A method and device for quickly generating a large-scale network scene of a network target range

Abstract

This invention discloses a method and apparatus for rapidly generating large-scale network scenarios in a network range. During network scenario rendering, the invention parses the atomic instructions required for topology instantiation and, through the combined effect of a hierarchical scheduling strategy and a dependency-based parallel processing mechanism, prioritizes critical tasks, fully utilizes the parallel computing capabilities between nodes, and achieves efficient resource utilization. It can efficiently process and analyze network topology and system configuration data, and can issue commands for building large-scale network environments in a short time, thereby enabling efficient and rapid construction of virtual nodes. It can also quickly adjust and expand network scenarios according to different training or testing needs.

Description

Technical Field

[0001] This invention relates to the field of network security testing technology, and in particular to a method and apparatus for rapidly generating large-scale network scenarios in a network test range. Background Technology

[0002] In the current context of increasingly important network technology and information security, network ranges, as key platforms for conducting network security experiments and tests, are rapidly expanding in scale. This rapid expansion brings a significant problem: as network ranges grow in size, the number of network elements increases, and the configuration complexity rises, more time and resources are required to build experimental network environments. The construction time of target networks is becoming increasingly longer, thus extending the preparation cycle for experiments and testing and affecting the timeliness of test results. Summary of the Invention

[0003] This invention provides a method and apparatus for rapidly generating large-scale network scenarios in network test ranges, in order to solve the problem of low testing efficiency caused by the high dependence of existing embedded software testing on physical hardware.

[0004] In a first aspect, the present invention provides a method for rapidly generating large-scale network scenarios in a network test range, comprising: drawing a network scenario based on the simulation training or testing requirements of a task to be processed, and parsing the topology of the network scenario to obtain the set of atomic instructions to be executed for topology instantiation; generating a task to be deployed based on the dependencies between the atomic instructions in the set of atomic instructions and a preset concurrent execution strategy; dynamically adjusting and deploying the number of scheduling centers and regional scheduling nodes according to the task to be deployed, and controlling the deployed scheduling centers and regional scheduling nodes to perform network scenario instantiation on a simulation cloud platform through a hierarchical scheduling strategy to obtain the required network test range, so as to use the network test range to perform simulation training or testing on the task to be processed; wherein, the scheduling center and the regional scheduling node are a hierarchical distributed architecture, and the concurrent execution strategy is to determine the number of atomic instructions executed in parallel by the scheduling center and the regional scheduling node according to the size of the thread pool of the scheduling center and the regional scheduling node, and the hierarchical scheduling strategy includes: proximity storage rules, load balancing principles, and burst response rules.

[0005] Optionally, the proximity-based storage rule is based on the target images processed by the computing nodes in the simulation cloud platform, and sends the atomic instructions involved in the tasks of the same target image to the same computing node; The load balancing principle intelligently distributes different types of atomic instructions to different scheduling centers and regional scheduling nodes in the cluster through load balancing. The emergency response rules are based on the current actual needs to dynamically deploy the scheduling center and regional scheduling nodes, so as to meet the high load requirements through the dynamic expansion of the underlying cloud platform resources.

[0006] Optionally, the hierarchical scheduling strategy further includes priority processing rules; The priority processing rule is to set the priority of atomic instructions according to the needs of the task and process the atomic instructions according to the priority.

[0007] Optionally, the process of parsing the topology of the network scenario to obtain the atomic instruction set required for topology instantiation includes: parsing the topology of the network scenario, decomposing instructions using an incremental instruction decomposition mechanism, and obtaining the atomic instruction set required for topology instantiation.

[0008] Optionally, the step of controlling the deployed scheduling center and regional scheduling nodes to instantiate the network scenario on the simulation cloud platform through a hierarchical scheduling strategy includes: On the deployed scheduling center and regional scheduling nodes, instructions are distributed and buffered layer by layer according to the hierarchical scheduling strategy in order to instantiate network scenarios on the simulation cloud platform.

[0009] Optionally, the method further includes: setting the atomic instructions with dependency requirements as the main thread, and setting the atomic instructions that the main thread depends on as child threads, so that the main thread can ensure the execution effectiveness of its own thread by monitoring the completion status of the child threads in real time.

[0010] Optionally, after completing the instruction task, each computing node publishes the task completion status and execution status; the top-level scheduling center subscribes to the task execution status published by each computing node, obtains the task execution status of all computing nodes in real time, and after collecting the task execution status of all computing nodes, it summarizes and analyzes the data, and by comparing the task completion status of each computing node, it can comprehensively control the construction progress and status of the entire network topology.

[0011] Optionally, during the resource recycling phase, a power outage is performed, and the released task-related computing and network resources are temporarily stored in a recycling pool; the resources in the recycling pool are recycled in batches according to a preset period using a polling method.

[0012] In a second aspect, the present invention provides a large-scale network scene rapid generation apparatus for implementing the network range described in any of the above methods, the apparatus comprising: The first processing unit is used to draw the network scene according to the simulation training or testing requirements of the task to be processed, and parse the topology of the network scene to obtain the set of atomic instructions to be executed for topology instantiation. The second processing unit is used to generate and distribute tasks based on the dependencies between atomic instructions in the atomic instruction set and a preset concurrent execution strategy. The third processing unit is used to dynamically adjust and deploy the number of scheduling centers and regional scheduling nodes according to the issued tasks, and to control the deployed scheduling centers and regional scheduling nodes to instantiate network scenarios on the simulation cloud platform through a hierarchical scheduling strategy to obtain the required network target range, so as to use the network target range to perform simulation training or testing on the task to be processed; wherein, the scheduling center and the regional scheduling node are hierarchical distributed architectures, and the concurrent execution strategy determines the number of atomic instructions executed in parallel by the scheduling center and the regional scheduling node based on the size of the thread pool of the scheduling center and the regional scheduling node. The hierarchical scheduling strategy includes: proximity storage rules, load balancing principles, and burst response rules.

[0013] Thirdly, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method for rapidly generating large-scale network scenarios in any of the above-described network ranges.

[0014] The beneficial effects of this invention are as follows: This invention parses the atomic instructions required for topology instantiation during network scene rendering. Through the combined effect of a hierarchical scheduling strategy and a dependency-based parallel processing mechanism, it prioritizes critical tasks, fully utilizes the parallel computing capabilities between nodes, and achieves efficient resource utilization. It can efficiently process and analyze network topology and system configuration data, and can issue commands to build large-scale network environments in a short time. This enables efficient and rapid construction of virtual nodes and allows for quick adjustment and expansion of network scenes according to different training or testing needs.

[0015] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0016] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 This is a flowchart illustrating a method for rapidly generating large-scale network scenarios in a network range, as provided in an embodiment of the present invention. Figure 2This is a comparative diagram of existing atomic instruction splitting and incremental atomic instruction splitting in this embodiment of the invention; Figure 3 This is a schematic diagram of the distributed architecture deployment and hierarchical scheduling provided in the embodiments of the present invention; Figure 4 This is a schematic diagram of the dependency-based parallel processing mechanism provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of the principle architecture of the method and device for rapidly generating large-scale network scenes provided in the embodiments of the present invention; Figure 6 This is a schematic diagram of the structure of a large-scale network scene rapid generation device for a network range provided in an embodiment of the present invention. Detailed Implementation

[0017] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not limit the scope of the invention.

[0018] The increasing scale, number of network elements, and configuration complexity of existing network test ranges have led to greater time and resource demands for building experimental network environments. Target network construction times are becoming increasingly long, and the frequent creation and destruction of tasks within network test ranges has become a significant efficiency bottleneck. Each creation of a new scenario or destruction of an old one requires substantial configuration and cleanup work, consuming significant time and increasing operational complexity and the likelihood of errors. This time waste is particularly pronounced during large-scale, high-frequency testing, severely limiting the efficiency and application scope of network test ranges. To address these issues, this invention provides a method for rapidly generating large-scale network scenarios in network test ranges. (See [link to relevant documentation]). Figure 1 The method includes: S101. Draw the network scene according to the simulation training or testing requirements of the task to be processed, and parse the topology of the network scene to obtain the set of atomic instructions to be executed for topology instantiation. Specifically, this invention employs an incremental instruction splitting mechanism during network scene rendering, extracting the instructions required for topology instantiation at the initial rendering stage. Furthermore, this incremental instruction splitting mechanism allows for rapid response to on-the-spot debugging needs, unlike existing implementations that only begin parsing the entire network topology and extracting the necessary instructions at the start of network scene instantiation. In large-scale network scenarios with numerous nodes, this step is time-consuming, taking 5-10 minutes in scenarios with hundreds of thousands of nodes. This invention eliminates this time consumption, thereby improving testing efficiency.

[0019] Understandably, this invention uses an incremental instruction splitting design to decompose complex tasks into multiple smaller tasks. Combined with a distributed architecture deployment, these smaller tasks can be executed simultaneously on multiple nodes, thereby improving the overall processing capacity.

[0020] S102. Generate and distribute tasks based on the dependencies between atomic instructions in the atomic instruction set and the preset concurrent execution strategy; That is, in this embodiment of the invention, the atomic instructions that have dependent requirements are set as the main thread, and the atomic instructions that the main thread depends on are set as child threads. By setting this up, the main thread can ensure the execution effectiveness of its own thread by monitoring the completion status of the child threads in real time.

[0021] In this embodiment of the invention, the concurrent execution strategy determines the number of atomic instructions that the scheduling center and the regional scheduling nodes execute in parallel based on the size of the thread pools of the scheduling center and the regional scheduling nodes. It is understood that the method described in this embodiment of the invention executes the decomposed atomic instructions concurrently through a dependency-based parallel processing mechanism and sends tasks to the scheduling center node. The number of instructions executed simultaneously can be determined by the thread pool size, and this mechanism can meet the needs of rapid construction in large-scale scenarios.

[0022] S103. Dynamically adjust and deploy the number of scheduling centers and regional scheduling nodes according to the issued task, and control the deployed scheduling centers and regional scheduling nodes to instantiate network scenarios on the simulation cloud platform through a hierarchical scheduling strategy to obtain the required network target range, so as to use the network target range to perform simulation training or testing on the task to be processed. Specifically, in this embodiment of the invention, a hierarchical scheduling center set and a regional scheduling set are set up. Each scheduling center set and regional scheduling set includes multiple nodes. On the deployed scheduling center and regional scheduling nodes, instructions are distributed and buffered layer by layer according to the hierarchical scheduling strategy, so as to instantiate the network scenario on the simulation cloud platform.

[0023] In this embodiment of the invention, the scheduling center and the regional scheduling nodes are in a hierarchical distributed architecture. The concurrent execution strategy determines the number of atomic instructions executed in parallel by the scheduling center and the regional scheduling nodes based on the size of their thread pools. The hierarchical scheduling strategy includes: proximity storage rules, load balancing principles, and burst response rules.

[0024] Specifically, in this embodiment of the invention, the proximity-based storage rule is based on the target images processed by the computing nodes in the simulation cloud platform, and the atomic instructions involved in the tasks of the same target image are sent to the same computing node; the load balancing principle is to intelligently distribute different types of atomic instructions to different scheduling centers and regional scheduling nodes in the cluster through load balancing; the emergency response rule is to dynamically deploy the scheduling center and regional scheduling nodes according to the current actual needs, so as to meet the high load demand through the dynamic expansion of the underlying cloud platform resources.

[0025] Of course, in specific implementation, those skilled in the art can also set hierarchical scheduling strategies, including priority processing rules and other rules, according to actual needs, so as to set the priority of atomic instructions according to the needs of the task and process atomic instructions according to priority. In this way, on the basis of instruction parallelism, priority instructions are processed in advance to maximize test efficiency.

[0026] In summary, the method described in this invention, at the virtualization implementation level, receives tasks instantiated from network scenarios through distributed architecture deployment and hierarchical scheduling strategies. It dynamically expands the number of nodes within different service clusters using a distributed architecture to meet large-scale demands. The hierarchical architecture deployment includes a scheduling center cluster, regional scheduling clusters, and a simulation cloud platform. Newly deployed regional scheduling clusters can manage the simulation cloud platform and newly deployed scheduling center clusters, enabling dynamic expansion of underlying cloud platform resources to meet large-scale needs. Furthermore, at the virtualization implementation level, during scheduling, proximity storage rules prioritize the distribution of processing instructions to corresponding computing nodes, improving the efficiency of network scenario construction.

[0027] Meanwhile, in this embodiment of the invention, during the execution of network scenario instantiation instructions, real-time monitoring of tasks is achieved through a distributed task monitoring and status feedback mechanism. After receiving an instruction, the regional scheduling node publishes the task status changes of the instruction, and the scheduling center subscribes to the task status change information published by the regional scheduling node. The scheduling center node provides real-time task execution status information to the network scenario instantiation execution, supporting its high-concurrency invocation.

[0028] Understandably, the distributed architecture design of this invention allows the system to flexibly add or remove nodes as needed to handle tasks of different scales. Furthermore, through the combined effect of hierarchical scheduling strategies and dependency-based parallel processing mechanisms, critical tasks are prioritized, fully utilizing the parallel computing capabilities among nodes to achieve efficient resource utilization. Additionally, by using proximity-based storage rules, data transfer is reduced, improving processing speed.

[0029] Specifically, in this embodiment of the invention, after each computing node completes its task, it publishes the task completion status and execution status. The top-level scheduling center subscribes to the task execution status published by each computing node, obtains the task execution status of all computing nodes in real time, and summarizes and analyzes the collected task execution statuses. By comparing the task completion status of each computing node, the system can comprehensively control the construction progress and status of the entire network topology. A distributed task monitoring and status feedback mechanism provides the system with real-time operational status information.

[0030] Furthermore, in the resource recycling phase, this embodiment of the invention performs a power outage and temporarily stores the released task-related computing and network resources in a recycling pool; the resources in the recycling pool are recycled in batches according to a preset period through a polling method.

[0031] In other words, this invention performs resource reclamation through a hierarchical asynchronous reclamation mechanism. First, the simulation objects instantiated from the network scenario are powered off and placed in a reclamation pool for resource recycling. This hierarchical asynchronous reclamation mechanism ensures that subsequent tasks can acquire the necessary resources in a timely manner.

[0032] The following will combine Figures 2-5 The method described in the embodiments of the present invention will be explained and illustrated in detail through a specific example: To efficiently generate large-scale network scenarios for network test ranges, this invention provides a method for rapidly generating large-scale network scenarios for network test ranges, the method comprising: Step 1: Perform incremental instruction splitting In network test ranges, the drawing and parsing of the target network topology is a crucial step in the simulation process. Traditional holistic parsing methods, after topology drawing, break down the entire scene's data structure into atomic simulation commands and send them to the simulation cloud platform all at once. As the complexity and scale of the network topology increase, the parsing and splitting time also increases significantly. To address this issue, this invention designs and implements an incremental splitting command method, such as... Figure 2 As shown, in each step of drawing the network topology, the system performs partial parsing and splitting in real time, distributing the originally centralized parsing and splitting tasks across various stages of the drawing process. The key to this approach is real-time feedback and step-by-step processing, making the parsing and splitting of large-scale, complex network topologies faster and more efficient. Furthermore, the incremental splitting instruction method of this invention makes the design and implementation of ad-hoc debugging easier. Through real-time parsing and splitting and the issuance of atomic instructions, the system can quickly respond to and adjust the network topology to meet the needs of ad-hoc debugging. This mechanism improves the flexibility of debugging and shortens the issuance and execution time of debugging instructions.

[0033] Step 2: Distribute instructions and tasks using a distributed architecture and a hierarchical scheduling strategy. The method described in this invention breaks down and parses complex network topologies into individual atomic instructions. A distributed architecture is used to deploy the scheduling module, such as... Figure 3 As shown, a unified scheduling cluster is formed, and these atomic instructions are then distributed to the simulation cloud platform. Specifically, this invention can intelligently distribute different types of instructions to different processing nodes in the cluster through load balancing, avoiding single-point overload problems. It can also be dynamically expanded according to actual needs to increase processing capacity to cope with sudden high load problems. And when some nodes in the cluster fail, the system can self-detect the problem and eliminate the faulty node, ensuring that the remaining nodes can continue to operate normally.

[0034] Furthermore, this invention can employ a hierarchical scheduling strategy to distribute atomic instructions. This scheduling method effectively smooths out peak loads from large-scale concurrent requests through multi-layered, ordered processing queues. By distributing and buffering instructions at different levels, request pressure can be mitigated, preventing sudden surges in requests from causing excessive impact on the system. This distributed architecture deployment and hierarchical scheduling optimize the request processing flow and enhance the stability and reliability of the system.

[0035] Step 3: Execute instruction tasks using a dependency-based parallel processing mechanism. The network topology processing is broken down into multiple atomic instruction tasks, which are designed to be dispatched in parallel to the simulation cloud platform via different threads. This design allows multiple instructions to execute simultaneously, improving the efficiency of instruction task execution; it also identifies and manages the dependencies between instruction tasks. Instructions with dependencies are designated as the main thread, such as establishing connections, while their dependent instructions, such as creating virtual nodes and network nodes, are started as child threads. This management method ensures the execution order of tasks. Furthermore, the main thread ensures the effectiveness of its own thread by monitoring the completion status of child threads in real time. This synchronization mechanism guarantees that the main thread can only continue or terminate after all dependent child thread instructions have been completed, thus ensuring the satisfaction of dependencies from the program's execution mechanism. See details. Figure 4 This design improves the efficiency and accuracy of network scenario generation by splitting instruction tasks, managing dependencies, and implementing thread synchronization.

[0036] Step 4: Optimize the execution of instruction tasks through the nearest storage rule. The core idea of ​​proximity storage is to reduce data transmission, thereby improving overall processing efficiency. The creation of virtual nodes depends on a specific target image. If a computing node processing a certain instruction has already synchronized with that target image, the scheduling system will try to distribute the task of that virtual node to that node, effectively reducing unnecessary data migration. The use of proximity storage rules in the system further optimizes the instruction distribution process and improves the efficiency of network scenario construction.

[0037] Step 5: Implement real-time monitoring of tasks through a distributed task monitoring and status feedback mechanism. After completing its task, each node publishes its task completion status and execution status. This information includes the task's progress, results, and other relevant data. The top-level scheduler subscribes to the task execution status published by each node, obtaining real-time information on the task execution status of all nodes. This subscription mechanism allows the top-level scheduler to maintain real-time tracking of the task execution status of each node. After collecting the task execution status of all nodes, the top-level scheduler summarizes and analyzes the data. By comparing the task completion status of each node, a comprehensive understanding of the construction progress and status of the entire network topology can be obtained. This summary is also used to monitor anomalies or delays in task execution and take necessary compensatory measures in a timely manner. The distributed task monitoring and status feedback mechanism enables the system to monitor the task execution status of each node in real time, thereby ensuring the smooth construction of the network topology.

[0038] Step 6: Perform resource reclamation through a tiered asynchronous reclamation mechanism. During the initial phase of resource reclamation, the system performs a power outage to ensure that the reclamation process does not adversely affect running simulation tasks. This step releases computing and network resources associated with the simulation tasks.

[0039] Simulation objects created by network topology splitting commands are placed in a dedicated recycling pool after a power outage. This recycling pool acts as a temporary storage area for managing resources awaiting recycling. Unlike traditional one-time recycling methods, this invention employs a background polling approach to gradually recycle resources. The system periodically checks the resources in the recycling pool in the background and recycles them in batches, rather than recycling all resources at once. Furthermore, the entire recycling process in this invention is asynchronous, ensuring that the recycling operation does not block the execution of other simulation tasks.

[0040] See Figure 5 This invention enables the rapid, stable, and efficient generation of required content in large-scale network scenarios.

[0041] Accordingly, embodiments of the present invention also provide a large-scale network scene rapid generation apparatus for network ranges used to implement any of the methods described above, see [link to relevant documentation]. Figure 6 The device includes: The first processing unit is used to draw the network scene according to the simulation training or testing requirements of the task to be processed, and parse the topology of the network scene to obtain the set of atomic instructions to be executed for topology instantiation. Understandably, this invention uses an incremental instruction splitting design to decompose complex tasks into multiple smaller tasks. Combined with a distributed architecture deployment, these smaller tasks can be executed simultaneously on multiple nodes, thereby improving the overall processing capacity.

[0042] The second processing unit is used to generate and distribute tasks based on the dependencies between atomic instructions in the atomic instruction set and a preset concurrent execution strategy. Specifically, the method described in this embodiment of the invention uses a dependency-based parallel processing mechanism to concurrently execute the decomposed atomic instructions and send tasks to the scheduling center node. The number of instructions executed simultaneously can be determined by the thread pool size, and this mechanism can meet the needs of rapid construction in large-scale scenarios.

[0043] The third processing unit is used to dynamically adjust and deploy the number of scheduling centers and regional scheduling nodes according to the issued task, and control the deployed scheduling centers and regional scheduling nodes to instantiate network scenarios on the simulation cloud platform through a hierarchical scheduling strategy to obtain the required network target range, so as to use the network target range to perform simulation training or testing on the task to be processed. Specifically, in this embodiment of the invention, a hierarchical scheduling center set and a regional scheduling set are set up. Each scheduling center set and regional scheduling set includes multiple nodes. On the deployed scheduling center and regional scheduling nodes, instructions are distributed and buffered layer by layer according to the hierarchical scheduling strategy, so as to instantiate the network scenario on the simulation cloud platform.

[0044] The method described in this invention, at the virtualization implementation level, receives tasks instantiated from network scenarios through distributed architecture deployment and hierarchical scheduling strategies. It dynamically expands the number of nodes within different service clusters using a distributed architecture to meet large-scale demands. The hierarchical architecture deployment includes a scheduling center cluster, regional scheduling clusters, and a simulation cloud platform. Newly deployed regional scheduling clusters can manage the simulation cloud platform and newly deployed scheduling center clusters, enabling dynamic expansion of underlying cloud platform resources to meet large-scale requirements. Furthermore, at the virtualization implementation level, during scheduling, proximity-based storage rules prioritize the distribution of processing instructions to corresponding computing nodes, improving the efficiency of network scenario construction.

[0045] Meanwhile, in this embodiment of the invention, during the execution of network scenario instantiation instructions, real-time monitoring of tasks is achieved through a distributed task monitoring and status feedback mechanism. After receiving an instruction, the regional scheduling node publishes the task status changes of the instruction, and the scheduling center subscribes to the task status change information published by the regional scheduling node. The scheduling center node provides real-time task execution status information to the network scenario instantiation execution, supporting its high-concurrency invocation.

[0046] Understandably, the distributed architecture design of this invention allows the system to flexibly add or remove nodes as needed to handle tasks of different scales. Furthermore, through the combined effect of hierarchical scheduling strategies and dependency-based parallel processing mechanisms, critical tasks are prioritized, fully utilizing the parallel computing capabilities among nodes to achieve efficient resource utilization. Additionally, by using proximity-based storage rules, data transfer is reduced, improving processing speed.

[0047] The scheduling center and the regional scheduling nodes form a hierarchical distributed architecture. The concurrent execution strategy determines the number of atomic instructions that the scheduling center and the regional scheduling nodes can execute in parallel based on the size of their thread pools. The hierarchical scheduling strategy includes: proximity storage rules, load balancing principles, and burst response rules.

[0048] In addition, in specific implementation, the embodiments of the present invention may also set up a resource recycling unit, which performs power-off processing during the resource recycling phase and temporarily stores the released task-related computing and network resources in the recycling pool; and recycles the resources in the recycling pool in batches according to a preset period by polling.

[0049] In other words, this invention performs resource reclamation through a hierarchical asynchronous reclamation mechanism. First, the simulation objects instantiated from the network scenario are powered off and placed in a reclamation pool for resource recycling. This hierarchical asynchronous reclamation mechanism ensures that subsequent tasks can acquire the necessary resources in a timely manner.

[0050] This invention also provides a computer-readable storage medium storing a computer program. When executed by a processor, the program implements any of the hardware-in-the-loop simulation methods based on digital models in this invention's method embodiments. For details, please refer to the method embodiments of this invention; further elaboration is not provided here.

[0051] The relevant content of the device embodiment and storage medium embodiment of the present invention can be understood by referring to the method embodiment of the present invention, and will not be discussed in detail here.

[0052] Although preferred embodiments of the invention have been disclosed for illustrative purposes, those skilled in the art will recognize that various modifications, additions, and substitutions are possible, and therefore the scope of the invention should not be limited to the embodiments described above.

Claims

1. A method for rapidly generating large-scale network scenarios in a network test range, characterized in that, The method includes: Based on the simulation training or testing requirements of the task to be processed, the network scene is drawn, and the topology of the network scene is parsed to obtain the set of atomic instructions to be executed for topology instantiation. The task is generated and distributed based on the dependencies between atomic instructions in the atomic instruction set and the preset concurrent execution strategy. The number of scheduling centers and regional scheduling nodes is dynamically adjusted and deployed according to the issued task. The deployed scheduling centers and regional scheduling nodes are controlled by a hierarchical scheduling strategy to instantiate network scenarios on the simulation cloud platform to obtain the required network range, so as to use the network range to perform simulation training or testing on the task to be processed. The scheduling center and the regional scheduling nodes form a hierarchical distributed architecture. The concurrent execution strategy determines the number of atomic instructions executed in parallel by the scheduling center and the regional scheduling nodes based on the size of their thread pools. The hierarchical scheduling strategy includes: proximity storage rules, load balancing principles, and burst response rules.

2. The method according to claim 1, characterized in that, The analysis of the network topology yields the set of atomic instructions required for topology instantiation, including: The network topology is analyzed, and the instructions are decomposed using an incremental instruction decomposition mechanism to obtain the atomic instruction set required for topology instantiation.

3. The method according to claim 1, characterized in that, The proximity-based storage rule is based on the target images processed by the computing nodes in the simulation cloud platform, and sends the atomic instructions involved in the tasks of the same target image to the same computing node. The load balancing principle intelligently distributes different types of atomic instructions to different scheduling centers and regional scheduling nodes in the cluster through load balancing. The emergency response rules are based on the current actual needs to dynamically deploy the scheduling center and regional scheduling nodes, so as to meet the high load requirements through the dynamic expansion of the underlying cloud platform resources.

4. The method according to claim 3, characterized in that, The hierarchical scheduling strategy also includes priority processing rules; The priority processing rule is to set the priority of atomic instructions according to the needs of the task and process the atomic instructions according to the priority.

5. The method according to claim 3, characterized in that, The process of instantiating network scenarios on the simulation cloud platform by controlling the deployed scheduling center and regional scheduling nodes through a hierarchical scheduling strategy includes: On the deployed scheduling center and regional scheduling nodes, instructions are distributed and buffered layer by layer according to the hierarchical scheduling strategy in order to instantiate network scenarios on the simulation cloud platform.

6. The method according to claim 5, characterized in that, The method further includes: The atomic instructions that have dependencies are set as the main thread, and the atomic instructions that the main thread depends on are set as child threads. By setting this up, the main thread can ensure the effectiveness of its own thread execution by monitoring the completion status of the child threads in real time.

7. The method according to claim 3, characterized in that, After completing the instruction task, each computing node publishes the task completion status and execution status. The top-level scheduling center subscribes to the task execution status published by each computing node, obtains the task execution status of all computing nodes in real time, and summarizes and analyzes the collected task execution status of all computing nodes. By comparing the task completion status of each computing node, it can comprehensively control the construction progress and status of the entire network topology.

8. The method according to any one of claims 1-7, characterized in that, During the resource recycling phase, a power outage is performed, and the released task-related computing and network resources are temporarily stored in the recycling pool. Resources in the recycling pool are recycled in batches according to a preset period and through a polling method.

9. A device for rapidly generating large-scale network scenes for implementing the network range described in any one of claims 1-8, characterized in that, The device includes: The first processing unit is used to draw the network scene according to the simulation training or testing requirements of the task to be processed, and parse the topology of the network scene to obtain the set of atomic instructions to be executed for topology instantiation. The second processing unit is used to generate and distribute tasks based on the dependencies between atomic instructions in the atomic instruction set and a preset concurrent execution strategy. The third processing unit is used to dynamically adjust and deploy the number of scheduling centers and regional scheduling nodes according to the issued tasks, and to control the deployed scheduling centers and regional scheduling nodes to instantiate network scenarios on the simulation cloud platform through a hierarchical scheduling strategy to obtain the required network target range, so as to use the network target range to perform simulation training or testing on the task to be processed; wherein, the scheduling center and the regional scheduling node are hierarchical distributed architectures, and the concurrent execution strategy determines the number of atomic instructions executed in parallel by the scheduling center and the regional scheduling node based on the size of the thread pool of the scheduling center and the regional scheduling node. The hierarchical scheduling strategy includes: proximity storage rules, load balancing principles, and burst response rules.

10. A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method for rapidly generating large-scale network scenarios for a network range according to any one of claims 1-8.

Images