A satellite-borne computing power elastic pooling method and system based on a dynamic resource mapping ratio, a terminal, and a storage medium

Abstract

The application discloses a kind of based on dynamic resource mapping ratio's satellite computing power elasticity pooling method, system, terminal and storage medium, the method includes: obtaining the running state of the service deployed in resource pool, according to the running state actual load of current resource pool is obtained, and according to satellite transit trajectory and the distribution law of the ground area of satellite transit trajectory, predicted load is obtained;According to the actual load and the predicted load, the resource dynamic mapping ratio currently used is determined, and the guarantee level quota and the best effort level quota are output;Based on the actual load, the predicted load, the guarantee level quota and the best effort level quota, the scheduling of two levels of priority is carried out to service, and service scheduling information is obtained;According to the resource dynamic mapping ratio, the guarantee level quota, the best effort level quota and the service scheduling information, the actual action of the running of relevant service deployment is executed.The application significantly improves the utilization rate of scarce satellite physical resources.

Description

Technical Field

[0001] This invention relates to the field of spaceborne computing and resource virtualization technology, and particularly to a method, system, and computer-readable storage medium for elastic pooling of spaceborne computing power based on dynamic resource mapping ratio. Background Technology

[0002] With the rapid expansion of satellite constellations and the improvement of onboard computing capabilities, the "satellite cloud-native" model, which uses satellite constellations as the computing power foundation to provide cloud services, is attracting increasing attention. In this model, onboard computing power is organized into a computing resource pool, and virtual computing resources are provided externally for business deployment through resource virtualization technology. The number of virtual resources that the resource pool can provide externally depends on the mapping ratio between physical and virtual resources, i.e., the number of virtual resources mapped to a unit of physical resource; this is commonly referred to as the over-allocation ratio in terrestrial cloud computing.

[0003] Existing resource mapping and over-allocation technologies have significant shortcomings in satellite scenarios. First, existing over-allocation technologies use static or slowly varying mapping ratios, which cannot adapt to the drastic time-varying loads in satellite scenarios. Second, existing over-allocation technologies are mainly geared towards static or slowly varying scenarios, and lack an effective mechanism for dynamically adjusting mapping ratios at large and high frequencies while simultaneously implementing tiered management of over-committed resources to ensure critical business operations.

[0004] Therefore, existing technologies still need to be improved and developed. Summary of the Invention

[0005] The main objective of this invention is to provide a method, system, and computer-readable storage medium for elastic pooling of onboard computing power based on dynamic resource mapping ratio. This aims to solve the problems in existing satellite scenarios where fixed mapping ratios cannot adapt to drastic time-varying loads, idle physical resources cannot be released and reused, and the risk of resource exhaustion introduced by dynamic over-division is difficult to manage.

[0006] To achieve the above objectives, this invention provides a method for elastic pooling of onboard computing power based on dynamic resource mapping ratio. This method includes the following steps: Obtain the operational status of deployed services in the resource pool, determine the current actual load of the resource pool based on the operational status, and obtain the predicted load based on the satellite transit trajectory and the distribution pattern of the ground areas traversed by the satellite transit trajectory. Based on the actual load and the predicted load, determine the current dynamic resource mapping ratio and output the guarantee level quota and the best effort level quota. Based on the actual load, the predicted load, the guaranteed level limit, and the best effort level limit, the service is scheduled with two levels of priority to obtain service scheduling information. Based on the dynamic resource mapping ratio, the guarantee level limit, the best effort level limit, and the service scheduling information, execute all actual actions of the relevant service deployment.

[0007] Optionally, the aforementioned method for elastic pooling of onboard computing power based on dynamic resource mapping ratio, wherein obtaining the operating status of deployed services in the resource pool, obtaining the current actual load of the resource pool based on the operating status, and obtaining the predicted load based on the satellite transit trajectory and the distribution pattern of the ground areas traversed by the satellite transit trajectory, specifically includes: The actual occupancy of physical computing resources by deployed services in the onboard computing resource pool is periodically collected, and the actual occupancy is aggregated and normalized to obtain the measured utilization rate of physical resources by the current satellite load. ; Based on known ephemeris parameters, the satellite transit area is obtained through orbit calculations. Then, based on the distribution patterns of on-orbit users and service types within this area, the trend of physical resource utilization for deployed services over a future period is predicted. ; Based on the measured usage rate and the aforementioned trend of change The predicted load is obtained. .

[0008] Optionally, the onboard computing power elastic pooling method based on dynamic resource mapping ratio, wherein determining the currently used dynamic resource mapping ratio based on the actual load and the predicted load, and outputting the guaranteed level quota and the best-effort level quota, specifically includes: Obtain the total physical computing power resources of the satellite nodes, and configure the resource baseline mapping ratio, initial resource dynamic mapping ratio, and dynamic mapping ratio level for the satellite nodes; Based on the total physical computing power resources and the resource baseline mapping ratio, calculate and output the guarantee level quota: ; Wherein, P represents the total amount of physical computing resources. This represents the resource baseline mapping ratio. This indicates the coverage level limit; Based on the actual load and the predicted load, the initial resource dynamic mapping ratio is adjusted according to the dynamic mapping ratio level to obtain the resource dynamic mapping ratio. Based on the total physical computing resources, the resource baseline mapping ratio, and the resource dynamic mapping ratio, calculate and output the best-effort quota: ; in, This indicates the dynamic mapping ratio of the resources. This indicates the effort level limit.

[0009] Optionally, the onboard computing power elastic pooling method based on dynamic resource mapping ratio, wherein the step of adjusting the initial dynamic resource mapping ratio according to the dynamic mapping ratio level based on the actual load and the predicted load to obtain the dynamic resource mapping ratio specifically includes: When the measured usage rate The predicted load is below a first preset threshold. At that time, according to the preset adjustment rules, the initial resource dynamic mapping ratio is adjusted upward based on the dynamic mapping ratio, to obtain the resource dynamic mapping ratio used to expand the best effort level quota. When the measured usage rate The predicted load is below a first preset threshold. At the same time, the initial dynamic resource mapping ratio is kept unchanged, and the initial dynamic resource mapping ratio is used as the dynamic resource mapping ratio to reserve room for the realization of the next stage of load increase; When the measured usage rate The predicted load is higher than the first preset threshold. At the same time, the initial dynamic resource mapping ratio is kept unchanged, and the initial dynamic resource mapping ratio is used as the dynamic resource mapping ratio to maintain the efficient state of full utilization of existing physical resources; When the measured usage rate The predicted load is higher than the first preset threshold. At that time, according to the preset adjustment rules, the initial resource dynamic mapping ratio is adjusted downward based on the dynamic mapping ratio, so as to obtain the resource dynamic mapping ratio used to shrink the best effort level quota.

[0010] Optionally, the spaceborne computing power elastic pooling method based on dynamic resource mapping ratio, wherein the step of performing two-level priority scheduling of services based on the actual load, the predicted load, the guaranteed level quota, and the best-effort level quota to obtain service scheduling information specifically includes: Based on the reliability and real-time requirements of the business, business instances are classified into guaranteed-level business and best-effort-level business. Based on the actual load, the predicted load, the guaranteed level quota, the best-effort level quota, the latency requirements of the service deployment location, and satellite orbital motion factors, service scheduling information is generated for the guaranteed level service and the best-effort level service.

[0011] Optionally, the spaceborne computing power elastic pooling method based on dynamic resource mapping ratio, wherein the step of executing all actual actions for the operation of relevant service deployments according to the dynamic resource mapping ratio, the guaranteed level quota, the best-effort level quota, and the service scheduling information specifically includes: Based on the resource dynamic mapping ratio, fixed physical resources are mapped to a corresponding number of virtual resources. Based on the guaranteed level quota and the best-effort level quota, the virtual resources are used to establish and maintain a guaranteed level resource pool and a best-effort level resource pool. Based on the service scheduling information, the service instance is deployed to the guaranteed resource pool or the best-effort resource pool for operation; When both guaranteed-level and best-effort-level services exist simultaneously, adjust cgroup resource limits to prioritize the needs of guaranteed-level services. When the demand for guaranteed services increases further, the deployed best-effort services will be temporarily suspended to free up physical computing resources. If the best-effort services are suspended for a period of time exceeding a second preset threshold, the best-effort services will be migrated to other satellite nodes with available resources.

[0012] Optionally, the spaceborne computing power elastic pooling method based on dynamic resource mapping ratio further includes: When physical resources are scarce, based on the underlying resource isolation and preemption mechanism, at least one of the following actions is performed on best-effort level services: restriction action, suspension action, eviction action, and reclamation action, to ensure that guarantee level services have priority in occupying physical resources. The limiting action is to adjust cgroup resource limits to reduce the resource consumption of best-effort level services; The suspension action is to temporarily pause the deployed best-effort level services to free up physical computing resources; The eviction action is to forcibly terminate the best-effort level service instance to free up the occupied physical resources. The recycling action involves reclaiming the virtual resources already allocated to the best-effort level services and correspondingly reducing the best-effort level resource pool.

[0013] Furthermore, to achieve the above objectives, the present invention also provides a spaceborne computing power elastic pooling system based on dynamic resource mapping ratio, wherein the spaceborne computing power elastic pooling system based on dynamic resource mapping ratio includes: The load perception and prediction module is used to obtain the operating status of the deployed services in the resource pool, obtain the current actual load of the resource pool based on the operating status, and obtain the predicted load based on the satellite transit trajectory and the distribution pattern of the ground areas traversed by the satellite transit trajectory. The dynamic mapping ratio decision module is used to determine the current dynamic resource mapping ratio based on the actual load and the predicted load, and output the guarantee level quota and the best effort level quota. The hierarchical resource scheduling module is used to perform two-level priority scheduling of services based on the actual load, the predicted load, the guaranteed level quota, and the best effort level quota to obtain service scheduling information. The elastic mapping execution module is used to execute all actual actions of the relevant business deployment based on the resource dynamic mapping ratio, the guarantee level quota, the best effort level quota, and the business scheduling information.

[0014] Furthermore, to achieve the above objectives, the present invention also provides a terminal, wherein the terminal includes: a memory, a processor, and a spaceborne computing power elastic pooling program based on dynamic resource mapping ratio stored in the memory and executable on the processor. When the spaceborne computing power elastic pooling program based on dynamic resource mapping ratio is executed by the processor, it implements the steps of the spaceborne computing power elastic pooling method based on dynamic resource mapping ratio as described above.

[0015] Furthermore, to achieve the above objectives, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a spaceborne computing power elastic pooling program based on dynamic resource mapping ratio, and the spaceborne computing power elastic pooling program based on dynamic resource mapping ratio implements the steps of the spaceborne computing power elastic pooling method based on dynamic resource mapping ratio as described above when executed by a processor.

[0016] In this invention, the operational status of deployed services in the resource pool is obtained. Based on the operational status, the current actual load of the resource pool is determined, and the predicted load is obtained based on the satellite transit trajectory and the distribution pattern of the ground areas traversed by the satellite transit trajectory. Based on the actual load and the predicted load, the currently used dynamic resource mapping ratio is determined, and the guaranteed-level quota and the best-effort-level quota are output. Based on the actual load, the predicted load, the guaranteed-level quota, and the best-effort-level quota, services are scheduled with two levels of priority to obtain service scheduling information. Based on the dynamic resource mapping ratio, the guaranteed-level quota, the best-effort-level quota, and the service scheduling information, all actual actions for the deployment and operation of the relevant services are executed. This invention significantly improves the utilization rate of scarce onboard physical resources while ensuring that the service quality of critical services is not affected by resource constraints. Attached Figure Description

[0017] Figure 1 This is a flowchart of a preferred embodiment of the spaceborne computing power elastic pooling method based on dynamic resource mapping ratio of the present invention; Figure 2 This is the overall architecture diagram of the technical principle of the spaceborne computing power elastic pooling method based on dynamic resource mapping ratio of the present invention; Figure 3 This is a structural diagram of a preferred embodiment of the spaceborne computing power elastic pooling system based on dynamic resource mapping ratio of the present invention; Figure 4 This is a structural diagram of a preferred embodiment of the terminal of the present invention. Detailed Implementation

[0018] This application provides a method, system, and terminal for elastic pooling of onboard computing power based on dynamic resource mapping ratio. To make the objectives, technical solutions, and effects of this application clearer and more explicit, the following detailed description is provided with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only for explaining this application and are not intended to limit this application.

[0019] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0020] Furthermore, if the embodiments of this invention involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by this invention.

[0021] The preferred embodiment of the present invention describes a spaceborne computing power elastic pooling method based on dynamic resource mapping ratio, such as... Figure 1 and Figure 2 As shown, the onboard computing power elastic pooling method based on dynamic resource mapping ratio includes the following steps: Step S10: Obtain the operating status of the deployed services in the resource pool, obtain the current actual load of the resource pool based on the operating status, and obtain the predicted load based on the satellite transit trajectory and the distribution pattern of the ground areas traversed by the satellite transit trajectory.

[0022] Specifically, the actual occupancy of physical computing resources by deployed services in the onboard computing resource pool is periodically collected, and the actual occupancy is aggregated and normalized to obtain the measured utilization rate of physical resources by the current satellite's actual load. ; Based on known ephemeris parameters, the satellite transit area is obtained through orbit calculations. Then, based on the distribution patterns of on-orbit users and service types within this area, the trend of physical resource utilization for deployed services over a future period is predicted. ; Based on the measured usage rate and the aforementioned trend of change The predicted load is obtained. .

[0023] In this embodiment, the load perception and prediction module is responsible for acquiring the operational status of deployed services in the resource pool, providing a load basis for mapping ratio decisions. First, it continuously collects real-time load data, gathering the actual physical computing resource occupancy of each service to obtain the current actual load level of the resource pool. Then, it performs periodic load prediction, forecasting the load change trend of deployed services over a future period based on the satellite's transit trajectory and the distribution of on-orbit users and service types in the ground areas it passes through. This module outputs the measured computational load status and the predicted computational load trend to the dynamic mapping ratio decision module and the hierarchical resource scheduling module.

[0024] Understandably, in the real-time load perception section, nodes periodically collect data on the actual physical computing resources occupied by the services deployed on the satellite, and after aggregation and normalization, obtain the actual utilization rate of physical resources for the current satellite load. In the load prediction section, the satellite passes over different ground areas based on its known ephemeris. Taking into account the predictability of the number of on-orbit users and the distribution of service types in different areas, the satellite obtains its passing trajectory and the prior knowledge of user / service distribution in the areas it passes through, predicting the future trend of physical resource utilization. Prediction methods include, but are not limited to, table lookup prediction based on transit area load mapping tables, time series prediction, and machine learning-based prediction.

[0025] In the real-time load perception part, a load model can be built in advance. Based on the constellation ephemeris and the statistics of on-orbit users and services in various ground areas, a mapping prior between satellite transit trajectory and load level can be constructed for load prediction during the online phase.

[0026] Furthermore, although this invention uses CPU computing power as an example, the measured utilization rate and mapping ratio can be replaced or extended to memory, storage, and onboard AI accelerator (GPU / NPU) computing power, maintaining the mapping ratio and two-level quota for each type of resource. The load prediction method can be arbitrarily replaced or combined among table lookup prediction based on the transit area load mapping table, time series prediction, machine learning-based prediction, and queuing theory-based analytical prediction. In addition to satellite transit trajectories and regional user / service distribution, the prediction basis can also include historical concurrent load, periodic service patterns, temporary event forecasts, and neighboring satellite load status. The prediction duration can also be improved from a fixed value to adaptively adjust according to the severity of load fluctuations or transit speed.

[0027] Step S20: Based on the actual load and the predicted load, determine the current dynamic resource mapping ratio and output the guarantee level quota and the best effort level quota.

[0028] Specifically, the total physical computing power resources of the satellite nodes are obtained, and the resource baseline mapping ratio, initial dynamic resource mapping ratio, and dynamic mapping ratio level are configured for the satellite nodes; Based on the total physical computing power resources and the resource baseline mapping ratio, calculate and output the guarantee level quota: ; Wherein, P represents the total amount of physical computing resources. This represents the resource baseline mapping ratio. This indicates the coverage level limit; Based on the actual load and the predicted load, the initial resource dynamic mapping ratio is adjusted according to the dynamic mapping ratio level to obtain the resource dynamic mapping ratio. Based on the total physical computing resources, the resource baseline mapping ratio, and the resource dynamic mapping ratio, calculate and output the best-effort quota: ; in, This indicates the dynamic mapping ratio of the resources. This indicates the effort level limit.

[0029] It is understandable that the dynamic mapping ratio decision module is the core module of this invention. It takes the current measured load and predicted load output by the load perception and prediction module as input to determine the current dynamic resource mapping ratio, so that the mapping ratio is linked to the load in reverse - the mapping ratio is increased when the load is low and decreased when the load is high.

[0030] This embodiment involves two levels of mapping ratios, including the resource baseline mapping ratio. And resource dynamic mapping ratio Among them, the resource benchmark mapping ratio This ensures that even if the load of all services within the resource pool simultaneously reaches its peak, the virtual resources committed based on this ratio can still be fully realized from physical resources without a run on the resource pool. Note that the baseline mapping ratio itself allows for a certain degree of over-division (based on the statistical reuse pattern of multiple service peaks occurring at different times), corresponding to the static over-division level conservatively reserved based on service peaks in existing technologies. This represents the safety level committed by the resource pool. Dynamic Resource Mapping Ratio This is the actual mapping ratio dynamically increased based on the current load. When the load is at its lowest point, the actual load is much lower than the peak load, and this module is adjusted upwards. This module maps more idle physical resources to virtual resources to support more services; when the load recovers, this module lowers its settings. This would cause them to tighten their over-commitment to dynamically increase.

[0031] Accordingly, this module can output two levels of resource quotas. Mapping ratio with baseline The corresponding virtual resources constitute the guarantee level quota. ; This portion of resources is calculated based on the total physical resources and peak-level service availability, and is a guaranteed virtual resource. The virtual resources corresponding to the portion where the dynamic resource mapping ratio exceeds the baseline resource mapping ratio constitute the best-effort level quota. ; Resources allocated through dynamic over-allocation are not guaranteed to be continuously fulfilled and will scale with load changes. The guaranteed-level quota remains stable, while the best-effort level quota changes dynamically with load. This module outputs the quota indicators for both levels to the hierarchical resource scheduling module and the elastic mapping execution module.

[0032] Further, the step of adjusting the initial resource dynamic mapping ratio based on the actual load and the predicted load, and according to the dynamic mapping ratio level, to obtain the resource dynamic mapping ratio specifically includes: When the measured usage rate The predicted load is below a first preset threshold. At that time, according to the preset adjustment rules, the initial resource dynamic mapping ratio is adjusted upward based on the dynamic mapping ratio, to obtain the resource dynamic mapping ratio used to expand the best effort level quota. When the measured usage rate The predicted load is below a first preset threshold. At the same time, the initial dynamic resource mapping ratio is kept unchanged, and the initial dynamic resource mapping ratio is used as the dynamic resource mapping ratio to reserve room for the realization of the next stage of load increase; When the measured usage rate The predicted load is higher than the first preset threshold. At the same time, the initial dynamic resource mapping ratio is kept unchanged, and the initial dynamic resource mapping ratio is used as the dynamic resource mapping ratio to maintain the efficient state of full utilization of existing physical resources; When the measured usage rate The predicted load is higher than the first preset threshold. At that time, according to the preset adjustment rules, the initial resource dynamic mapping ratio is adjusted downward based on the dynamic mapping ratio, so as to obtain the resource dynamic mapping ratio used to shrink the best effort level quota.

[0033] It is understandable that the principle for adjusting the initial dynamic resource mapping ratio is... It is linked to the load; it increases when the load is low. Increase the maximum resource limit; reduce it if the load is high. Reduce the limit at the minimum level.

[0034] In this embodiment, a tiered adjustment method is used to determine the current mapping ratio. That is, preset several gears ,in, Indicates the total number of gears. Indicates the first gear. Indicates the second gear. This indicates the highest gear level. Gear switching is based on the currently measured physical resource utilization rate. With the next phase of predicted load changes Two factors determine it together. Specifically, when the measured utilization rate... The value is too low (below the first preset threshold), and the predicted load is stable or decreasing, i.e. ( )hour, Increase several levels, actively expand the best-effort level quota, and fully map and utilize idle physical resources; when the actual utilization rate Low, but predicted load is rising, i.e. ( )hour, Maintain the current gear level to allow for potential increases in load during the next phase, avoiding excessive over-distribution; when the actual utilization rate is measured... High, but the predicted load is stable or decreasing, i.e. ( )hour, Maintain the gear level unchanged, ensuring efficient utilization of existing physical resources; when the actual utilization rate... The load is high, and the predicted load is increasing, i.e. ( )hour, By lowering the threshold by several levels and reducing the maximum effort level, physical resources are made available in advance for high loads. The specific thresholds and level changes used in this mechanism can be flexibly set according to the characteristics of satellite service load changes. The two-factor combination method can be implemented using a lookup table, or any equivalent form such as weighted scoring, rule engine, or state machine. This invention does not impose any restrictions on this.

[0035] Furthermore, in addition to tiered adjustments, it can be replaced by continuous function mapping that directly calculates the continuous mapping ratio, or any equivalent method such as fuzzy control, rule engine, state machine, or reinforcement learning-based strategy. The number of tiers can be arbitrarily set and can be uniformly or non-uniformly distributed. The values ​​of each tier and the switching threshold can be statically preset or adjusted online through learning. In addition to the two factors of "actual utilization rate + predicted load change", decision factors can be further incorporated, such as the current occupancy of guaranteed-level services, the queue length of best-effort-level services, power status, and migration costs. In addition to hysteresis intervals and minimum hold time, anti-jitter measures can include switching cooldown time, rate of change limiting, and multi-cycle confirmation.

[0036] Step S30: Based on the actual load, the predicted load, the guaranteed level limit, and the best effort level limit, perform two-level priority scheduling of services to obtain service scheduling information.

[0037] Specifically, based on the reliability and real-time requirements of the business, business instances are classified into guaranteed-level businesses and best-effort-level businesses. Based on the actual load, the predicted load, the guaranteed level quota, the best-effort level quota, the latency requirements of the service deployment location, and satellite orbital motion factors, service scheduling information is generated for the guaranteed level service and the best-effort level service.

[0038] In this embodiment, the hierarchical resource scheduling module takes the load status of each satellite and the two-level quota indicators as input to perform two-level priority scheduling of services, and is the entity that actually schedules services in the system. This module determines "which service is deployed at which node and with which priority". First, it performs service classification and matching, determining its level according to the reliability and real-time requirements of the service itself. High-reliability, high-real-time critical services are scheduled to the guarantee-level resource pool, while low-reliability, low-real-time, interruptible, or delayed services are scheduled to the best-effort-level resource pool. Second, it performs two independent scheduling layers: critical services are scheduled and deployed in the guarantee-level resource pool, providing reliable services with physical resource backing; elastic services are scheduled and deployed in the best-effort-level resource pool, providing best-effort services. This module outputs service scheduling information to the elastic mapping execution module.

[0039] What is understandable is that the core mechanism of the hierarchical resource scheduling module is to classify services and make two-level scheduling decisions within a two-tiered resource pool across multiple nodes. The service classification strategy can categorize services into guaranteed-level or best-effort-level services based on their reliability and real-time requirements. The two-level scheduling decision can independently schedule guaranteed-level and best-effort-level services based on factors such as satellite node load, latency requirements of service deployment locations, and satellite orbital motion.

[0040] Furthermore, in terms of resource tiering, in addition to the guaranteed and best-effort tiers, it can be expanded to three or more tiers, assigning different fulfillment commitments and preemption priorities to different tiers, achieving more granular service quality stratification. Service tiers can be explicitly declared upon access, automatically determined by the system based on service type, historical behavior, SLA (Service Level Agreement) tags, or a combination of both. The boundary between guaranteed and best-effort tiers can be dynamically adjusted based on the actual usage of guaranteed-tier services, allowing guaranteed-tier quotas to be flexibly reserved as needed, with the remainder released for best-effort tiers, further improving utilization.

[0041] In terms of two-layer scheduling, in addition to considering resource levels, the two-layer scheduling can comprehensively consider factors such as service latency requirements, deployment location, satellite orbital motion, inter-satellite link quality, and remaining node power. Appropriate scheduling strategies can be adopted for different objectives such as load balancing, optimal latency, optimal energy consumption, and minimal migration. The scheduling scope can be improved from centralized scheduling across the entire network to distributed collaborative scheduling by region and cluster, to adapt to large-scale constellation scenarios.

[0042] Step S40: Based on the resource dynamic mapping ratio, the guarantee level quota, the best effort level quota, and the service scheduling information, execute all actual actions for the operation of the relevant service deployment.

[0043] Specifically, based on the resource dynamic mapping ratio, fixed physical resources are mapped to a corresponding number of virtual resources. Based on the guaranteed level quota and the best-effort level quota, the virtual resources are used to establish and maintain a guaranteed level resource pool and a best-effort level resource pool. Based on the service scheduling information, the service instance is deployed to the guaranteed resource pool or the best-effort resource pool for operation; When both guaranteed-level and best-effort-level services exist simultaneously, adjust cgroup resource limits to prioritize the needs of guaranteed-level services. When the demand for guaranteed services increases further, the deployed best-effort services will be temporarily suspended to free up physical computing resources. If the best-effort services are suspended for a period of time exceeding a second preset threshold, the best-effort services will be migrated to other satellite nodes with available resources.

[0044] Understandably, the core mechanism of the elastic mapping execution module is to map physical resources to virtual resources based on the calculation formulas for guaranteed and best-effort quotas, establish and maintain guaranteed and best-effort resource pools for each node, and deploy service instances to designated resource pools for operation based on scheduling decisions. When both guaranteed and best-effort services exist simultaneously, the cgroup resource limits are adjusted to prioritize the needs of guaranteed services. When the demand for guaranteed services increases further, the deployed best-effort services are temporarily suspended to release more physical computing resources. If the best-effort service remains suspended for an extended period, a rescheduling request should be sent to the hierarchical resource scheduling module to migrate the service to other satellite nodes with available resources.

[0045] Furthermore, the onboard computing power elastic pooling method based on dynamic resource mapping ratio also includes: When physical resources are scarce, based on the underlying resource isolation and preemption mechanism, at least one of the following actions is performed on best-effort level services: restriction action, suspension action, eviction action, and reclamation action, to ensure that guarantee level services have priority in occupying physical resources. The limiting action is to adjust cgroup resource limits to reduce the resource consumption of best-effort level services; The suspension action is to temporarily pause the deployed best-effort level services to free up physical computing resources; The eviction action is to forcibly terminate the best-effort level service instance to free up the occupied physical resources. The recycling action involves reclaiming the virtual resources already allocated to the best-effort level services and correspondingly reducing the best-effort level resource pool.

[0046] In this embodiment, based on the two-level quota indicators and business scheduling information, it executes all the actual actions of the relevant business deployment and is the only execution entity that actually operates on physical / virtual resources and business instances.

[0047] Elastic mapping execution: based on the current mapping ratio output by the dynamic mapping ratio decision module. It maps fixed physical computing resources to a corresponding number of virtual resources, enabling the total amount of external virtual resources to expand and contract elastically; and establishes and maintains two-level resource pools of the current node, namely the guaranteed level and the best-effort level, based on two-level quota indicators; when the mapping ratio is increased, the best-effort level resource pool is expanded, and when the mapping ratio is decreased, the best-effort level resource pool is reduced.

[0048] Business deployment and execution: Based on the two-tier scheduling decisions issued by the hierarchical resource scheduling module, business instances are actually deployed to the specified level resource pool for operation.

[0049] Prioritization and recovery of execution: By connecting to the underlying resource isolation and prioritization mechanism, when physical resources are scarce, actions such as restricting, suspending, evicting, and recovering best-effort level services are taken to ensure that guarantee level services have priority in occupying physical resources, thereby ensuring that critical business commitments are not affected by dynamic over-division and run-out under any load.

[0050] Furthermore, resource isolation methods, besides CPU quota / share limits based on cgroups, can be replaced by container runtime resource limits, virtual machine resource quotas, lightweight virtualization, or single-core resource constraints. In addition to the three levels of "quota reduction, suspension, and rescheduling migration," the yielding gradient can be modified based on business characteristics, such as direct rerouting for stateless services and checkpointing / recovery migration for checkpointable services. Cross-node migration of best-effort services can employ any method, such as virtual machine hot migration, container migration, task-level migration, or transfer of intermediate computation results. The selection order of reclaimed services can be based on single or combined criteria such as service priority, remaining execution time, runtime, and migration cost.

[0051] Furthermore, regarding the deployment architecture, the deployment of each module can be flexibly configured. Load awareness, dynamic mapping ratio decision-making, and elastic mapping execution are best deployed on each satellite node, while hierarchical resource scheduling can be centrally deployed at the system level or decentralized to a distributed manner. In scenarios where satellite-to-ground links are limited, more decision-making capabilities can be decentralized to the satellite to enhance system autonomy and resilience. This invention can be applied to the local resource elastic pooling of a single satellite, as well as to the pooling of network-wide computing resources composed of multiple satellites.

[0052] In terms of application scenarios, this invention is not limited to low-Earth orbit constellations, but can be extended to any scenario where nodes move periodically, loads vary over time, and physical resources are limited, such as medium-Earth orbit and mixed high-low orbit constellations. Its core concept of "load-driven dynamic mapping ratio + hierarchical management" can also be extended to other resource-constrained scenarios such as edge computing and mobile platform computing where loads vary drastically over time.

[0053] In other words, the core concept of this invention can be extensively improved and equivalently replaced in the above-mentioned stages. All these improvements and replacements do not deviate from the core idea of ​​this invention, namely, "load-driven dynamic resource mapping ratio + guaranteed or best-effort level resource hierarchical management", and should all fall within the protection scope of this invention.

[0054] Furthermore, after the elastic mapping is completed, the system returns to step S10 to enter the next scheduling cycle for new load sensing and prediction. As satellites continue to pass over the area and the load continues to change, the system cyclically executes the above steps S10-S40. This ensures that the fixed onboard physical resources continuously appear to the outside world as a virtual computing resource pool that elastically scales with the load and is hierarchically controllable.

[0055] This embodiment takes a low-Earth orbit satellite node with onboard computing capabilities as an example to demonstrate how to achieve elastic pooling of computing power through dynamic mapping ratio and hierarchical management as it passes over different ground areas and its load changes accordingly.

[0056] In this embodiment, the total physical computing resources of the satellite node are set to 100, and the baseline mapping ratio is... That is, it provides The guaranteed resource pool, with dynamic mapping levels set to That is, additional services can be provided or The best-effort level resource pool. Utilization rate Set a low usage threshold of 40% and a high usage threshold of 70%, and adjust them upwards or downwards each time. Afterwards, at least one scheduling cycle is maintained. Based on the satellite's ephemeris, its transit trajectory and the ground areas it passes through at each time period are determined. This embodiment sets a simplified transit sequence, with the satellite transiting the ocean (low service), nearshore (medium service), urban (high service), and then the suburbs (medium service) in sequence. Based on this, a priori mapping between transit time periods and physical resource utilization is constructed for online prediction. When determining the service level, high-reliability, high-real-time services (such as emergency communication support) are classified as guaranteed-level services; low-reliability, low-real-time, and interruptible services (such as remote sensing data batch processing) are classified as best-effort-level services.

[0057] This example uses the gear selection decision table shown in Table 1 to determine the gear position. Gear shifting strategy: Table 1: Gear Selection Decision Table

[0058] The decision-making results for each time period when the satellite passes over the above-mentioned area are shown in Table 2.

[0059] Table 2: Decision Results for Each Time Period

[0060] As can be seen, using a dynamic mapping ratio, compared to using a fixed baseline mapping ratio, In this instance, an average of 66.7 additional effort pool resources were obtained, providing computing power support for effort-level services and effectively improving the utilization of physical resources.

[0061] The beneficial effects of this invention are as follows: 1. By using dynamic mapping ratios, the dilemma of "peak collapse and trough waste" caused by fixed mapping ratios can be resolved, significantly improving the utilization rate of scarce spaceborne physical resources.

[0062] This invention uses a dynamic mapping ratio decision mechanism to improve the resource mapping ratio. This invention features a reverse load-matching mechanism: the mapping ratio is increased during periods of low load and decreased during periods of high load. The underlying principle is that a fixed mapping ratio can only achieve a static compromise value. However, satellite transit causes a dramatic shift in load between low-density areas and high-density areas, resulting in opposing demands on the mapping ratio. Any fixed value would inevitably come at a cost at one extreme. This invention reverses the load-matching principle: when transiting low-density areas and the load decreases, the mapping ratio is increased to map previously idle physical resources into more virtual resources and access more services, fully utilizing idle resources. Conversely, when transiting high-density areas and the load increases, the mapping ratio is decreased to tighten the commitment to virtual resources and avoid exceeding physical limits. Thus, without increasing any physical resources, the fixed onboard physical resources are presented as a virtual resource supply that elastically scales with the load, significantly improving resource utilization.

[0063] 2. The decision-making process uses a two-factor approach of "actual utilization rate + predicted utilization rate" to achieve forward-looking adjustments and avoid misjudgments and fluctuations based solely on the current load.

[0064] The mapping ratio decision of this invention is based not only on the current measured physical resource utilization rate It also incorporates the predicted load changes for the next stage. The allocation adjustment is determined by two factors. The underlying principle is that adjusting solely based on current load can lead to misjudgments. While current utilization may be low, a sudden increase in the mapping ratio could force a large-scale recovery of best-effort level services in the next phase, causing frequent fluctuations and resource shortages. This invention introduces load prediction based on satellite transit patterns as a forward-looking factor. It only actively increases the allocation when actual utilization is low and the predicted utilization is stable or declining; conversely, it maintains the current allocation level and reserves room for future adjustments when actual utilization is low but the predicted utilization is rising. Since satellite transit trajectories are precisely known from ephemeris data and the load in transit areas is predictable, this forward-looking decision-making is well-founded. This approach pursues high utilization while avoiding excessive over-resolution and allocation fluctuations, ensuring smooth and reliable resource adjustments.

[0065] 3. Implement resource tiered management to dynamically control the risk of over-allocation and ensure both high utilization and high reliability of critical business operations.

[0066] This invention divides external virtual resources into two levels: a guaranteed level and a best-effort level. The guaranteed level limit is determined by a baseline mapping ratio. Corresponding physical resources will be guaranteed based on peak demand and will definitely be fulfilled; the best-effort level credit limit will be... Exceeding The dynamic over-provisioning component is not guaranteed to be continuously fulfilled. Its technical principle lies in the fact that while dynamically increasing the mapping ratio improves utilization, the over-provisioning commitment may fail to materialize when the load recovers, potentially triggering a run on the bank. This invention allocates all the benefits of dynamic over-provisioning to the best-effort level, while the guaranteed level remains consistently at... The safety level ensures reliability equivalent to existing conservative peak-reserve levels. When load increases and physical resources become scarce, the system prioritizes ensuring that essential services occupy physical resources, only yielding to best-effort services. Thus, the service commitment of critical services is unaffected by dynamic over-division overload under any load, achieving a balance between "high utilization" and "high reliability of critical services"—something existing static over-division technologies cannot simultaneously achieve.

[0067] 4. By employing a tiered runtime execution mechanism, critical business operations are protected while minimizing the impact on elastic business operations.

[0068] This invention employs a progressive approach to resource allocation for best-effort services under conditions of physical resource scarcity: adjusting cgroup resource limits → suspending → rescheduling and migration. The underlying principle is that terminating best-effort services immediately upon encountering resource constraints leads to frequent interruptions of elastic services and wasted computing power. This invention uses a progressive allocation method, first reducing the physical share of best-effort services by lowering cgroup resource limits; suspending them only when insufficient resources remain; and only initiating rescheduling after prolonged suspension to migrate them to other satellite nodes with available resources. This gradient increases the cost of resource allocation progressively with the level of scarcity, ensuring priority for essential services while maximizing the continuity of best-effort services and minimizing computing power loss.

[0069] 5. It integrates with the existing scheduling system in a minimally invasive manner, making the project highly feasible and easy to implement.

[0070] This invention only changes the available resource capacity seen by the scheduler at the dynamic mapping ratio level, and achieves resource tiering by leveraging mature underlying resource isolation and preemption mechanisms. Its technical principle lies in the innovation of this invention, which focuses on the upper-level mechanism of "dynamic mapping ratio decision + two-level quota + tiered transfer." The resource mapping, resource isolation, service suspension, and migration mechanisms upon which its implementation depends are all mature technologies; this invention organizes and applies them rather than reconstructing them. Therefore, this invention does not require rewriting the underlying virtualization and scheduling core, and can be incrementally deployed on existing onboard computing resource pools and cloud scheduling systems. It has minimal engineering invasiveness, high feasibility, and is easy to implement in actual constellation systems.

[0071] Furthermore, such as Figure 3 As shown, based on the above-mentioned spaceborne computing power elastic pooling method based on dynamic resource mapping ratio, the present invention also provides a spaceborne computing power elastic pooling system based on dynamic resource mapping ratio, wherein the spaceborne computing power elastic pooling system based on dynamic resource mapping ratio includes: The load perception and prediction module 51 is used to obtain the operating status of the deployed services in the resource pool, obtain the current actual load of the resource pool based on the operating status, and obtain the predicted load based on the satellite transit trajectory and the distribution pattern of the ground area traversed by the satellite transit trajectory. The dynamic mapping ratio decision module 52 is used to determine the current dynamic resource mapping ratio based on the actual load and the predicted load, and output the guarantee level quota and the best effort level quota. The hierarchical resource scheduling module 53 is used to perform two-level priority scheduling of services based on the actual load, the predicted load, the guaranteed level quota, and the best effort level quota to obtain service scheduling information. The elastic mapping execution module 54 is used to execute all actual actions of the relevant business deployment based on the resource dynamic mapping ratio, the guarantee level quota, the best effort level quota and the business scheduling information.

[0072] Furthermore, such as Figure 4 As shown, based on the above-mentioned method and system for elastic pooling of onboard computing power based on dynamic resource mapping ratio, the present invention also provides a terminal, which includes a processor 10, a memory 20 and a display 30. Figure 4 Only some of the terminal components are shown; however, it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0073] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard disk or memory. In other embodiments, the memory 20 may be an external storage device of the terminal, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the terminal. Further, the memory 20 may include both internal and external storage devices. The memory 20 is used to store application software and various types of data installed on the terminal, such as program code installed on the terminal. The memory 20 can also be used to temporarily store data that has been output or will be output. In one embodiment, the memory 20 stores a spaceborne computing power elastic pooling program 40 based on a dynamic resource mapping ratio. This spaceborne computing power elastic pooling program 40 based on a dynamic resource mapping ratio can be executed by the processor 10, thereby implementing the spaceborne computing power elastic pooling method based on a dynamic resource mapping ratio in this application.

[0074] In some embodiments, the processor 10 may be a central processing unit (CPU), a microprocessor, or other data processing chip, used to run program code stored in the memory 20 or process data, such as executing the spaceborne computing power elastic pooling method based on dynamic resource mapping ratio.

[0075] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display 30 is used to display information on the terminal and to display a visual user interface. The components of the terminal communicate with each other via a system bus.

[0076] In one embodiment, when the processor 10 executes the onboard computing power elastic pooling program 40 based on the dynamic resource mapping ratio in the memory 20, the following steps are implemented: Obtain the operational status of deployed services in the resource pool, determine the current actual load of the resource pool based on the operational status, and obtain the predicted load based on the satellite transit trajectory and the distribution pattern of the ground areas traversed by the satellite transit trajectory. Based on the actual load and the predicted load, determine the current dynamic resource mapping ratio and output the guarantee level quota and the best effort level quota. Based on the actual load, the predicted load, the guaranteed level limit, and the best effort level limit, the service is scheduled with two levels of priority to obtain service scheduling information. Based on the dynamic resource mapping ratio, the guarantee level limit, the best effort level limit, and the service scheduling information, execute all actual actions of the relevant service deployment.

[0077] The process of obtaining the operational status of deployed services in the resource pool, determining the current actual load of the resource pool based on the operational status, and obtaining the predicted load based on the satellite transit trajectory and the distribution patterns of the ground areas traversed by the satellite transit trajectory, specifically includes: The actual occupancy of physical computing resources by deployed services in the onboard computing resource pool is periodically collected, and the actual occupancy is aggregated and normalized to obtain the measured utilization rate of physical resources by the current satellite load. ; Based on known ephemeris parameters, the satellite transit area is obtained through orbit calculations. Then, based on the distribution patterns of on-orbit users and service types within this area, the trend of physical resource utilization for deployed services over a future period is predicted. ; Based on the measured usage rate and the aforementioned trend of change The predicted load is obtained. .

[0078] Specifically, determining the current dynamic resource mapping ratio based on the actual load and the predicted load, and outputting the guarantee level quota and the best-effort level quota, includes: Obtain the total physical computing power resources of the satellite nodes, and configure the resource baseline mapping ratio, initial resource dynamic mapping ratio, and dynamic mapping ratio level for the satellite nodes; Based on the total physical computing power resources and the resource baseline mapping ratio, calculate and output the guarantee level quota: ; Wherein, P represents the total amount of physical computing resources. This represents the resource baseline mapping ratio. This indicates the coverage level limit; Based on the actual load and the predicted load, the initial resource dynamic mapping ratio is adjusted according to the dynamic mapping ratio level to obtain the resource dynamic mapping ratio. Based on the total physical computing resources, the resource baseline mapping ratio, and the resource dynamic mapping ratio, calculate and output the best-effort quota: ; in, This indicates the dynamic mapping ratio of the resources. This indicates the effort level limit.

[0079] Specifically, the step of adjusting the initial dynamic resource mapping ratio based on the actual load and the predicted load, and according to the dynamic mapping ratio level, to obtain the dynamic resource mapping ratio includes: When the measured usage rate The predicted load is below a first preset threshold. At that time, according to the preset adjustment rules, the initial resource dynamic mapping ratio is adjusted upward based on the dynamic mapping ratio, to obtain the resource dynamic mapping ratio used to expand the best effort level quota. When the measured usage rate The predicted load is below a first preset threshold. At the same time, the initial dynamic resource mapping ratio is kept unchanged, and the initial dynamic resource mapping ratio is used as the dynamic resource mapping ratio to reserve room for the realization of the next stage of load increase; When the measured usage rate The predicted load is higher than the first preset threshold. At the same time, the initial dynamic resource mapping ratio is kept unchanged, and the initial dynamic resource mapping ratio is used as the dynamic resource mapping ratio to maintain the efficient state of full utilization of existing physical resources; When the measured usage rate The predicted load is higher than the first preset threshold. At that time, according to the preset adjustment rules, the initial resource dynamic mapping ratio is adjusted downward based on the dynamic mapping ratio, so as to obtain the resource dynamic mapping ratio used to shrink the best effort level quota.

[0080] Specifically, the process of scheduling services based on the actual load, the predicted load, the guaranteed load limit, and the best-effort load limit to obtain service scheduling information includes: Based on the reliability and real-time requirements of the business, business instances are classified into guaranteed-level business and best-effort-level business. Based on the actual load, the predicted load, the guaranteed level quota, the best-effort level quota, the latency requirements of the service deployment location, and satellite orbital motion factors, service scheduling information is generated for the guaranteed level service and the best-effort level service.

[0081] Specifically, the step of executing all actual actions related to the deployment of relevant services based on the dynamic resource mapping ratio, the guarantee level limit, the best-effort level limit, and the service scheduling information includes: Based on the resource dynamic mapping ratio, fixed physical resources are mapped to a corresponding number of virtual resources. Based on the guaranteed level quota and the best-effort level quota, the virtual resources are used to establish and maintain a guaranteed level resource pool and a best-effort level resource pool. Based on the service scheduling information, the service instance is deployed to the guaranteed resource pool or the best-effort resource pool for operation; When both guaranteed-level and best-effort-level services exist simultaneously, adjust cgroup resource limits to prioritize the needs of guaranteed-level services. When the demand for guaranteed services increases further, the deployed best-effort services will be temporarily suspended to free up physical computing resources. If the best-effort services are suspended for a period of time exceeding a second preset threshold, the best-effort services will be migrated to other satellite nodes with available resources.

[0082] The onboard computing power elastic pooling method based on dynamic resource mapping ratio further includes: When physical resources are scarce, based on the underlying resource isolation and preemption mechanism, at least one of the following actions is performed on best-effort level services: restriction action, suspension action, eviction action, and reclamation action, to ensure that guarantee level services have priority in occupying physical resources. The limiting action is to adjust cgroup resource limits to reduce the resource consumption of best-effort level services; The suspension action is to temporarily pause the deployed best-effort level services to free up physical computing resources; The eviction action is to forcibly terminate the best-effort level service instance to free up the occupied physical resources. The recycling action involves reclaiming the virtual resources already allocated to the best-effort level services and correspondingly reducing the best-effort level resource pool.

[0083] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a spaceborne computing power elastic pooling program based on dynamic resource mapping ratio, and the spaceborne computing power elastic pooling program based on dynamic resource mapping ratio implements the steps of the spaceborne computing power elastic pooling method based on dynamic resource mapping ratio as described above when executed by a processor.

[0084] In summary, this invention provides a method, system, terminal, and storage medium for elastic pooling of onboard computing power based on dynamic resource mapping ratio. The method includes: acquiring the operational status of deployed services in the resource pool; obtaining the current actual load of the resource pool based on the operational status; and obtaining the predicted load based on the satellite transit trajectory and the distribution pattern of the ground areas traversed by the satellite transit trajectory; determining the currently used dynamic resource mapping ratio based on the actual load and the predicted load, and outputting the guaranteed level quota and the best-effort level quota; performing two-level priority scheduling of services based on the actual load, the predicted load, the guaranteed level quota, and the best-effort level quota to obtain service scheduling information; and executing all actual actions for the deployment and operation of relevant services based on the dynamic resource mapping ratio, the guaranteed level quota, the best-effort level quota, and the service scheduling information. This invention significantly improves the utilization rate of scarce onboard physical resources while ensuring that the service quality of critical services is not affected by resource constraints.

[0085] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal that includes that element.

[0086] Of course, those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0087] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A method for elastic pooling of onboard computing power based on dynamic resource mapping ratio, characterized in that, The onboard computing power elastic pooling method based on dynamic resource mapping ratio includes: Obtain the operational status of deployed services in the resource pool, determine the current actual load of the resource pool based on the operational status, and obtain the predicted load based on the satellite transit trajectory and the distribution pattern of the ground areas traversed by the satellite transit trajectory. Based on the actual load and the predicted load, determine the current dynamic resource mapping ratio and output the guarantee level quota and the best effort level quota. Based on the actual load, the predicted load, the guaranteed level limit, and the best effort level limit, the service is scheduled with two levels of priority to obtain service scheduling information. Based on the dynamic resource mapping ratio, the guarantee level limit, the best effort level limit, and the service scheduling information, execute all actual actions of the relevant service deployment.

2. The spaceborne computing power elastic pooling method based on dynamic resource mapping ratio according to claim 1, characterized in that, The process involves obtaining the operational status of deployed services in the resource pool, determining the current actual load of the resource pool based on the operational status, and obtaining the predicted load based on the satellite transit trajectory and the distribution patterns of the ground areas traversed by the satellite transit trajectory. Specifically, this includes: The actual occupancy of physical computing resources by deployed services in the onboard computing resource pool is periodically collected, and the actual occupancy is aggregated and normalized to obtain the measured utilization rate of physical resources by the current satellite load. ; Based on known ephemeris parameters, the satellite transit area is obtained through orbit calculations. Then, based on the distribution patterns of on-orbit users and service types within this area, the trend of physical resource utilization for deployed services over a future period is predicted. ; Based on the measured usage rate and the aforementioned trend of change The predicted load is obtained. .

3. The spaceborne computing power elastic pooling method based on dynamic resource mapping ratio according to claim 2, characterized in that, The step of determining the current dynamic resource mapping ratio based on the actual load and the predicted load, and outputting the guaranteed level quota and the best-effort level quota, specifically includes: Obtain the total physical computing power resources of the satellite nodes, and configure the resource baseline mapping ratio, initial resource dynamic mapping ratio, and dynamic mapping ratio level for the satellite nodes; Based on the total physical computing power resources and the resource baseline mapping ratio, calculate and output the guarantee level quota: ； Where P represents the total amount of physical computing resources, This represents the resource baseline mapping ratio. This indicates the coverage level limit; Based on the actual load and the predicted load, the initial resource dynamic mapping ratio is adjusted according to the dynamic mapping ratio level to obtain the resource dynamic mapping ratio. Based on the total physical computing resources, the resource baseline mapping ratio, and the resource dynamic mapping ratio, calculate and output the best-effort level quota: ； in, This indicates the dynamic mapping ratio of the resources. This indicates the effort level limit.

4. The spaceborne computing power elastic pooling method based on dynamic resource mapping ratio according to claim 3, characterized in that, The step of adjusting the initial dynamic resource mapping ratio based on the actual load and the predicted load, and according to the dynamic mapping ratio level, to obtain the dynamic resource mapping ratio, specifically includes: When the measured usage rate The predicted load is below a first preset threshold. At that time, according to the preset adjustment rules, the initial resource dynamic mapping ratio is adjusted upward based on the dynamic mapping ratio, to obtain the resource dynamic mapping ratio used to expand the best effort level quota. When the measured usage rate The predicted load is below a first preset threshold. At the same time, the initial dynamic resource mapping ratio is kept unchanged, and the initial dynamic resource mapping ratio is used as the dynamic resource mapping ratio to reserve room for the realization of the next stage of load increase; When the measured usage rate The predicted load is higher than the first preset threshold. At the same time, the initial dynamic resource mapping ratio is kept unchanged, and the initial dynamic resource mapping ratio is used as the dynamic resource mapping ratio to maintain the efficient state of full utilization of existing physical resources; When the measured usage rate The predicted load is higher than the first preset threshold. At that time, according to the preset adjustment rules, the initial resource dynamic mapping ratio is adjusted downward based on the dynamic mapping ratio, so as to obtain the resource dynamic mapping ratio used to shrink the best effort level quota.

5. The spaceborne computing power elastic pooling method based on dynamic resource mapping ratio according to claim 1, characterized in that, The process of scheduling services based on the actual load, the predicted load, the guaranteed load limit, and the best-effort load limit, using a two-tier priority system, yields service scheduling information, specifically including: Based on the reliability and real-time requirements of the business, business instances are classified into guaranteed-level business and best-effort-level business. Based on the actual load, the predicted load, the guaranteed level quota, the best-effort level quota, the latency requirements of the service deployment location, and satellite orbital motion factors, service scheduling information is generated for the guaranteed level service and the best-effort level service.

6. The spaceborne computing power elastic pooling method based on dynamic resource mapping ratio according to claim 1, characterized in that, The process of executing all actual actions related to the deployment of relevant services based on the dynamic resource mapping ratio, the guarantee level limit, the best effort level limit, and the service scheduling information specifically includes: Based on the resource dynamic mapping ratio, fixed physical resources are mapped to a corresponding number of virtual resources. Based on the guaranteed level quota and the best-effort level quota, the virtual resources are used to establish and maintain a guaranteed level resource pool and a best-effort level resource pool. Based on the service scheduling information, the service instance is deployed to the guaranteed resource pool or the best-effort resource pool for operation; When both guaranteed-level and best-effort-level services exist simultaneously, adjust cgroup resource limits to prioritize the needs of guaranteed-level services. When the demand for guaranteed services increases further, the deployed best-effort services will be temporarily suspended to free up physical computing resources. If the best-effort services are suspended for a period of time exceeding a second preset threshold, the best-effort services will be migrated to other satellite nodes with available resources.

7. The spaceborne computing power elastic pooling method based on dynamic resource mapping ratio according to claim 6, characterized in that, The onboard computing power elastic pooling method based on dynamic resource mapping ratio further includes: When physical resources are scarce, based on the underlying resource isolation and preemption mechanism, at least one of the following actions is performed on best-effort level services: restriction action, suspension action, eviction action, and reclamation action, to ensure that guarantee level services have priority in occupying physical resources. The limiting action is to adjust cgroup resource limits to reduce the resource consumption of best-effort level services; The suspension action is to temporarily pause the deployed best-effort level services to free up physical computing resources; The eviction action is to forcibly terminate the best-effort level service instance to free up the occupied physical resources. The recycling action involves reclaiming the virtual resources already allocated to the best-effort level services and correspondingly reducing the best-effort level resource pool.

8. A spaceborne computing power elastic pooling system based on dynamic resource mapping ratio, characterized in that, The onboard computing power elastic pooling system based on dynamic resource mapping ratio includes: The load perception and prediction module is used to obtain the operating status of the deployed services in the resource pool, obtain the current actual load of the resource pool based on the operating status, and obtain the predicted load based on the satellite transit trajectory and the distribution pattern of the ground areas traversed by the satellite transit trajectory. The dynamic mapping ratio decision module is used to determine the current dynamic resource mapping ratio based on the actual load and the predicted load, and output the guarantee level quota and the best effort level quota. The hierarchical resource scheduling module is used to perform two-level priority scheduling of services based on the actual load, the predicted load, the guaranteed level quota, and the best effort level quota to obtain service scheduling information. The elastic mapping execution module is used to execute all actual actions of the relevant business deployment based on the resource dynamic mapping ratio, the guarantee level quota, the best effort level quota, and the business scheduling information.

9. A terminal, characterized in that, The terminal includes: a memory, a processor, and a spaceborne computing power elastic pooling program based on dynamic resource mapping ratio stored in the memory and executable on the processor. When the spaceborne computing power elastic pooling program based on dynamic resource mapping ratio is executed by the processor, it implements the steps of the spaceborne computing power elastic pooling method based on dynamic resource mapping ratio as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a spaceborne computing power elastic pooling program based on dynamic resource mapping ratio. When the spaceborne computing power elastic pooling program based on dynamic resource mapping ratio is executed by a processor, it implements the steps of the spaceborne computing power elastic pooling method based on dynamic resource mapping ratio as described in any one of claims 1-8.

Images