. Then, according to the resource requirement characteristics of the experimental tasks in the task queue, the path heat weights are dynamically calculated. Next, combining the initial state values, path heat weights, and historical task execution feedback data, dynamic path selection probabilities are generated, and positive or negative adjustments are made according to the task completion time to generate an updated set of path state values. Finally, the system filters out the optimal path set based on the updated set of path state values and calculates the link performance coefficient to ensure the efficiency and stability of the data transmission path.

[0053] The following is a specific embodiment: In a bioinformatics research project, researchers need to process a large amount of gene sequencing data. To improve the optimization accuracy and dynamic adjustment ability of the data transmission path, the system adopts the above optional solution. First, the system initializes the initial state values of the data transmission paths across node groups based on a multi-dimensional resource status map, and these values reflect the link delay coefficients and data access heat of each path. For example, when processing large-scale data analysis tasks, the system monitors the link delay fluctuations and storage access request distributions in real time to ensure the accuracy of the initial state values. Then, the system dynamically calculates the path heat weights according to the resource requirement characteristics of the experimental tasks in the task queue. For example, for compute-intensive tasks, the system preferentially selects those paths where the real-time operation load rate of the compute node group is relatively low and the data access heat matching degree of the storage node group is relatively high. In addition, the system also considers the adaptability between the task network bandwidth requirement and the remaining bandwidth of the path to ensure the efficiency of data transmission. Subsequently, the system combines the initial state values, path heat weights, and historical task execution feedback data to generate dynamic path selection probabilities. For example, if the completion time of a task on a certain path is lower than the preset threshold, the system will positively reinforce the initial state value of this path, and vice versa for negative attenuation, thus continuously optimizing the path selection strategy. Finally, the system filters out the paths that meet the collaborative constraints of real-time operation load rate, data access heat, and link delay coefficient based on the updated set of path state values to generate the optimal path set. Based on the ratio of the remaining bandwidth of the path to the total bandwidth, the number of paths, and the instantaneous delay fluctuations, the system calculates the link performance coefficient to ensure the efficiency and stability of the data transmission path. In this way, the system not only improves the resource utilization efficiency, but also enhances the security and stability of data transmission, ensuring the efficient execution of complex experimental tasks.

[0054] In order to further improve the dynamic adjustment ability of the path selection probability and optimize the update mechanism of the path state value, in some embodiments, in step 303, the initial state value, the path heat weight, and the historical task execution feedback data are combined to generate a dynamic path selection probability. Based on the dynamic path selection probability, when the task completion time is lower than a preset threshold, the initial state value is positively reinforced, and vice versa for negative attenuation, to generate an updated set of path state values. The historical task execution feedback data includes the normalized ratio of the task completion time to the number of paths, and further includes: Based on the initial state value and the preset real-time operation load rate tolerance interval, a path state dynamic baseline parameter is generated. The path state dynamic baseline parameter dynamically scales according to the distribution ratio of compute-intensive tasks and storage-intensive tasks in the current experimental task queue, and is inversely correlated with the fluctuation range of the link delay coefficient; the task completion time and the number of paths in the historical task execution feedback data are normalized to generate a path feedback reinforcement factor and a decay factor. Based on the path feedback reinforcement factor and the decay factor, when the task completion time is lower than a preset threshold, a positive reinforcement step size is calculated based on the adaptability between the task computing intensity and the remaining bandwidth of the path, and vice versa, a negative attenuation step size is calculated according to the product of the number of paths and the instantaneous delay fluctuation; the initial state value is bimodally updated according to the positive reinforcement step size or the negative attenuation step size. If the path feedback reinforcement factor is higher than the dynamic baseline parameter, the initial state value is enhanced using an exponentially weighted superposition method. If the decay factor triggers a threshold condition, the initial state value is reduced using a piecewise linear attenuation method to generate an updated set of path state values.

[0055] In this embodiment, the path state dynamic baseline parameter is generated based on the initial state value and the real-time operation load rate tolerance interval, reflecting the benchmark performance level of the path under different task types, and dynamically adjusted according to the ratio of compute-intensive and storage-intensive tasks. The path feedback reinforcement factor and the decay factor are generated by normalizing the historical task execution feedback data (task completion time and number of paths), used to evaluate the performance of the path, and guide subsequent positive reinforcement or negative attenuation. The positive reinforcement step size and the negative attenuation step size are calculated based on the task computing intensity, the remaining bandwidth of the path, as well as the number of paths and the instantaneous delay fluctuation, used to adjust the initial state value of the path to optimize path selection.

[0056] In the embodiments of the present application, first, the system generates dynamic baseline parameters of the path state based on the initial state value and the real-time operation load rate tolerance interval, and this parameter will be dynamically adjusted according to the distribution ratio of different types of tasks in the current experimental task queue. Then, the system normalizes the task completion time and the number of paths in the historical task execution feedback data to generate a path feedback reinforcement factor and a decay factor. Next, the system calculates a positive reinforcement step or a negative decay step based on these factors. If the task completion time is lower than the preset threshold, the system calculates the positive reinforcement step based on the adaptability between the task calculation intensity and the remaining bandwidth of the path; otherwise, it calculates the negative decay step based on the product of the number of paths and the instantaneous delay fluctuation. Finally, the system performs a bimodal update on the initial state value according to the calculated step, and uses the exponential weighted superposition or the piecewise linear decay method to generate the updated set of path state values.

[0057] The following is a specific embodiment: In a bioinformatics research project, researchers need to efficiently process a large amount of gene sequencing data. To optimize path selection and state update, the system adopts the above optional solution. First, the system generates dynamic baseline parameters of the path state based on the initial state value and the real-time operation load rate tolerance interval to ensure that the path can adapt to different types of computing and storage requirements. For example, when processing large-scale data analysis tasks, the system dynamically adjusts the path state baseline according to the ratio of compute-intensive and storage-intensive tasks in the task queue. Next, the system normalizes the task completion time and the number of paths in the historical task execution feedback data to generate a path feedback reinforcement factor and a decay factor. For example, for those tasks that are completed in a short time, the system calculates the positive reinforcement step to increase the state value of the relevant path; while for those tasks with a long completion time, it calculates the negative decay step to reduce its state value. Finally, the system performs a bimodal update on the initial state value according to the calculated step. If the path feedback reinforcement factor is higher than the dynamic baseline parameter, the system uses the exponential weighted superposition method to increase the state value of the path; if the decay factor triggers the threshold condition, it uses the piecewise linear decay method to reduce the state value of the path. In this way, the system not only improves the accuracy and flexibility of path selection, but also enhances the stability and efficiency of the system, ensuring the efficient execution of complex experimental tasks.

[0058] To further improve the flexibility and accuracy of the path state value update mechanism, in some embodiments, for the bimodal update of the initial state value according to the positive reinforcement step or the negative decay step in step 403, if the path feedback reinforcement factor is higher than the dynamic baseline parameter, the exponential weighted superposition method is used to increase the initial state value, and if the decay factor triggers the threshold condition, the piecewise linear decay method is used to reduce the initial state value to generate the updated set of path state values, and it further includes: According to the difference between the path feedback reinforcement factor and the dynamic baseline parameter of the path state, a positive reinforcement interval is divided. The positive reinforcement interval includes a rapid improvement area, a smooth transition area, and a saturation convergence area. The rapid improvement area corresponds to the scenario where the difference is higher than the first preset threshold, and the exponential weighted superposition method is used to increase the initial state value. The smooth transition area corresponds to the scenario where the difference is between the first preset threshold and the second preset threshold, and the linear weighted superposition method is used to adjust the initial path state value. The saturation convergence area corresponds to the scenario where the difference is lower than the second preset threshold, and the current level of the initial state value is maintained. According to the ratio of the attenuation factor to the dynamic baseline parameter of the path state, a negative attenuation interval is divided. The negative attenuation interval includes a rapid attenuation area, a smooth attenuation area, and a protection stagnation area. The rapid attenuation area corresponds to the scenario where the ratio is higher than the third preset threshold, and the piecewise linear attenuation method is used to decrease the initial state value. The smooth attenuation area corresponds to the scenario where the ratio is between the third preset threshold and the fourth preset threshold, and the linear attenuation method is used to adjust the initial path state value. The protection stagnation area corresponds to the scenario where the ratio is lower than the fourth preset threshold, and the current level of the initial state value is maintained. Based on the positive reinforcement interval and the negative attenuation interval, a dual-mode update of the initial state value is performed. When the path feedback reinforcement factor is higher than the dynamic baseline parameter of the path state, the exponential weighted superposition method is selected according to the positive reinforcement interval to increase the initial state value. When the attenuation factor triggers the threshold condition, the piecewise linear attenuation method is selected according to the negative attenuation interval to decrease the initial state value, and a set of updated path state values is generated.

[0059] In this embodiment, the positive reinforcement interval is divided based on the difference between the path feedback reinforcement factor and the dynamic baseline parameter of the path state, and is divided into a rapid improvement area, a smooth transition area, and a saturation convergence area. Different weight adjustment strategies (such as exponential weighted superposition, linear weighted superposition, or remaining unchanged) are used in each area to optimize the path state value. The negative attenuation interval is divided based on the ratio of the attenuation factor to the dynamic baseline parameter of the path state, and is divided into a rapid attenuation area, a smooth attenuation area, and a protection stagnation area. Different attenuation strategies (such as piecewise linear attenuation, linear attenuation, or remaining unchanged) are used in each area to meet the adjustment requirements of the path state value in different situations. The dual-mode update combines the strategies of the positive reinforcement interval and the negative attenuation interval to flexibly adjust the initial state value, ensuring that the path state value can be optimally adjusted according to the actual situation.

[0060] In the embodiments of the present application, first, the system divides a positive reinforcement interval, including a rapid improvement area, a smooth transition area, and a saturation convergence area, according to the difference between the path feedback reinforcement factor and the dynamic baseline parameter of the path state. For the rapid improvement area, the system uses an exponentially weighted superposition method to increase the initial state value; for the smooth transition area, a linearly weighted superposition method is used for adjustment; and for the saturation convergence area, the current state value is maintained unchanged. Next, the system divides a negative attenuation interval, including a rapid attenuation area, a smooth attenuation area, and a protection stagnation area, according to the ratio of the attenuation factor to the dynamic baseline parameter of the path state. For the rapid attenuation area, the system uses a piecewise linear attenuation method to reduce the initial state value; for the smooth attenuation area, a linear attenuation method is used for adjustment; and for the protection stagnation area, the current state value is maintained unchanged. Finally, the system performs a dual-mode update on the initial state value based on the strategies of the positive reinforcement interval and the negative attenuation interval. If the path feedback reinforcement factor is higher than the dynamic baseline parameter of the path state, an appropriate weight adjustment strategy is selected according to the positive reinforcement interval to increase the initial state value; if the attenuation factor triggers the threshold condition, an appropriate attenuation strategy is selected according to the negative attenuation interval to reduce the initial state value, thereby generating an updated set of path state values.

[0061] The following is a specific example: In a bioinformatics research project, researchers need to process a large amount of gene sequencing data. To optimize the update mechanism of the path state value, the system adopts the above optional solution. First, the system divides a positive reinforcement interval according to the difference between the path feedback reinforcement factor and the dynamic baseline parameter of the path state. For example, when processing high-priority tasks, if the path feedback reinforcement factor is much higher than the dynamic baseline parameter, the system classifies this path as the rapid improvement area and uses the exponentially weighted superposition method to significantly increase its initial state value, ensuring that high-priority tasks can obtain more resource support. Next, the system divides a negative attenuation interval according to the ratio of the attenuation factor to the dynamic baseline parameter of the path state. For example, for those tasks with a long completion time and low resource utilization rate, the system classifies them as the rapid attenuation area and uses the piecewise linear attenuation method to reduce their initial state values, so as to release more resources for other tasks. Finally, the system performs a dual-mode update on the initial state values of all paths based on the strategies of the positive reinforcement interval and the negative attenuation interval. In this way, the system not only improves the accuracy and flexibility of path selection, but also enhances the stability and efficiency of the system, ensuring the efficient execution of complex experimental tasks. For example, a computationally intensive task has a significant increase in the state value in the rapid improvement area, while another storage-intensive task obtains a moderate adjustment in the smooth transition area, ensuring the balance and efficient operation of the entire system.

[0062] To further improve the security of data transmission and the reliability of path selection, in some embodiments, the multi-level security protection tunnel deployed in step 103 generates a dynamic authentication factor through device identification features and establishes an encryption link generation rule based on the data transmission path, including: Generate a set of device identification features based on the device hardware fingerprint and software configuration information of the multi-dimensional resource status map, encode the set of device identification features through an asymmetric encryption algorithm to generate a device identification ciphertext; generate a dynamic authentication factor based on the device identification ciphertext and the resource allocation weight, and bind the dynamic authentication factor to the data transmission path through a key derivation function to generate a path-bound authentication token; construct an encryption link generation rule according to the optimal path set of the data transmission path and the link performance coefficient.

[0063] In this embodiment, the set of device identification features includes device hardware fingerprints (such as CPU serial numbers, hard disk serial numbers, etc.) and software configuration information (such as operating system versions, installed software lists, etc.), which are used to uniquely identify each device. The set is encoded through an asymmetric encryption algorithm to generate a device identification ciphertext to ensure the security of the device identity. The dynamic authentication factor is generated based on the device identification ciphertext and the resource allocation weight, and is combined with the data transmission path through a key derivation function to form a path-bound authentication token, ensuring that data can only be accessed by legitimate devices during transmission. The encryption link generation rule is constructed according to the optimal path set of the data transmission path and the link performance coefficient, ensuring that data is not only efficient but also secure during transmission.

[0064] In the embodiments of the present application, first, the system generates a set of device identification features based on the device hardware fingerprint and software configuration information in the multi-dimensional resource status map, and encodes it through an asymmetric encryption algorithm to generate a device identification ciphertext. Then, the system generates a dynamic authentication factor based on the generated device identification ciphertext and the resource allocation weight of the current task, and binds the factor to the data transmission path through a key derivation function to generate a path-bound authentication token. Next, the system constructs an encryption link generation rule according to the optimal path set of the data transmission path and the link performance coefficient to ensure that the data transmission path is both efficient and secure. In this way, the system can effectively prevent unauthorized devices from accessing sensitive data and ensure the security and integrity of data transmission.

[0065] The following is a specific embodiment: In a bioinformatics research project, researchers need to process a large amount of gene sequencing data, and it is crucial to ensure the security and efficiency of data transmission. The system adopts the above optional solutions to enhance security. First, based on the device hardware fingerprints (such as CPU serial numbers and hard disk serial numbers) and software configuration information (such as operating system versions and installed software lists) in the physical server cluster, the system generates a set of device identification features, and encodes them through an asymmetric encryption algorithm to generate device identification ciphertext. This step ensures that the identity of each device can be uniquely identified and is difficult to forge. Then, based on the generated device identification ciphertext and the resource allocation weights of the current experimental task, the system generates a dynamic authentication factor, and binds it to the data transmission path through a key derivation function to generate a path-bound authentication token. This measure ensures that only verified legitimate devices can participate in data transmission, preventing unauthorized access. Finally, the system constructs an encryption link generation rule according to the optimal path set of the data transmission path and the link performance coefficient. For example, when processing large-scale data analysis tasks, the system will select those paths with higher remaining bandwidth and lower latency, and set appropriate encryption protocols for them to ensure the efficiency and security of data transmission. Through this multi-level security protection mechanism, researchers can process sensitive gene data with confidence without worrying about the risk of data leakage or tampering.

[0066] In order to further improve the resource utilization efficiency and the optimization ability of the experimental process, in some embodiments, the intelligent experimental resource orchestrator described in step 104, the intelligent experimental resource orchestrator generates a resource association graph by parsing the multi-dimensional resource status graph, and constructs a cross-platform experimental process optimization path, including: Based on the multi-dimensional resource status graph, extract the real-time operation load rate of the computing node group, the data access heat of the storage node group, and the link delay coefficient of the network node group, generate a resource status feature vector, compress the resource status feature vector through a dimensionality reduction algorithm to generate a resource status feature set; based on the resource status feature set, construct a resource association graph, and optimize the resource association graph through a graph neural network to generate an optimized resource association graph; according to the task type and resource requirement characteristics of the experimental task queue, construct a cross-platform experimental process model, the cross-platform experimental process model includes a task sharding strategy, a resource allocation strategy, and a task scheduling strategy; based on the cross-platform experimental process model and the optimized resource association graph, generate a cross-platform experimental process optimization path.

[0067] In this embodiment, the resource status feature vector includes information such as the real-time operation load rate of the computing node group, the data access heat of the storage node group, and the link delay coefficient of the network node group, which is used to describe the current status of each node. These data are compressed through a dimensionality reduction algorithm to generate a resource status feature set for more efficient processing and analysis. The resource association graph is constructed based on the resource status feature set, showing the topological connection relationship and its dependence weights among the computing nodes, storage nodes, and network nodes. By optimizing this graph through a graph neural network, the complex relationship between nodes can be more accurately reflected, providing a basis for resource allocation. The cross-platform experimental process model is constructed according to the task type and resource requirement characteristics of the experimental task queue, including a task sharding strategy (decomposing large tasks into multiple subtasks), a resource allocation strategy (reasonably allocating computing, storage, and network resources), and a task scheduling strategy (determining the task execution order) to ensure the efficient operation of the experimental process.

[0068] In the embodiment of the present application, the system first extracts the real-time operation load rate of the computing node group, the data access heat of the storage node group, and the link delay coefficient of the network node group based on the multi-dimensional resource status graph, generates a resource status feature vector, and compresses it through a dimensionality reduction algorithm to generate a resource status feature set. Then, a resource association graph is constructed based on the resource status feature set, and this graph is optimized through a graph neural network to generate an optimized resource association graph. Next, according to the task type and resource requirement characteristics of the experimental task queue, a cross-platform experimental process model is constructed, and this model includes a task sharding strategy, a resource allocation strategy, and a task scheduling strategy. Finally, based on the cross-platform experimental process model and the optimized resource association graph, a cross-platform experimental process optimization path is generated to ensure the high efficiency of the experimental process and the maximization of resource utilization.

[0069] The following is a specific embodiment: In a bioinformatics research project, researchers need to process a large amount of gene sequencing data. Ensuring the efficient utilization of resources and optimizing the experimental process is crucial. The system first extracts the real-time computing load rate of the computing node group, the data access heat of the storage node group, and the link delay coefficient of the network node group based on the device hardware fingerprints and software configuration information in the physical server cluster, generates a resource status feature vector, and compresses it through a dimensionality reduction algorithm to generate a resource status feature set. Then, based on the generated resource status feature set, a resource association graph is constructed, and the graph is optimized through a graph neural network to generate an optimized resource association graph. Subsequently, according to the task type and resource requirement characteristics of the experimental task queue, the system constructs a cross-platform experimental process model. When processing large-scale data analysis tasks, the large task is decomposed into multiple sub-tasks using the task sharding strategy, and the computing, storage, and network resources are reasonably allocated according to the optimized resource association graph. At the same time, a task scheduling strategy is formulated to ensure that high-priority tasks are processed first. Finally, based on the cross-platform experimental process model and the optimized resource association graph, the system generates a cross-platform experimental process optimization path, selects the optimal combination of computing nodes, storage nodes, and network nodes, sets appropriate resource allocation strategies and task scheduling strategies, ensures the efficient operation of the entire experimental process and the maximization of resource utilization, enables researchers to efficiently process complex gene data, and improves the efficiency and quality of scientific research work.

[0070] Figure 2 The following is a schematic structural diagram of a cloud laboratory architecture design system provided by an embodiment of the present application, as Figure 2 shown. The system includes: A building module 21, configured to build a laboratory resource pool with a hybrid cloud architecture, divide the physical server cluster into a computing node group, a storage node group, and a network node group, and construct a dynamic perception matrix. The dynamic perception matrix generates a multi-dimensional resource status graph by collecting the real-time computing load rate of the computing node group, the data access heat of the storage node group, and the link delay coefficient of the network node group; An execution module 22, configured to perform adaptive load balancing scheduling based on the multi-dimensional resource status graph. The adaptive load balancing scheduling generates a resource allocation weight by fusing resource correlation parameters and dynamically adjusts the data transmission path across node groups; A deployment module 23, configured to deploy a multi-level security protection tunnel. The multi-level security protection tunnel generates a dynamic authentication factor based on device identification features and establishes an encryption link generation rule based on the data transmission path; A generation module 24, configured to construct an intelligent experimental resource orchestrator. The intelligent experimental resource orchestrator generates a resource association graph by parsing the multi-dimensional resource status graph and constructs a cross-platform experimental process optimization path; Implementation module 25, which is used to implement the heterogeneous device collaborative verification mechanism. The heterogeneous device collaborative verification mechanism generates a trusted execution verification code by capturing device interaction features and synchronizes the trusted execution verification code to the dynamic authentication factor of the multi-level security protection tunnel. Mapping module 26, which is used to deploy a visual experimental environment monitoring interface. The visual experimental environment monitoring interface dynamically maps the topological relationship of the resource association graph and displays the real-time effective status of the encryption link generation rule.

[0071] Figure 2 The described cloud laboratory architecture design system can execute Figure 1 The cloud laboratory architecture design method described in the illustrated embodiment, and its implementation principle and technical effects will not be elaborated further. For the cloud laboratory architecture design system in the above embodiment, the specific ways for each module and unit to execute operations have been described in detail in the embodiment related to the method, and will not be elaborated here.

[0072] In a possible design, Figure 2 The cloud laboratory architecture design system of the illustrated embodiment can be implemented as a computing device, such as Figure 3 As shown, the computing device can include a storage component 31 and a processing component 32; The storage component 31 stores one or more computer instructions, where the one or more computer instructions are called and executed by the processing component 32.

[0073] The processing component 32 above Figure 1 The cloud laboratory architecture design method of the above embodiment.

[0074] Among them, the processing component 32 can include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component can also be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components for executing the above method.

[0075] The storage component 31 is configured to store various types of data to support operations on the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0076] Of course, the computing device must also necessarily include other components, such as input / output interfaces, display components, communication components, etc.

[0077] The input / output interface provides an interface between the processing component and the peripheral interface module, and the above-mentioned peripheral interface module can be an output device, an input device, etc.

[0078] The communication component is configured to facilitate communication between the computing device and other devices in a wired or wireless manner, etc.

[0079] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. At this time, the computing device can refer to a cloud server, and the above-mentioned processing component, storage component, etc. can be basic server resources leased or purchased from a cloud computing platform.

[0080] The embodiment of the present application also provides a computer storage medium storing a computer program, and when the computer program is executed by a computer, it can implement the above Figure 1 a cloud laboratory architecture design method shown in the above embodiments.

[0081] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the above-described systems, devices, and units can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0082] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative labor.

[0083] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0084] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A cloud laboratory architecture design method, characterized in that: include: Establish a laboratory resource pool of hybrid cloud architecture, divide the physical server cluster into computing node group, storage node group and network node group, and construct a dynamic perception matrix, which generates a multi-dimensional resource status map by collecting the real-time computing load rate of the computing node group, the data access heat of the storage node group and the link delay coefficient of the network node group; Based on the multi-dimensional resource status graph, an adaptive load balancing scheduling is performed, wherein the adaptive load balancing scheduling generates a resource allocation weight by fusing resource association parameters and dynamically adjusts a data transmission path across node groups; Deploy a multi-level security protection tunnel, the multi-level security protection tunnel generates a dynamic authentication factor through device identification characteristics, and establishes an encryption link generation rule based on the data transmission path; Constructing an intelligent experiment resource orchestrator, which generates a resource association map by parsing the multi-dimensional resource status map and constructing a cross-platform experiment process optimization path; Implementing a heterogeneous device collaborative verification mechanism, the heterogeneous device collaborative verification mechanism generates a trusted execution verification code by capturing device interaction features, and synchronizing the trusted execution verification code to the dynamic authentication factor of the multi-level security protection tunnel; Deploy a visual experiment environment monitoring interface, which dynamically maps the topological relationship of the resource association map and displays the real-time effectiveness status of the encryption link generation rules.

2. The method according to claim 1, characterized in that The adaptive load balancing scheduling is performed based on the multi-dimensional resource state map, wherein the adaptive load balancing scheduling generates resource allocation weights by fusing resource association parameters and dynamically adjusts data transmission paths across node groups, including: According to the multi-dimensional resource state map, a resource association parameter set is generated through weighted fusion, wherein the resource association parameter set includes a dynamically adjusted weight coefficient, and the weight coefficient is dynamically modified based on historical task execution records and current experimental task types to generate a resource allocation weight; Based on the resource allocation weight, the data transmission path across the node group is dynamically optimized, and a path selection probability matrix is ​​constructed by real-time monitoring of the link delay coefficient and the data access heat. The path selection probability matrix updates the state value of the node in the data transmission path according to the task completion time and the number of path hops, and outputs the optimal path set and the corresponding link performance coefficient. The link performance coefficient is associated with the ratio of the remaining bandwidth to the total bandwidth; When it is detected that the real-time computing load rate of the computing node group exceeds a first preset threshold, the elastic contraction strategy is triggered, the resource allocation weights are reallocated according to the ratio of the number of idle nodes to the total number of nodes, and the concurrency threshold of the optimal path set is adjusted based on the updated resource allocation weights and the link performance coefficient. When the link performance coefficient is lower than a second preset threshold, the network bypass strategy is triggered, the data transmission path is switched to a preset backup logical channel and the data transmission path across the node group is updated.

3. The method according to claim 2, characterized in that The method dynamically optimizes the data transmission path across the node group based on the resource allocation weight, builds a path selection probability matrix by real-time monitoring of the link delay coefficient and the data access heat, and feeds back the state value of the path node according to the task completion time and the number of path hops, outputs the optimal path set and the corresponding link performance coefficient, and the link performance coefficient is associated with the ratio of the remaining bandwidth to the total bandwidth, including: Based on the real-time resource status data of the multi-dimensional resource status map, the initial state value of the data transmission path across the node group is initialized, the initial state value is proportional to the link delay coefficient and the weighted inverse of the data access heat, and the instantaneous delay fluctuation and storage access request distribution of the data transmission path are monitored in real time; According to the resource demand characteristics of the experimental tasks in the task queue, the path heat weight is dynamically calculated using the initial state value, and the path heat weight is jointly determined by the difference between the task calculation intensity and the real-time computing load rate of the path calculation node group, the data access heat matching degree of the path storage node group, and the adaptability of the task network bandwidth demand and the path remaining bandwidth; Combine the initial state value, the path heat weight and historical task execution feedback data to generate a dynamic path selection probability. Based on the dynamic path selection probability, when the task completion time is lower than a preset threshold, the initial state value is positively reinforced, otherwise it is negatively attenuated to generate an updated path state value set. The historical task execution feedback data includes a normalized ratio of the task completion time to the number of paths. According to the updated path state value set, the paths that meet the collaborative constraints of the real-time computing load rate, the data access heat and the link delay coefficient are screened to generate an optimal path set. Based on the ratio of the path remaining bandwidth to the total bandwidth, the number of paths and the instantaneous delay fluctuation, the link performance coefficient is calculated. The link performance coefficient is dynamically bound to the path state value in the optimal path set.

4. The method according to claim 3, characterized in that: The dynamic path selection probability is generated by combining the initial state value, the path heat weight and the historical task execution feedback data. Based on the dynamic path selection probability, when the task completion time is lower than the preset threshold, the initial state value is positively strengthened, otherwise it is negatively attenuated to generate an updated path state value set. The historical task execution feedback data includes the normalized ratio of the task completion time to the number of paths, including: Based on the initial state value and a preset tolerance interval of the real-time computing load rate, a path state dynamic baseline parameter is generated, the path state dynamic baseline parameter is dynamically scaled according to the distribution ratio of computing-intensive tasks and storage-intensive tasks in the current experimental task queue, and is inversely correlated with the fluctuation range of the link delay coefficient; Normalizing the task completion time and the number of paths in the historical task execution feedback data to generate a path feedback reinforcement factor and an attenuation factor. Based on the path feedback reinforcement factor and the attenuation factor, when the task completion time is lower than a preset threshold, a positive reinforcement step length is calculated based on the adaptability of the task calculation intensity and the path remaining bandwidth. Otherwise, a negative attenuation step length is calculated based on the product of the number of paths and the instantaneous delay fluctuation. The initial state value is bimodally updated according to the positive reinforcement step size or the negative attenuation step size. If the path feedback reinforcement factor is higher than the dynamic baseline parameter, the initial state value is increased by exponential weighted superposition. If the attenuation factor triggers a threshold condition, the initial state value is reduced by piecewise linear attenuation to generate an updated path state value set.

5. The method according to claim 4, characterized in that The bimodal updating of the initial state value according to the positive reinforcement step length or the negative attenuation step length is performed. If the path feedback reinforcement factor is higher than the dynamic baseline parameter, the initial state value is increased by exponential weighted superposition. If the attenuation factor triggers a threshold condition, the initial state value is reduced by piecewise linear attenuation. An updated path state value set is generated, including: According to the difference between the path feedback reinforcement factor and the path state dynamic baseline parameter, a positive reinforcement interval is divided, and the positive reinforcement interval includes a rapid improvement area, a smooth transition area and a saturated convergence area. The rapid improvement area corresponds to a scenario where the difference is higher than a first preset threshold, and an exponential weighted superposition method is used to improve the initial state value. The smooth transition area corresponds to a scenario where the difference is between the first preset threshold and the second preset threshold, and a linear weighted superposition method is used to adjust the path initial state value. The saturated convergence area corresponds to a scenario where the difference is lower than the second preset threshold, and the current level of the initial state value is maintained; According to the ratio of the attenuation factor to the dynamic baseline parameter of the path state, a negative attenuation interval is divided, and the negative attenuation interval includes a rapid attenuation area, a stable attenuation area, and a protection stagnation area. The rapid attenuation area corresponds to a scenario where the ratio is higher than a third preset threshold, and the initial state value is reduced by a piecewise linear attenuation method. The stable attenuation area corresponds to a scenario where the ratio is between the third preset threshold and the fourth preset threshold, and the path initial state value is adjusted by a linear attenuation method. The protection stagnation area corresponds to a scenario where the ratio is lower than the fourth preset threshold, and the current level of the initial state value is maintained; Based on the positive reinforcement interval and the negative attenuation interval, the initial state value is bimodally updated. When the path feedback reinforcement factor is higher than the path state dynamic baseline parameter, the exponential weighted superposition method is selected according to the positive reinforcement interval to increase the initial state value. When the attenuation factor triggers the threshold condition, the piecewise linear attenuation method is selected according to the negative attenuation interval to reduce the initial state value, thereby generating an updated path state value set.

6. The method according to claim 2, characterized in that The deploying of a multi-level security protection tunnel, wherein the multi-level security protection tunnel generates a dynamic authentication factor through a device identification feature and establishes an encryption link generation rule based on the data transmission path, includes: Based on the device hardware fingerprint and software configuration information of the multi-dimensional resource state map, a device identification feature set is generated, and the device identification feature set is encoded by an asymmetric encryption algorithm to generate a device identification ciphertext; Generate a dynamic authentication factor based on the device identification ciphertext and the resource allocation weight, bind the dynamic authentication factor to the data transmission path through a key derivation function, and generate a path-bound authentication token; An encryption link generation rule is constructed based on the optimal path set of the data transmission path and the link performance coefficient.

7. The method according to claim 1, characterized in that The intelligent experiment resource orchestrator is constructed, wherein the intelligent experiment resource orchestrator generates a resource association map by parsing the multi-dimensional resource status map and constructs a cross-platform experiment process optimization path, including: Based on the multi-dimensional resource state graph, extract the real-time computing load rate of the computing node group, the data access heat of the storage node group, and the link delay coefficient of the network node group to generate a resource state feature vector, compress the resource state feature vector through a dimensionality reduction algorithm, and generate a resource state feature set; Based on the resource state feature set, a resource association graph is constructed, and the resource association graph is optimized through a graph neural network to generate an optimized resource association graph; According to the task type and resource demand characteristics of the experimental task queue, a cross-platform experimental process model is constructed, wherein the cross-platform experimental process model includes a task slicing strategy, a resource allocation strategy, and a task scheduling strategy; Based on the cross-platform experiment process model and the optimized resource association map, a cross-platform experiment process optimization path is generated.

8. A cloud laboratory architecture design system, characterized in that: include: Establish a module for establishing a laboratory resource pool of a hybrid cloud architecture, divide the physical server cluster into a computing node group, a storage node group and a network node group, and construct a dynamic perception matrix, which generates a multi-dimensional resource status map by collecting the real-time computing load rate of the computing node group, the data access heat of the storage node group and the link delay coefficient of the network node group; An execution module, configured to execute adaptive load balancing scheduling based on the multi-dimensional resource status graph, wherein the adaptive load balancing scheduling generates resource allocation weights by integrating resource correlation parameters and dynamically adjusts data transmission paths across node groups; A deployment module, used to deploy a multi-level security protection tunnel, wherein the multi-level security protection tunnel generates a dynamic authentication factor through device identification features and establishes an encryption link generation rule based on the data transmission path; A generation module, used to construct an intelligent experiment resource orchestrator, which generates a resource association map by parsing the multi-dimensional resource status map and constructs a cross-platform experiment process optimization path; An implementation module, used to implement a heterogeneous device collaborative verification mechanism, wherein the heterogeneous device collaborative verification mechanism generates a trusted execution verification code by capturing device interaction features, and synchronizes the trusted execution verification code to a dynamic authentication factor of the multi-level security protection tunnel; A mapping module is used to deploy a visual experimental environment monitoring interface, which dynamically maps the topological relationship of the resource association map and displays the real-time effectiveness status of the encryption link generation rules.

9. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a cloud laboratory architecture design method as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a computer, a cloud laboratory architecture design method as described in any one of claims 1 to 7 is implemented.

Images