Multi-dimensional performance evaluation method and simulation system for large-scale distributed space systems

Abstract

This invention discloses a multi-dimensional performance evaluation method and simulation system for large-scale distributed space systems, implemented using low Earth orbit satellites. The system includes end nodes and computing satellite nodes. The method specifically includes: S1. Constructing a distributed space system performance evaluation index system, which includes heterogeneous task support, massive task support, task real-time performance, resource utilization, load balancing, and application business indicators; S2. Constructing a relative matrix based on the analytic hierarchy process (AHP); S3. Obtaining a comprehensive score using a multi-dimensional fuzzy evaluation method; and S4. Designing the architecture and functional modules of the distributed space system simulation system. This invention, through kinematic simulation of satellite nodes, configuration and detection of node resources, design of various simulation scenarios and schemes, and multi-index evaluation of scenarios, can meet the performance evaluation and optimization needs of distributed space systems in complex application scenarios.

Description

Technical Field

[0001] This invention relates to the field of satellite network and performance evaluation technology, and proposes a multi-dimensional performance evaluation method and simulation system for large-scale distributed space systems. Background Technology

[0002] With the rapid development of satellite technology, large-scale distributed satellite space systems have been widely used in global communication, remote sensing, navigation, and positioning. Primarily composed of low-Earth orbit small satellites, these systems connect various space-based heterogeneous nodes to construct ultra-large-scale networks, achieving seamless global coverage and efficient communication, thus becoming an important direction for the future development of space networks. Accurately evaluating the performance of distributed space systems is crucial in their research and application. However, existing technologies have many limitations. The prior art most relevant to this invention is as follows:

[0003] A satellite performance evaluation method (CN202010486519.0): This paper presents a performance evaluation scheme based on environmental fuzzy factors in the performance evaluation technology of cognitive radio-based satellite communication systems. In the performance evaluation of cognitive satellite communication systems, this scheme innovatively adds environmental fuzzy factors to construct an environmental parameter system, based on the combination of the analytic hierarchy process (AHP) and fuzzy comprehensive evaluation. First, an index parameter system is constructed according to the AHP. Then, an environmental parameter system is constructed based on the surrounding environment and environmental fuzzy factors. The fuzzy operators of the fuzzy comprehensive evaluation method can compensate for the subjectivity of the weights in the AHP, and the AHP can simplify the cumbersome data preprocessing steps of the fuzzy comprehensive evaluation method. Furthermore, the newly added environmental parameter system can not only more realistically evaluate the performance of the satellite communication system, but also dynamically adjust the system input parameters, thereby achieving the function of a feedback system and improving the overall system performance.

[0004] A Distributed Computing Data Flow Scheduling and Routing Method for Low-Earth Orbit Satellite Networks (CN202210793602.1): Applicable to the field of information and communication technology improvement, this paper provides a distributed computing data flow scheduling and routing method for low-Earth orbit satellite networks, including: S1, constructing a satellite distributed computing framework based on SDN and simulating the orbit of the Giants constellation satellite using satellite simulation software to obtain a satellite network model; S2, establishing a virtualized network simulating the satellite network topology using Mininet based on the obtained satellite network model; S3, configuring routing algorithms using an SDN controller; S4, configuring the distributed computing framework Hadoop on the host nodes in the satellite network and deploying actual computing tasks to achieve data packet-level simulation of satellite distributed computing. The simulation of the Giants constellation orbit obtains the low-Earth orbit satellite network model. Combining the advantages of OpenFlow flow tables and the topology awareness and centralized control of the SDN controller, a flood-free routing establishment mechanism for the network topology is proposed, which can improve the forwarding efficiency of satellite network information.

[0005] Satellite Performance Evaluation Method, Apparatus, and Electronic Equipment (CN202111514065.4): This application provides a satellite performance evaluation method, apparatus, and electronic equipment. The method includes: determining satellite performance indicators based on satellite model; classifying the satellite performance indicators according to a system performance analysis method based on preset indicator categories; obtaining a dataset based on the classified satellite performance indicators; and processing the dataset using the analytic hierarchy process (AHP) to obtain performance evaluation results. This application constructs a classification model for satellite performance indicators using the system performance analysis method, refining satellite performance evaluation into a set of satellite performance indicators, thus improving the accuracy of the evaluation results. Furthermore, the data feedback source for satellite performance indicators in the system performance analysis method is based on actual usage conditions, avoiding overly idealized models. The consistency check in the AHP method significantly reduces the error of a single expert scoring mechanism.

[0006] The performance evaluation method of this invention relies on the modular architecture and end-to-end integration capabilities of large-scale distributed space system simulation software, integrating evaluation simulation and simulation verification. Compared with the isolated evaluation mode of traditional analytic hierarchy process (AHP), it can achieve closed-loop verification of "index definition - weight setting - scenario simulation". Through data linkage between the simulation module and the evaluation module, the real-time monitoring of Prometheus+Grafana in the simulation system can quickly locate the source of the problem and verify the correctness of the evaluation data. For example, the node resource utilization index is correlated with the CPU utilization of virtualized nodes in real time. "Low resource utilization" is caused by the limited link bandwidth in the simulation scenario, etc.

[0007] In addition to addressing the issues of traditional evaluation methods being limited to a single scenario and having poor evaluation results, this invention supports comparisons of multiple scenarios and methods. It sets differentiated performance evaluation matrices and weight matrices for multiple scenarios, such as appropriately increasing the weight of mission real-time performance in satellite on-orbit micro-cloud migration and increasing the weight of application layer indicators in space-based data processing, thereby maximizing the accuracy and reliability of the evaluation. Summary of the Invention

[0008] To address the problems existing in the prior art, the present invention aims to provide a multi-dimensional performance evaluation method and simulation system for large-scale distributed space systems. This simulation system covers a variety of technical applications, including but not limited to kinematic simulation of satellite nodes, configuration and detection of node resources, design of various simulation scenarios and schemes, and evaluation of multiple indicators of scenarios, which can meet the performance evaluation and optimization needs of distributed space systems in complex application scenarios.

[0009] To achieve the above objectives, this invention provides a method for multi-dimensional performance evaluation of a large-scale distributed space system, which is based on low Earth orbit satellites; the method specifically includes the following steps:

[0010] S1. Construct a performance evaluation index system for distributed space systems: This evaluation index system includes heterogeneous task support, massive task support, task real-time performance, resource utilization, load balancing, and application business indicators.

[0011] S2. Constructing a relative matrix based on the analytic hierarchy process;

[0012] S3. A comprehensive score is obtained using a multi-dimensional fuzzy evaluation method;

[0013] S4. Design the architecture and functional modules of a distributed space system simulation system.

[0014] Furthermore, step S2 specifically includes the following steps:

[0015] S2.1: For the performance evaluation index system of distributed space systems, a scoring method is used to compare the relative importance of each element of the index pairwise and score them accordingly; a judgment matrix is ​​then constructed based on this. ;

[0016] S2.2: Calculate the eigenvectors;

[0017] S2.3: Consistency check.

[0018] Furthermore, the pairwise scoring standard is the 1-9 scale; 1 represents equal or slightly equal, 3 represents the former being slightly more important than the latter, 5 represents significantly important, 7 represents extremely important, 9 represents strongly important, and 2, 4, 6, and 8 represent median values.

[0019] Furthermore, the eigenvector is calculated as follows: multiply the elements of each row of the matrix and take the nth root to obtain the eigenvector of the i-th index. :

[0020] eigenvectors Normalization is performed to obtain the weight vector. ;

[0021] Calculate the judgment matrix eigenvalues .

[0022] Furthermore, the consistency check aims to determine whether the weight values ​​derived from the judgment matrix A are reasonable.

[0023] Furthermore, the consistency verification is based on the consistency index CI. The closer the index value is to 0, the higher the reliability of the obtained weight value.

[0024] Furthermore, in the specific verification process, if the consistency verification passes, the result will be that the consistency ratio CR < 0.1, which indicates that the derived weight is within a reasonable range.

[0025] Furthermore, the formula for calculating the consistency index is as follows:

[0026] ;

[0027] RI is obtained by looking up a table, and CI is calculated using the following formula:

[0028] .

[0029] Furthermore, in step S3, each indicator is normalized, and the value of each indicator is transformed to a certain range so that it meets the conditions of normalization and dimensionlessness, and its fuzzy membership degree is determined; a corresponding fuzzy membership degree function is established for each indicator.

[0030] On the other hand, the present invention provides a simulation system for multi-dimensional performance evaluation of large-scale distributed space systems, which is implemented according to the above method.

[0031] The beneficial effects of this invention are as follows:

[0032] The performance evaluation index system established in this invention starts from the diverse application needs of distributed space systems, comprehensively covering multiple levels such as business, resources, and performance. It better reflects the task- and requirement-oriented nature of performance and closely aligns with the business direction and needs of the application parties. The designed multi-dimensional fuzzy evaluation method comprehensively considers multiple factors, enabling a scientific and objective evaluation of the degree to which distributed space systems meet user needs, with high accuracy and operability. The developed simulation software integrates the advantages of discrete event simulation and virtualization simulation, supporting the simulation of a large number of distributed space system nodes within limited simulation resources. This provides important tools and technical support for the research, design optimization, and operation management of distributed space systems. Attached Figure Description

[0033] Figure 1 This is a schematic diagram of the architecture design of the multi-dimensional performance evaluation method and simulation system for large-scale distributed space systems according to the present invention.

[0034] Figure 2 This is a flowchart of the multi-index comprehensive evaluation based on the present invention. Detailed Implementation

[0035] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0036] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0037] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0038] The following combination Figures 1-2 Specific embodiments of the present invention will be described in detail below. It should be understood that the specific embodiments described herein are for illustrative and explanatory purposes only and are not intended to limit the present invention.

[0039] The technical problem solved by this invention is to provide a multi-dimensional performance evaluation method and simulation system for large-scale distributed space systems, to meet the needs of quantitative evaluation and comparative analysis of the performance of multiple satellite nodes and various mission scenarios in distributed space systems, and to provide technical support for the design optimization of distributed space systems. The technical solution of this invention is: a multi-dimensional performance evaluation method for large-scale distributed space systems, specifically including the following steps:

[0040] S1. Constructing a Distributed Space System Performance Evaluation Index System: This evaluation index system covers multiple key dimensions to comprehensively measure the performance of the distributed space system. These include: Heterogeneous task support, reflecting the system's compatibility and efficient execution capability for different types of tasks (such as bandwidth-sensitive, computing resource-sensitive, and track resource-sensitive services), requiring the system to have flexible task scheduling and resource allocation mechanisms; Massive task support, reflecting the system's ability to handle a large number of concurrent tasks, involving efficient resource allocation algorithms, powerful parallel processing capabilities, and scalable architecture; Task real-time performance, focusing on the system's ability to respond to and complete tasks within a specified time, crucial for critical applications; Resource utilization, measuring the system's effective utilization of node, link, and track resources, requiring reasonable resource allocation strategies; Load balancing, ensuring even distribution of system workload, improving concurrent processing capabilities and system availability; and Application business indicators, directly related to specific application scenarios and user needs, reflecting the actual application effect of the system. The flowchart for the multi-index comprehensive evaluation of the performance evaluation module is as follows: Figure 2 As shown.

[0041] The various indicator parameters output from the system simulation results are used as input layers. These indicators include heterogeneous task support, massive task support, task implementation, resource utilization, load balancing, and application business indicators.

[0042] S2. Constructing relative matrices based on the analytic hierarchy process: specifically including:

[0043] S2.1: For the performance evaluation index system of distributed space systems, expert scoring is used to compare the relative importance of each element of the index pairwise. The pairwise scoring standard is a 1-9 scale. 1 represents equal or slightly equal, 3 represents the former being slightly more important than the latter, 5 represents significantly important, 7 represents extremely important, 9 represents strongly important, 2, 4, 6, and 8 represent intermediate values, and the reciprocals of 1-9 represent the reversed order of importance. A judgment matrix is ​​constructed based on this. For example, in one specific embodiment, the judgment matrix is ​​as follows:

[0044]

[0045] S2.2: Calculate the eigenvectors. For each row of the matrix... Multiplying the results and taking the nth root yields the eigenvector of the i-th index. :

[0046]

[0047] In a specific implementation, the product of the elements in each row of the given matrix is ​​calculated first, and then the nth root is calculated. (n = 6):

[0048] = 1.817

[0049] = 1.781

[0050] = 1.414

[0051] = 0.890

[0052] = 0.890

[0053] = 0.275

[0054] eigenvectors Normalization is performed to obtain the weight vector. :

[0055]

[0056] in, for The corresponding feature vector;

[0057] The normalization process is as follows:

[0058] = 7.070

[0059] ≈ 0.257

[0060] ≈ 0.252

[0061] ≈ 0.200

[0062] ≈ 0.126

[0063] ≈ 0.126

[0064] ≈ 0.038

[0065] ;

[0066] Calculate the judgment matrix eigenvalues :

[0067]

[0068] calculate =[1.646,1.620 1.269 0.806 0.864 0.243]

[0069] =38.700

[0070] =6.450.

[0071] S2.3: Consistency Check. This check aims to determine whether the weight values ​​derived from the judgment matrix A are reasonable. The key criterion is the consistency index (CI); the closer the CI value is to 0, the higher the reliability of the obtained weight values.

[0072] During the specific verification process, if the consistency check passes, the consistency ratio CR will be less than 0.1, indicating that the derived weights are reasonable. The formula for calculating the consistency index is as follows:

[0073]

[0074] RI is obtained by looking up a table, and CI is calculated according to the following formula.

[0075]

[0076] Calculate CI = 0.090.

[0077] The table for RI is as follows. From the table, we find that RI = 1.26 (n = 6).

[0078] n 1 2 3 4 5 6 7 8 9 10 11 12 RI 0.00 0.00 0.52 0.89 1.12 1.26 1.36 1.41 1.46 1.49 1.52 1.54

[0079] The corresponding CR is 0.072, and the consistency ratio CR < 0.1. This indicates that the weights we derived are relatively reasonable.

[0080] S3. Multi-dimensional Fuzzy Evaluation Method for Comprehensive Score: In human factors engineering evaluation, the evaluation standards, methods, and units of each indicator are different, lacking commensurability and making comparison and evaluation difficult. Therefore, this invention standardizes each indicator, transforming its value to a certain range to meet the conditions of normalization and dimensionlessness before determining its fuzzy membership degree. For each indicator, a corresponding fuzzy membership function is established based on its characteristics and practical application requirements. For example, for the task real-time performance indicator, corresponding membership degrees are set according to different task response time ranges. The relative matrix weights obtained by the analytic hierarchy process are combined with the calculation results of the fuzzy membership function to obtain a comprehensive score.

[0081] For quantitative indicators in the indicator system, this invention adopts a "bigger is better" approach, using a threshold method to determine the linear function. The dimensionless indicator value is obtained by comparing the actual indicator value with the threshold. The Analytic Hierarchy Process (AHP) is used to determine the weights, which typically characterize the "importance" of an indicator. Multiplying the evaluation value of the "bigger is better" indicator by its weight shows that indicators with larger values ​​contribute more significantly to the overall score, aligning with the evaluation logic of "important indicators need to be optimized." Therefore, this invention employs the "bigger is better" approach, specifically using the following calculation formula:

[0082] ;

[0083] in, Represents the original index in the method of this invention. The lower limit threshold represents the minimum acceptable level for this indicator. If the indicator value is lower than... If the value is zero, then the contribution of this indicator to the system performance is considered to be 0. The upper threshold of the i-th indicator represents the value at which the indicator reaches its ideal state. When the indicator value... Reaching or exceeding If the contribution of this indicator to the system efficiency is considered to be 1, then the system efficiency is considered to have achieved the full score. It is the evaluation value of the i-th indicator after transformation, and its value ranges from [0, 1].

[0084] Based on the task completion time obtained from the simulation results, the calculated load balancing rate and other parameters are used as input layer parameters, which are then transformed using a fuzzy membership function to obtain scores for each indicator.

[0085] ;

[0086] The overall score is obtained by multiplying each score by the weight vector:

[0087] ;

[0088] The evaluation score is a real number ranging from [0,1]. The closer the score is to 0, the worse the effect is, and it is not recommended to use it. The closer the score is to 1, the better the effect is, and it is recommended to use it.

[0089] S4. Store the evaluation results in a MySQL database and output them to the front-end simulation interface.

[0090] According to the present invention, a large-scale distributed space system multi-dimensional performance evaluation simulation system is used to design the architecture and functional modules of a distributed space system simulation system. For example... Figure 1 As shown, an architecture integrating discrete event simulation and virtualization simulation is adopted. The architecture is divided into three layers: virtual node and resource simulation layer, simulation scenario and performance evaluation layer, and application and data layer.

[0091] Virtual Node and Resource Simulation Layer: This layer provides the basic operating environment for virtualized nodes, offering underlying support for configuring node communication, computing, and storage resources. It provides node resource status information to the simulation scenario and performance evaluation layer, executes returned network decisions, and provides the hardware and software environment for business operations to the application and data layers. Specifically, it establishes Docker containers corresponding to satellite nodes, simulates the software operating environment by configuring the container's hard disk space, memory, CPU, and GPU, and further simulates virtual node resources by configuring the node's communication bandwidth, communication latency, and communication packet loss rate.

[0092] Simulation Scenario and Performance Evaluation Layer: This layer is specifically divided into two sub-layers: the scenario simulation layer and the performance evaluation layer. The scenario simulation layer first models and simulates the basic characteristics of the three system nodes: ground station, satellite, and end nodes, including the kinematic parameters and basic resources of the nodes. Based on this, the scenario simulation layer combines resource information provided by the virtual node and resource simulation layers to form a space network of infrastructure nodes, simulating network decisions and behaviors, including the selection of edge computing nodes, satellite orbit changes, and the data transmission process of inter-satellite communication. The simulation scenario layer receives device and resource status information from the virtual node and resource simulation layers, returns network decision instructions, and receives business requirements and execution status information from the application and data layers, returning task execution instructions. The relevant parameters in these task execution processes and results are further fed back to the performance evaluation layer. The performance evaluation comprehensively assesses factors such as heterogeneous task compatibility, support for massive tasks, task real-time performance, resource utilization, load balancing, and application business indicators. Specifically, the scenario simulation is conducted using the satellite micro-cloud simulation platform developed in this invention, encompassing four scenarios: on-orbit micro-cloud migration, massive data transmission back to the ground, space-based native data processing, and high-orbit maneuver interception. The performance evaluation layer is based on the performance evaluation method of this invention.

[0093] Application and Data Layer: This layer encapsulates space-based missions in containers. Deployment and computation are performed on virtual nodes based on decisions from the simulation scenario and performance evaluation layer, allowing for observation of the mission's execution effectiveness. This part runs within the environment of the virtual nodes and resource simulation layer, receiving mission execution instructions from the simulation scenario and performance evaluation layer and returning the mission execution status. Specifically, remote sensing satellite image recognition is implemented as a space-based distributed task, with the image recognition program running on Docker nodes corresponding to the satellite nodes.

[0094] To verify the effectiveness of the distributed space system simulation system, this invention uses an on-orbit dynamic edge computing scenario as an example to explain the specific implementation of the performance evaluation method.

[0095] Due to the unique advantage of ubiquitous access, space-based distributed systems can provide crucial edge computing services in remote areas with insufficient, congested, or nonexistent terrestrial network coverage. This service is particularly important for scenarios such as emergency communications, ocean voyages, and polar scientific expeditions, significantly expanding the boundaries of information services and improving operational efficiency and security in these special environments. The core component of space-based distributed systems—low Earth orbit (LEO) satellites—is constantly in high-speed motion. This not only requires the system to possess high flexibility and adaptability but also causes the entire space network topology, including the satellite-to-ground network, to be constantly and rapidly changing. Therefore, on-orbit edge cloud construction strategies must consider the motion characteristics of satellites and design dynamic resource allocation algorithms. In practical applications, maintaining service continuity and stability becomes a critical issue when the served end nodes (such as ground mobile devices, maritime platforms, or air vehicles) leave the coverage area of ​​the current edge cloud.

[0096] The distributed space system according to the present invention includes two types of nodes: end nodes and computing satellite nodes. When the computing demands generated by the end nodes reach the satellite nodes, the satellite nodes become the task initiating nodes. If the computing power and storage resources of the satellite nodes themselves cannot meet the task requirements, the system uses the satellite node as the master node, forming an on-orbit edge cloud with surrounding nodes to jointly complete the computing task. Due to the high-speed movement of LEO satellites and end nodes, the end nodes will move relative to the edge cloud. Therefore, after the end nodes leave the coverage area of ​​the edge cloud, the current edge cloud master node will maintain contact with the end nodes through service migration or inter-satellite data routing.

[0097] Current research on dynamic edge computing in space networks largely focuses on stationary ground-based endpoints or those with relatively slow movement speeds, such as smartphones and IoT devices. However, the endpoints in this invention are high-speed mobile endpoints, such as airplanes, high-speed trains, or self-driving cars. The relative motion between these high-speed mobile endpoints and satellite nodes is extremely drastic, leading to frequent and significant changes in the topology of the space-to-ground network, thus posing unprecedented challenges to the provision of edge computing services. The continuous connectivity between the edge cloud and endpoints, especially for latency-sensitive services such as real-time video processing, online games, and telemedicine, directly determines the total latency of the entire process from a node issuing a computing request to receiving the processing result, depending on the location of the edge cloud, migration strategies, and data routing. This, in turn, affects service quality.

[0098] This invention employs inter-satellite service migration to adapt to high-speed mobile end nodes. The specific implementation method of inter-satellite service migration is as follows:

[0099] When an end node issues a computation request, to reduce the access latency between the end node and the edge cloud, the end node broadcasts to the space nodes. The nearest computing satellite node in the distributed space system responds to the end node's computation request. If the nearest computing satellite node's own computing power is insufficient to handle the computation request, this computing satellite node will act as the master node and form an on-orbit edge cloud with surrounding computing satellite nodes. To reduce the edge cloud establishment time and the communication cost between master and slave nodes, nodes with smaller hop counts from the master node are preferentially selected to join the cloud. Specifically, a greedy algorithm is used to select the computing satellite node with the smallest hop count from the set of idle computing satellite nodes to join the cloud. Then, the computing power and storage requirements of the current edge cloud are checked to see if they meet the requirements. This process is repeated until the requirements are met.

[0100] When an end node leaves or is about to leave the coverage area of ​​the edge cloud, the service migration mechanism of the distributed space system will be triggered. Inter-satellite service migration can be divided into two steps: edge cloud service information data transmission and the establishment of a new edge cloud node network. During service migration, user services will be temporarily interrupted. To improve user experience, service migration time should be minimized as much as possible. Therefore, this invention uses transmission latency and propagation latency as weights when designing the transmission path. The algorithm for selecting new edge cloud nodes is designed based on the edge cloud establishment strategy. The service migration strategy of this invention ensures that end nodes are always under the coverage of the edge cloud, resulting in lower access latency. During service delivery, satellite nodes continuously take over the work, providing better load balancing.

[0101] The specific operation procedure of the simulation system is as follows:

[0102] ① After opening the distributed space system multidimensional performance evaluation simulation software, the user should first click [File] in the menu bar to create a new simulation scenario or open a .sc file to import the simulation scenario;

[0103] ②After completing the scene setup, click on the [Object] menu to insert and configure the distributed space node simulation object.

[0104] ③ After completing the configuration of the distributed space node simulation object, click on the task you want to simulate, the selected strategy, and configure the parameters in the task configuration and strategy selection display section.

[0105] ④ After completing the task configuration, the simulation initialization is completed, and the simulation log prompts "Simulation initialization is complete, simulation can start". The 3D display module displays the status of the initialized distributed space system.

[0106] ⑤ Click the [Start Simulation] button in the menu bar to start the simulation. The simulation log module records the real-time status of the task execution and records key information. The 3D display module visualizes information such as the computing cluster setup and data transmission path. During the simulation, Prometheus+Grafana can monitor the operation of satellite nodes in real time. After the simulation ends, the simulation log and performance evaluation module evaluates the performance indicators.

[0107] Any process or method described in the flowcharts of this invention or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, achievable on any computer-readable medium for use by an instruction execution system, apparatus, or device. The computer-readable medium can be any medium containing a program for storage, communication, propagation, or transmission for use by an execution system, apparatus, or device, including read-only memory, magnetic disks, or optical disks.

[0108] In the description of this specification, references to terms such as "embodiment," "example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, those skilled in the art can combine or combine the different embodiments or examples described in this specification and the features therein without causing contradiction.

[0109] While embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions, and alterations to the above embodiments within the scope of the present invention.

Claims

1. A multi-dimensional performance evaluation method for large-scale distributed space systems, characterized in that, The distributed space system is based on low Earth orbit satellites; the distributed space system includes end nodes and computing satellite nodes, and the method specifically includes the following steps: S1. Construct a performance evaluation index system for distributed space systems: This evaluation index system includes heterogeneous task support, massive task support, task real-time performance, resource utilization, load balancing, and application business indicators. S2. Constructing a relative matrix based on the analytic hierarchy process; S3. A comprehensive score is obtained using a multi-dimensional fuzzy evaluation method; S4. Store the evaluation results in a MySQL database and output them to the front-end simulation interface; In this process, the end nodes and satellite nodes move at relatively high speeds. When the computing needs generated by the end node reach the satellite node, the satellite node becomes the task initiating node. If the computing power and storage resources of the satellite node itself cannot meet the task requirements, the satellite node is used as the master node to form an on-orbit edge cloud with surrounding nodes to jointly complete the computing task. After the end node leaves the coverage area of ​​the edge cloud, the current edge cloud master node maintains contact with the end node through inter-satellite service migration or inter-satellite data routing to adapt to the high-speed movement of the end node. The method for implementing inter-satellite service migration is as follows: When an end node issues a computing request, to reduce the access latency between the end node and the edge cloud, the end node broadcasts to the space nodes. The computing satellite node closest to the end node in the distributed space system responds to the end node's computing request. If the computing power of the nearest computing satellite node is insufficient to handle the computing request, this computing satellite node will act as the master node and form an on-orbit edge cloud with surrounding computing satellite nodes. To reduce the edge cloud establishment time and the communication cost between master and slave nodes, nodes with smaller hop counts from the master node are preferentially selected to enter the cloud. When an end node leaves or is about to leave the coverage of the edge cloud, the service migration mechanism of the distributed space system will be triggered. The inter-satellite service migration is divided into two steps: edge cloud service information data transmission and new edge cloud node networking. When designing the transmission path, the transmission latency and propagation latency are used as weights for data communication latency. The new edge cloud nodes ensure that the end node is always under the coverage of the edge cloud, with low access latency. Through continuous relay work by satellite nodes, load balancing is achieved.

2. The method for multi-dimensional performance evaluation of large-scale distributed space systems according to claim 1, characterized in that, Step S2 specifically includes the following steps: S2.1: For the performance evaluation index system of distributed space systems, a scoring method is used to compare the relative importance of each element of the index pairwise and score them accordingly; a judgment matrix is ​​then constructed based on this. ; S2.2: Calculate the eigenvectors; S2.3: Consistency check.

3. The method for multi-dimensional performance evaluation of large-scale distributed space systems according to claim 2, characterized in that, The pairwise scoring standard is based on the 1-9 scale. 1 represents equal importance, 3 represents slightly more important than the latter, 5 represents significantly important, 7 represents extremely important, 9 represents very important, and 2, 4, 6, and 8 represent intermediate values.

4. The method for multi-dimensional performance evaluation of large-scale distributed space systems according to claim 2, characterized in that, The eigenvector is calculated as follows: multiply the elements of each row of the matrix and take the nth root to obtain the eigenvector of the i-th index. : eigenvectors Normalization is performed to obtain the weight vector. ; Calculate the judgment matrix eigenvalues .

5. The method for multi-dimensional performance evaluation of large-scale distributed space systems according to claim 2, characterized in that, Consistency checks aim to determine whether the weight values ​​derived from the judgment matrix A are reasonable.

6. The method for multi-dimensional performance evaluation of large-scale distributed space systems according to claim 5, characterized in that, The consistency check is based on the consistency index CI. The closer the index value is to 0, the higher the reliability of the obtained weight value.

7. The method for multi-dimensional performance evaluation of large-scale distributed space systems according to claim 6, characterized in that, In the specific verification process, if the consistency verification passes, the result will be that the consistency ratio CR < 0.1, which indicates that the derived weight is within a reasonable range.

8. The method for multi-dimensional performance evaluation of large-scale distributed space systems according to claim 7, characterized in that, The formula for calculating the consistency index is as follows: ； RI is obtained by looking up a table, and CI is calculated using the following formula: 。 9. The method for multi-dimensional performance evaluation of large-scale distributed space systems according to claim 8, characterized in that, In step S3, each indicator is normalized, and the value of each indicator is transformed to a certain range so that it meets the conditions of normalization and dimensionlessness, and its fuzzy membership degree is determined; a corresponding fuzzy membership degree function is established for each indicator.

10. A multi-dimensional performance simulation system for large-scale distributed space systems, characterized in that, The simulation system is used to implement the multi-dimensional performance evaluation method for large-scale distributed space systems according to any one of claims 1-9, wherein the distributed space system includes end nodes and computing satellite nodes, and the end nodes are high-speed moving nodes.

Images