Space Super Intelligent Fusion Computing System

Abstract

The present invention relates to the field of intelligent computing and discloses a space super-intelligent fusion computing power system, which includes: a computing subsystem, which is used to determine computing resources, receive target jobs, and process the target jobs based on the computing resources if the computing resources meet the computing requirements of the target jobs; a storage subsystem, which is used to determine read-write characteristics and corresponding read-write paths when executing the target jobs, and store data corresponding to the target jobs based on the read-write paths; a communication subsystem, which is used to establish a first communication channel between subsystems and a second communication channel with external systems; a global management subsystem, which is used to monitor the operating status of each subsystem. If the operating status of at least one subsystem is a fault state, the task instructions corresponding to the fault state are determined based on a rule engine, and the instructions are allocated to the corresponding intelligent bodies, thereby achieving the effect of integrating powerful computing power, efficient storage, stable communication, reliable energy supply and heat dissipation, and a solid hardware foundation.

Description

Technical Field

[0001] The present invention relates to the field of intelligent computing, and in particular to a space super-intelligent fusion computing system. Background Art

[0002] In astronomy and cosmology research, it is necessary to process massive amounts of celestial data to explore the mysteries of the universe, such as analyzing the evolution of galaxies and the distribution of dark matter. Earth observation and environmental monitoring rely on powerful computing power to conduct real-time analysis of satellite images and other data in order to keep abreast of natural disasters, climate change, and other situations. Space science experiments and simulations, deep space exploration science, and other fields also place extremely high demands on computing performance.

[0003] However, existing space computing system technologies have many computing bottlenecks in terms of computing power, storage, communication, etc., and face profound technical contradictions between high performance and high reliability, heat dissipation and power consumption, communication delay and autonomous intelligence, making it difficult to fully and effectively carry out super-intelligent computing in a space environment.

[0004] In view of this, the present invention is proposed. Summary of the Invention

[0005] In order to solve the above technical problems, the present invention provides a space super-intelligent fusion computing power system, which integrates powerful computing power, efficient storage, stable communication, reliable energy supply and heat dissipation, and a solid hardware foundation.

[0006] An embodiment of the present invention provides a space super-intelligent fusion computing system, the system comprising:

[0007] The computing subsystem is used to determine the computing resources in the current deployment space, receive the target job, and process the target job based on the computing resources if the computing resources meet the computing requirements of the target job;

[0008] The storage subsystem is configured to, when executing a target job, determine read / write characteristics corresponding to the target job, determine a read / write path based on the read / write characteristics, and store data corresponding to the target job based on the read / write path; wherein the read / write characteristics are determined based on the read / write throughput, read / write burstiness, and source data operation volume of the target job;

[0009] The communication subsystem is used to establish a first communication channel between the subsystems and a second communication channel between the Space Super Intelligent Fusion Computing System and external systems;

[0010] The global management subsystem is used to monitor the operating status of each subsystem, and when the operating status of at least one subsystem is a fault state, it determines the task instructions corresponding to the fault state based on the rule engine, and executes each task instruction according to the intelligent agent corresponding to each task instruction to change the operating status to a normal state.

[0011] The embodiments of the present invention have the following technical effects: the computing resources in the current deployment space are determined through the computing subsystem, and the target job is received. When the computing resources meet the computing requirements of the target job, the target job is processed based on the computing resources. When executing the target job, the storage subsystem determines the read-write characteristics corresponding to the target job, and determines the read-write path based on the read-write characteristics, and stores the data corresponding to the target job based on the read-write path. The communication subsystem establishes a first communication channel between each subsystem and a second communication channel between the space super-intelligent fusion computing power system and the external system. The global management subsystem monitors the operating status of each subsystem, and when the operating status of at least one subsystem is a fault state, the rule engine determines the task instructions corresponding to the fault state, and executes each task instruction according to the intelligent agent corresponding to each task instruction to change the operating status to a normal state, thereby realizing the effect of integrating powerful computing power, efficient storage, stable communication, reliable energy supply and heat dissipation, and a solid hardware foundation in space, and effectively improving the performance of space intelligent computing. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0013] Figure 1 This is a schematic diagram of the structure of a space super-intelligent fusion computing system provided by an embodiment of the present invention;

[0014] Figure 2 This is a schematic diagram of the operating mode of a space super-intelligent fusion computing power system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0015] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention are described clearly and completely below. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort are also within the scope of protection of the present invention.

[0016] The space super-intelligent fusion computing system provided by the embodiments of the present invention is primarily suitable for performing sufficient computation, storage, communication, and other processing on target tasks in a space environment. The space super-intelligent fusion computing system provided by the embodiments of the present invention can be deployed in space stations or satellites of different structural types.

[0017] Figure 1 This is a schematic diagram of the structure of a space super-intelligent fusion computing system provided by an embodiment of the present invention. Figure 1 The space super-intelligent fusion computing power system specifically includes: a computing subsystem 110, a storage subsystem 120, a communication subsystem 130 and a global management subsystem 140.

[0018] The computing subsystem 110 is used to determine the computing resources in the current deployment space, receive the target job, and process the target job based on the computing resources if the computing resources meet the computing requirements of the target job.

[0019] The current deployment space refers to the space where the Space Super Intelligent Fusion Computing System is deployed, which can be within a space station or satellite. Computing resources refer to the resources available to the Space Super Intelligent Fusion Computing System for computing. Target jobs refer to jobs that require super intelligent computing processing by the Space Super Intelligent Fusion Computing System. Computing requirements refer to the resources required to complete the target job.

[0020] Specifically, the computing subsystem 110 can determine the computing resources within the current deployment space in real time and can receive target jobs that need to be processed in space. After receiving the target job, it is necessary to determine whether the computing resources within the current deployment space meet the computing requirements of the target job. If so, the target job can be executed within the current deployment space, that is, the target job can be processed based on the computing resources within the current deployment space. If not, it is necessary to coordinate with the ground scheduling server to reschedule and execute the target job.

[0021] The computing subsystem 110 provides powerful computing power and is the core functional component of the space super-intelligent fusion computing power system. In the ground-based and integrated construction methods, the computing subsystem 110 is composed of multiple computing modules. Each computing module contains a network server blade, a computing server blade, a monitoring blade and a power supply backplane. The network server blade is generally set in the middle of the module, and the monitoring blade is at the bottom. Among them, the network server blade is used to exchange data from the computing board in this module to the core switch component, and then interact with other computing modules or components; the computing server blade is set in a heterogeneous form within the board, and is composed of high-performance computing chips such as CPU (Central Processing Unit, Central Processing Unit), GPU (Graphics Processing Unit, Graphics Processing Unit), etc. Liquid cooling covers core components such as CPU, GPU chips and memory, providing super-intelligent fusion computing power. Each independent blade server has a traditional monitoring architecture that can detect internal and external performance indicators, which are collected uniformly and regularly by the monitoring blades within the module. At the same time, each node is equipped with high-capacity HBM (High Bandwidth Memory) and CXL SSD (Compute Express Link Solid-State Drive), but accessories are optional. It uses an optical communication network based on RDMA (Remote Direct Memory Access).

[0022] In the management of nodes, in order to ensure the high availability and reliability of the cluster, different computing queues can be set up to build a computing resource pool according to the different characteristics of space computing applications. A job feature-aware resource scheduling method can be used to allocate appropriate resources based on job characteristics. For example, computing resources can be divided into different computing queues such as computing area, training area, and inference area.

[0023] The storage subsystem 120 is used to determine the read / write characteristics corresponding to the target job when executing the target job, determine the read / write path based on the read / write characteristics, and store the data corresponding to the target job based on the read / write path.

[0024] Read and write characteristics are determined based on the target job's read and write throughput, read and write burstiness, and the source data operation volume. Read and write characteristics characterize data stability, specifically whether frequent updates are required. The read and write path is the path used to read and write data corresponding to the target job, which can be understood as different storage hardware and read and write processes.

[0025] Specifically, when executing a target job, the storage subsystem 120 can first determine the read / write characteristics of the target job based on the target job's read / write throughput, read / write burstiness, and source data operation volume. The read / write characteristics can be obtained through calculation, model processing, or other methods. Based on the read / write characteristics, the read / write characteristics corresponding to the target job are determined from the pre-established correspondence between the read / write characteristics and the read / write paths. Furthermore, data corresponding to the target job, such as initial data and data generated during execution, is stored based on the read / write paths.

[0026] The storage subsystem 120 includes an acceleration storage layer, a persistent storage layer, and a cold data storage layer. The number of storage blade servers in the persistent storage layer is smaller than the number of storage blade servers in the acceleration storage layer.

[0027] The read / write path includes a first path and a second path; the acceleration storage layer and the persistent storage layer serve as the first path, and the persistent storage layer serves as the second path; when the read / write path is the first path, a parallel file system is constructed based on the acceleration storage layer, data is stored based on the parallel file system, and after the target job is completed, the data in the parallel file system is copied to the persistent storage layer, and the parallel file system is destroyed;

[0028] The cold data storage layer is used to store packaged data corresponding to the generated data in the persistent storage layer when the stabilization time of the stored data in the persistent storage layer reaches a first time threshold.

[0029] The accelerated storage layer is a storage structure for fast read and write speeds and can be used to store hot data. The persistent storage layer is a storage structure for reading and writing data and has a large storage space, and can be used to store warm data. The cold data storage layer is a storage structure for infrequently accessed data and is used to store cold data. The stabilization time is the time the data has been in the persistent storage layer since its last read or write, that is, the time it has not been read or written. The first time threshold is the time value used to determine whether data needs to be transferred and stored in the cold data storage layer.

[0030] Specifically, when the read / write path is the first path, it is necessary to improve the data read / write efficiency through the acceleration storage layer, that is, to build a parallel file system based on the acceleration storage layer, store data based on the parallel file system, and stop using the acceleration storage layer after the target job is completed, that is, copy the data in the parallel file system to the persistent storage layer and destroy the parallel file system. When the read / write path is the second path, data is stored directly through the persistent storage layer. For the data stored in the persistent storage layer, the time is reset after each read / write to obtain a new stable time. When the stable time reaches the first time threshold, the data is determined to be cold data, and the corresponding packaged data is generated and sent to the cold data storage layer for storage.

[0031] The storage subsystem 120 further includes: a software image library.

[0032] The software image library is used to store the software image corresponding to each application so that when the application cannot run, the corresponding software image can be pulled from the software image library; when the update completion time of the software image library reaches the preset update cycle, the communication subsystem 130 obtains a new software image from the ground data server to update the software image library.

[0033] The update completion duration is the time interval since the last software image library update was completed. The preset update period is the period used to trigger software image library updates. The ground data server is a server located on the ground for storing various software images, which can be transmitted via the communication subsystem 130.

[0034] Exemplarily, the hardware of storage subsystem 120 is constructed with a multi-tiered storage structure consisting of an accelerated storage layer consisting of compute node HBM, a persistent storage layer with multiple backups of compute node local CXL SSDs, and an optical disk-based cold data storage layer. HBM is used within the compute nodes to accelerate data processing efficiency, and local high-speed CXL SSDs are mounted. Storage subsystem 120 includes separate storage modules installed in storage blade servers, which can be equipped with high-capacity NVME SSDs (Non-Volatile Memory Express Solid-State Drives).

[0035] Each space module is loaded with the same number of storage blade servers, and metadata management and configuration information management services are deployed on each storage blade server. The persistent storage capacity owned by all space modules is divided into multiple storage resource pools (N) with equal storage capacity. One of the storage resource pools is selected as the acceleration storage layer, one as the cold data storage layer, and the remaining N-2 as the persistent storage layer, which are used to build a multi-copy distributed file system and mount it on demand. The metadata management and configuration information management services of all space modules share metadata and configuration information.

[0036] At the software level, the accelerated storage layer can be constructed and destroyed during the target job cycle. A unified storage media controller manages the allocation and management of multi-layered storage media. When the target job is ready to execute, the storage media controller adjusts the read and write paths based on the target job's read and write characteristics, selecting resources from the resource pool with appropriate capacity for the target job. If the accelerated storage layer is used, a parallel file system is constructed and the original data is copied into the parallel file system. When the target job completes, the data is copied to the persistent storage layer, and the file system is destroyed.

[0037] The optical disk-based cold storage layer dynamically sets data migration rules for each job based on the last access time and data volume of all output files (data) of the target job. For example, after a certain time t (the stabilization time reaches the first time threshold), the output data of the target job is packaged and compressed, marked as cold data, and copied from the persistent storage layer to the cold data storage layer for storage. If the data of the target job is read or written again within time t (the first time threshold), the time is reset and the set data migration rules are executed.

[0038] Because updating software in space is difficult, storage subsystem 120 maintains a software image library. When an application fails due to a missing system library, the scheduler automatically pulls the software image from the library. Upon reboot, the kernel is restored to its original state. The software image library must regularly communicate with the ground (ground data server) for updates.

[0039] The communication subsystem 130 is used to establish a first communication channel between the subsystems and a second communication channel between the space super-intelligent fusion computing power system and the external system.

[0040] The first communication channel is used for internal communication within the Space Super Intelligent Fusion Computing System. The second communication channel is used for external communication within the Space Super Intelligent Fusion Computing System. This external system can be a Space Super Intelligent Fusion Computing System in another space station or satellite, or a ground-based system.

[0041] Based on the above example, the second communication channel includes an inter-satellite communication channel and a satellite-to-ground communication channel. The communication subsystem 130 is further configured to:

[0042] Based on the encrypted time-sensitive network, the data corresponding to the target job is transmitted.

[0043] The encrypted time-sensitive network (ETSN) extends encryption information to the network identification number field in Ethernet link layer data frames and adds a security tag field and an integrity check value field to the data frame. Intersatellite communication channels are used to communicate with other space stations or satellites in space. Satellite-to-ground communication channels are used to communicate with ground systems.

[0044] Specifically, when using the communication subsystem 130 for data transmission, an encrypted time-sensitive network is used for transmission to ensure the timeliness and security of inter-satellite communication and inter-satellite-ground communication.

[0045] The communication subsystem 130 actually consists of two parts: an external communication interface responsible for intersatellite and satellite-to-ground communications; and an internal communication interface primarily responsible for low-latency, high-speed communications within the satellite's large-scale, hyper-intelligent, converged computing system. The communication subsystem 130 is closely connected to the computing subsystem 110 and the storage subsystem 120, and collaborates with the global management subsystem 140 to ensure stable and reliable system communications.

[0046] For inter-satellite and satellite-to-ground communications, external communication modules, including radio frequency and laser communication modules, are deployed. These modules are adaptively matched based on performance and functional requirements to enable data upload, download, and command exchange. For intra-satellite communications, low-latency communication between chips, high-speed and reliable interconnection between nodes, and high-speed network interconnection between units are used to ensure efficient data transmission. The cluster network structure adopts a spine-leaf network architecture. For a uniform construction approach, each space module will be equipped with two customized switches. All devices will first connect to the customized leaf switches, which will then be interconnected with the spine switches within the module. Finally, the spine switches between modules are fully interconnected, achieving a modular effect. When a module needs to be replaced, all links related to the module are first logically isolated. After the load is removed, they are physically disconnected before the module is replaced.

[0047] In terms of communication protocols, the Cybersecurity Time Sensitive Network (CTSN) or Encrypted Time Sensitive Network (CTSN) was proposed to ensure the timeliness and security of inter-satellite and inter-satellite-ground communications. Through data encryption, communication protocols, and access control, it comprehensively ensures secure and efficient communications, preventing data leaks and unauthorized access. The CTSN network is an extension of the Ethernet link layer (Layer 2) protocol. It encrypts frame content by extending the PCP (Priority Code Point) / DEI (Drop Eligible Indicator) / VID (VLAN Identifier) bits in the VLAN Tag (Virtual Local Area Network Tag) field of Layer 2 packets, and then adding the SecTAG (Security Tag) and ICV (Integrity Check Value) fields.

[0048] Example, original layer 2 packet format:

[0049] │Dest MAC│Src MAC│Type / Length│VLAN Tag│Payload│FCS│;

[0050] CTSN Layer 2 packet format:

[0051] │Dest MAC│Src MAC│Type / Length│SecTAG│VLAN+Payload(encryption)│ICV│FCS│;

[0052] Among them, Dest MAC (Destination MAC Address), Src MAC (Source MAC Address), Type / Length (Type / Length field), FCS (Frame Check Sequence).

[0053] The computing subsystem 110 is further configured to: if the computing resources do not meet the computing requirements of the target job, send the computing resources and the computing requirements to the communication subsystem;

[0054] The communication subsystem 130 is also used to: receive computing resources and computing requirements, and send the computing resources and computing requirements to the ground scheduling server so that the ground scheduling server can determine the space collaboration system based on the computing resources and computing requirements; receive the space collaboration system corresponding to the computing requirements fed back by the ground scheduling server, and establish a second communication channel with the space collaboration system to collaboratively process the target operation based on the second communication channel, the space super-intelligent fusion computing power system in the current deployment space, and the space collaboration systems in other deployment spaces.

[0055] The space collaboration system is a hyper-intelligent fusion computing system in another deployment space, designed to collaborate with the hyper-intelligent fusion computing system in the current deployment space to complete target tasks. The ground dispatch server is a ground-based dispatch server that is used to dispatch the hyper-intelligent fusion computing system in the current deployment space to execute target tasks and allocate space collaboration systems for collaboration.

[0056] Specifically, if computing subsystem 110 determines that the current system's computing resources do not meet the computational requirements of the target job, this indicates that systems in other deployment spaces need to collaborate to process the target job. Therefore, the computing resources and computational requirements are sent to communication subsystem 130. Communication subsystem 130 receives the computing resources and computational requirements sent by computing subsystem 110 and sends them to the ground dispatch server. The ground dispatch server then determines, based on the computing resources and computational requirements, whether the space hyper-intelligent fusion computing power systems in other deployment spaces can collaborate to complete the target task. The ground dispatch server then determines which space collaborative systems are capable of collaboration and provides feedback to the communication subsystem 130 of the space hyper-intelligent fusion computing power system in the current deployment space. The ground dispatch server then receives the space collaborative system corresponding to the computational requirements from the ground dispatch server and establishes a second communication channel with the space collaborative system. Communication via this second communication channel allows the space hyper-intelligent fusion computing power system in the current deployment space and the space collaborative systems in other deployment spaces to collaboratively process the target job, thereby achieving multi-system collaboration.

[0057] For example, a unified scheduling server (ground scheduling server) is set up on the ground. Each deployment space is treated as a unit resource pool, which is added to the overall resource pool managed by the unified resource scheduling server. Each unit resource pool can independently provide computing services. When a new target job is ready to run in any unit resource pool, the unit resource pool first determines whether the job requires computing services from other resource pools. If so, it submits the job to the unified resource scheduler, which determines the specific resource pool location and prepares data. If communication is abnormal or other resource pools are not needed, the unit resource pool allocates its own computing resources for the calculation.

[0058] The global management subsystem 140 is used to monitor the operating status of each subsystem, and when the operating status of at least one subsystem is a fault state, based on the rule engine, determine the task instructions corresponding to the fault state, and execute each task instruction according to the intelligent agent corresponding to each task instruction to change the operating status to a normal state.

[0059] Among them, each subsystem includes a computing subsystem 110, a storage subsystem 120 and a communication subsystem 130. The operating status is used to describe whether the subsystem can work normally, including normal status and fault status. The rule engine receives data input, interprets rules and outputs decision results, and is often used to process complex and frequently changing logic. Task instructions are instructions to solve the current fault state. Each intelligent agent is an intelligent agent with embodied intelligent functions used to process different task instructions. Different intelligent agents have different functions, for example, they can include system perception agents, strategy planning agents, command operation execution agents, and hardware replacement track robots.

[0060] Specifically, the global management subsystem 140 monitors the operating status of the computing subsystem 110, storage subsystem 120, and communication subsystem 130. If at least one subsystem experiences a fault, the rules engine analyzes the cause of the fault and determines the task instructions to resolve the fault, which serve as the corresponding task instructions. Furthermore, the agent that executes each task instruction is identified as the corresponding agent. By executing each task instruction, the fault is automatically resolved, restoring the operating status from the faulty state to a normal state.

[0061] The global management subsystem 140 is further configured to:

[0062] Collect monitoring data from each subsystem regularly, analyze the monitoring data based on the rule engine, and determine the target processing instructions;

[0063] Based on the commander agent, it receives the target processing instructions pushed by the rule engine, analyzes and disassembles the target processing instructions, obtains task instructions, and sends each task instruction to the corresponding functional agent so that each functional agent executes the corresponding task instruction.

[0064] Monitoring data is collected from each subsystem to assess its operational status. Target processing instructions are instructions for resolving faults found in the monitoring data. The commander agent is responsible for overall planning, and the functional agents are agents that perform different functions according to the commander agent's requirements.

[0065] Specifically, the global management subsystem 140 regularly collects monitoring data from each subsystem. When at least one subsystem experiences a fault, the rules engine analyzes the monitoring data and identifies target processing instructions to resolve the corresponding fault. The commander agent receives target processing instructions pushed by the rules engine and analyzes and breaks them down into task instructions executable by different functional agents. Each task instruction is then sent to the corresponding functional agent, enabling each agent to execute the corresponding task instruction and resolve the fault.

[0066] Optionally, when determining the target processing instructions and disassembling the target processing instructions to obtain the task instructions, a large language model and a knowledge retrieval enhancement strategy may be used for determination.

[0067] The global management subsystem 140 is responsible for monitoring and scheduling, and is the core hub for ensuring the stable and efficient operation of the Space Super Intelligent Fusion Computing System. A system intelligence platform is deployed within the Space Super Intelligent Fusion Computing System to monitor the status of each component (monitoring data) in real time, using intelligent diagnostic technology to promptly detect and resolve faults, ensuring system reliability. This system intelligence platform regularly collects data from the monitoring blades of each computing module in each computing subsystem 110. Probes are deployed at each node in the modules of other subsystems to regularly obtain data from these probes.

[0068] The global management subsystem 140 can use a multi-agent system with embodied intelligent functions to perform intelligent system management, set up multiple agents, set up a commander agent on the system intelligent platform, receive target processing instructions pushed by the system intelligent platform through the rule engine, and the commander agent analyzes and disassembles the instructions to obtain various task instructions. According to task requirements, the commander agent distributes each task instruction to the system perception agent, strategy planning agent, command operation execution agent, and hardware replacement track robot, etc., so as to ensure the normal operation of the system.

[0069] Build a network security management system through a software-defined approach. Establish a hierarchical intelligent management and control system, utilizing system intelligence algorithms to achieve real-time monitoring of the status of each system component, predict faults, and automatically resolve them, ensuring stable and efficient system operation. Upon receiving event notifications, an on-orbit robot will perform knowledge retrieval based on a large language model to enhance policy reasoning and identify the most appropriate solution. This will automatically handle online software and hardware failures, eliminating manual maintenance and reducing reliance on expensive space-based hardware reliability measures such as radiation hardening.

[0070] In order to enable the space super-intelligent fusion computing system to be adapted for deployment in space stations or satellites of different structural types, three different system composition methods are set up, which can be called: integral type, ground-based type and uniform type.

[0071] The integrated construction method does not use the space module blocks in the space station or satellite, but assembles the server into a whole using a customized rack on the ground, and then places the whole in the space facility (deployment space). This method requires little modification to the ground computing system and is relatively low in cost. It is suitable for placement in a cabin-type space station with a good environment and sufficient space. However, due to the overall placement, it is not suitable for satellites such as deployable structures.

[0072] Generally speaking, deployable satellites have lower launch and transportation costs than space stations, making them a more cost-effective option for large-scale computing systems. However, for space-based super-intelligent fusion computing systems, the monolithic construction method cannot be used for deployable satellites due to its monolithic nature. Ground-based construction methods are more suitable and have moderate modification costs. The ground-based construction method involves functionally dividing the functions of each space module, placing equipment with different functions in different space modules. Before launch, the equipment is pre-installed on an available mounting surface (for ease of description, the term "space module" will be used hereafter). It is then folded and placed within the satellite's main structure. After entering orbit and completing attitude capture, the satellite unfolds to its operational state.

[0073] Assume that there are M space modules, each of which is used to place the components of each subsystem of the space super-intelligent fusion computing system. According to the preset computing power demand C and the computing power (computing amount) of a single computing blade server C s The number of computing blade servers Q that can be placed in a single space module, the number of modules N required for the computing subsystem 110 required for calculation c : .

[0074] Based on the preset storage demand S (including the capacity provided by all types of storage servers such as SSD servers and HDD servers) and the storage capacity list L of the storage blade server (the capacity provided by different servers varies depending on the storage medium), L=[s1, s2...s n ] and the number of storage blade servers R that can be placed in a single space module, calculate the number of space modules N required for the storage subsystem 120 p : .in, The total number of servers with different storage media required can be determined.

[0075] According to the preset computing power demand C, the computing capacity of a single computing blade server C s , preset storage requirement S, storage capacity list L and the number of network blade servers T that can be placed in a single space module, determine the number of space modules N required for the network core switch (communication subsystem 130) s : Where g(a, b) represents the number of required network core switches calculated based on the network structure and port number, a represents the number of computing servers, and b represents the number of storage servers.

[0076] In order to ensure that the space super intelligent fusion computing system has sufficient robustness, set N m (N m ≥2) Space Modules are used for the global management module to house the hardware required for the global management subsystem. The aforementioned number of servers that can be placed in a single space module is based on the assumption that each space module also requires a switch blade server. This blade server is used to communicate with the core switch and interconnect with the network structure in a fat tree. The number of space modules required for each subsystem must also meet the following restrictions: .

[0077] Based on the above examples, taking the foundation type as an example, there is a system rack in the current deployment space, and the system rack is evenly distributed with multiple space modules, and each subsystem is placed on at least one space module; the first number of space modules corresponding to the computing subsystem 110 is determined according to the preset computing power requirement, the computing amount corresponding to a single computing blade server, and the maximum number of computing blade servers corresponding to each space module; the second number of space modules corresponding to the storage subsystem 120 is determined according to the preset storage requirement, the storage capacity of the storage blade servers corresponding to each storage medium, and the maximum number of storage blade servers corresponding to each space module; the third number of space modules corresponding to the communication subsystem 130 is determined according to the preset computing power requirement, the computing amount corresponding to a single computing blade server, the preset storage requirement, the storage capacity of the storage blade servers corresponding to each storage medium, and the maximum number of network blade servers corresponding to each space module; the fourth number of space modules corresponding to the global management subsystem 140 is at least two; the sum of the first number, the second number, the third number, and the fourth number is less than or equal to the total number of space modules on the system rack.

[0078] The first number is the , the second quantity is in the above example, the third number is in the above example , the fourth number is the .

[0079] Based on the above example, we can rationally plan the subsystems for each space module in the Space Super Intelligent Fusion Computing System. Specifically, we can:

[0080] The space modules in the middle area of the system rack are used to place the communication subsystem 130, the space modules in the outermost area of the system rack are used to place the global management subsystem 140, the space modules in the second outermost area of the system rack are used to place the storage subsystem 120, and the space modules in other areas of the system rack are used to place the computing subsystem 110.

[0081] In order to achieve high performance in a small space, a high-density assembly method is adopted as the whole. The servers involved in all modules are installed in the form of server blades. Each module is provided with a backplane, which provides communication and power supply access.

[0082] For the uniform construction method, each space module has completely consistent computing power, storage capacity, and network capacity, and these three types of blade servers need to be deployed simultaneously in one space module. Since it is difficult to maintain and update equipment in space, a modular design is usually adopted to improve maintainability. This requires that the machine groups in each space module need to be structurally independent and have standardized interfaces, that is, each machine group needs to contain equivalent computing units, storage units, network units, and management units, etc. In this way, if a module can no longer provide services, it only needs to be replaced with a new module. If there are M space modules, such as the preset computing power demand C, the preset storage demand S, and the network interaction capability N of the entire space super-intelligent fusion computing power system, each module should have a computing power of C. s =C / M, storage capacity S s =S / M, network interaction capability N s =N / M.

[0083] When adopting the monolithic and foundation-based construction, the global management node is placed in the global management module. However, for the normalized approach, the global management module hardware is deployed in each spatial module, and the resource scheduling service is deployed in the global management module. Each service manages the computing resources in all modules, and the status information required by the service is guaranteed based on the RAFT protocol.

[0084] Specifically, there are three hardware configuration options: one is to directly use the entire space super-intelligent fusion computing power system as a module, which can be deployed in a module-type space station; one is to build space modules with different functions on available installation planes based on the operational needs of the space super-intelligent fusion computing power system, and then combine them into a complete system, which can be deployed in a module-type space station or a satellite with an expandable structure; the last is to build a unique space module installed on all available installation planes, with all space modules having the same function, suitable for deployment in satellites with expandable structures such as triangular prisms and stacked configurations. Logically, various subsystem modules are divided into various types according to their functions, including hardware subsystems such as the computing subsystem 110, storage subsystem 120, communication subsystem 130, global management subsystem 140, energy and thermal control subsystem and supporting infrastructure, and software subsystems such as the application software subsystem, system software subsystem, and global management subsystem 140. Each subsystem works together and is closely related to ensure the stable operation and efficient performance of the space computing power system, filling the gap in the current method of building a centralized large-scale space super-intelligent fusion computing power system.

[0085] Based on the above example, the Space Super Intelligent Fusion Computing System also includes: an energy and thermal control management subsystem; the energy and thermal control management subsystem is connected to the energy system and cooling system of the current deployment space; the energy and thermal control management subsystem is used to connect the liquid inlet and outlet of the liquid cooling plate on each blade server in each subsystem to the cooling system of the current deployment space; the energy and thermal control management subsystem also includes an external air cooling device to perform thermal management of each subsystem.

[0086] Specifically, the energy and thermal control subsystem ensures the stable operation of the Space Super Intelligent Fusion Computing System and is crucial for its continued operation. Space infrastructure (currently deployed in space), such as space stations or satellites, already has energy systems. For example, solar panels harvest solar energy and batteries store energy. This provides continuous power to the Space Super Intelligent Fusion Computing System, prioritizing critical components. The Space Super Intelligent Fusion Computing System utilizes multiple power supply channels to ensure reliable energy. Regarding thermal control, the space infrastructure (currently deployed in space) employs a multi-layered closed-loop thermal control system. Radiative heat dissipation is utilized to rapidly dissipate heat generated by the equipment into space. A two-phase capillary pump heat pipe array is integrated into the baseplate of each module to ensure uniform heating. However, due to the unique characteristics of the Space Super Intelligent Fusion Computing System, which exhibits localized high-temperature conditions, further thermal control measures are required for high-heat components such as chips. At the chip level, for monolithic build methods, the computing system is integrated into the cooling system (liquid cooling circuit) provided by the deployment space. The liquid cooling plate on each blade server provides liquid inlets and outlets, interconnected with the cooling system provided by the deployment space. Locally high-heat components are covered with liquid cooling plates and memory clips. For uniform and foundation-based build methods, a directional, highly thermally conductive graphene composite film (with an in-plane thermal conductivity greater than 1500W / (m·K)) is applied to the chip surface. Through optimized heat conduction path design, a three-dimensional heat flow channel is established, transferring local hotspots with a heat flux density greater than 50W / cm² to a secondary heat sink system. Heat diffusion is achieved through the isothermal properties of the heat pipe array.

[0087] Building on the above examples, the energy and thermal management subsystem can also control computing power consumption, specifically for:

[0088] According to the moon shadow status, the preset minimum main frequency and the preset maximum main frequency are determined;

[0089] According to the target operation, the estimated number of computing nodes is determined, and the number of retained nodes is determined based on the estimated number of computing nodes and the idle node retention threshold; the switch state of each computing node in the computing subsystem is controlled according to the retained node amount; for each computing node with a switch state of on, if the computing node does not receive an operation signal notification, the computing node is controlled to operate at a preset minimum main frequency; if the computing node receives an operation signal notification, the computing node is controlled to operate at a preset maximum main frequency.

[0090] Compute nodes correspond to compute blade servers. A lunar shadow is formed when the moon blocks some sunlight when it is between the sun and the earth, casting a shadow over certain areas of the earth. This shadow is called a lunar shadow, and the lunar shadow status indicates whether a lunar shadow exists. The idle node retention threshold is the number of nodes reserved to ensure normal system operation. The estimated number of compute nodes is the number of compute nodes expected to be used to execute the target job. The retained node count is the sum of the estimated number of compute nodes and the idle node retention threshold, representing the number of available compute nodes reserved by the control system. Switch states include on and off. Job signal notification is a notification to execute the target job. The preset minimum main frequency and preset maximum main frequency are the preset minimum and maximum main frequencies that the compute node can operate at.

[0091] Specifically, the state of the moon shadow is observed, and the corresponding preset minimum main frequency and preset maximum main frequency are determined according to the different states of the moon shadow. The target job is analyzed to predict the number of computing nodes that need to be used, which is the estimated number of computing nodes. The sum of the estimated number of computing nodes and the idle node retention threshold is determined as the retained node quantity. In the control computing subsystem, each computing node of the retained node quantity is selected to be turned on, that is, the switch state is turned on, and the switch state of other computing nodes is turned off, so as to minimize the energy consumption generated by turning on redundant computing nodes. For each computing node with a switch state turned on, if the computing node does not receive an operation signal notification, in order to reduce energy consumption, the computing node can be controlled to operate at the preset minimum main frequency. If the computing node receives an operation signal notification, the computing node can be controlled to operate at the preset maximum main frequency to execute the target job.

[0092] For example, in terms of computing power control, for computing nodes, or computing blade servers, the computing load can be estimated based on the target job (estimated computing node quantity). A threshold for retaining idle nodes is set, and the number of retained nodes is determined. Computing nodes exceeding the reserved node quantity are all shut down (the open state is turned off). For processors not participating in computing and computing nodes in an idle state, a preset minimum main frequency is set to maintain basic operation. After receiving a job signal notification, the computing node is set to a preset maximum main frequency to ensure job computing, thereby saving energy. Because the heat dissipation capacity per unit area in current deployment space is still limited, and excessively high temperatures can lead to reduced chip operating efficiency, a power consumption-performance model is established to ensure limited performance degradation while controlling power consumption. Before the moon shadow state occurs, power consumption requirements are reduced, and after the moon shadow state occurs, the maximum load is restored.

[0093] Solar wings collect solar energy, and batteries are installed to maintain power during periods of lunar shadow. A multi-layered thermal control loop is employed, utilizing high-thermal conductivity interface materials, heat pipe arrays, loop heat pipes, and radiative cooling to maintain efficient computing power for high-density servers in multiple spatial modules. Detailed chip energy conservation is meticulously controlled, and the system automatically identifies and predicts periods of lunar shadow, reducing loads in advance and maintaining a constant temperature even in low ambient temperatures.

[0094] Based on the above examples, the Space Super Intelligent Fusion Computing System also includes:

[0095] System software subsystem, used to store and start the system software required by each subsystem;

[0096] The application software subsystem is used to store and integrate application software corresponding to various technical fields.

[0097] Specifically, the system software subsystem includes the host operating system, resource scheduling tools, compilation environment, file system software, and parallel environment, supporting high-speed and efficient application execution. The application software subsystem (packaging, pre-installed software sources, and Blu-ray cold storage) integrates computing needs across multiple domains, leveraging the unique advantages of the space environment to provide critical support for each field. In intelligent training and inference, the space environment is leveraged to achieve efficient model training, assisting with autonomous navigation and target recognition for spacecraft, and improving the intelligence of space missions. In astronomy and cosmology, the system rapidly processes massive amounts of data captured by astronomical equipment, helping scientists explore the mysteries of the universe and advancing research on cosmic evolution and dark matter. In Earth observation and environmental monitoring, real-time analysis of satellite data enables rapid early warning of natural disasters and climate change, providing data support for ecological protection and disaster response. Space science experiments and simulations simulate space environments to provide data and theoretical support for actual space experiments, reducing experimental costs and risks. Low-altitude economic space management and aircraft scheduling leverage space supercomputing to achieve precise scheduling and airspace management, promoting the orderly development of the low-altitude economy. Providing a reference for national defense and military industry, fields such as high-energy physics and particle research, space mission planning and autonomous navigation, biology and life sciences, and deep space exploration can all benefit from the powerful computing capabilities of space supercomputing, driving in-depth research in various fields. The diverse demands of the application layer are the driving force behind the development of the entire space supercomputing technology architecture, driving continuous optimization and innovation at other levels.

[0098] A distributed system software technology is proposed to support the operation of space super-intelligent fusion computing power systems, which can support space computing power systems with different hardware composition modes. The software includes a distributed scheduling center, a multi-level storage media management and control system, etc., which can support cross-space module computing resource scheduling and data sharing, and can be expanded to multi-star computing power resource scheduling and data sharing according to needs.

[0099] Unlike previous space computing systems that focused on small computing power, single satellites, and single computing architectures, the technology described in this embodiment proposes new collaborative service technologies for space-based and ground-based computing systems. This technology leverages federated scheduling for multi-space-based computing systems, along with collaborative scheduling and management of computing resources across different architectures, such as ultra-high-precision computing and intelligent computing. This technology, combined with parallel and distributed computing frameworks, manages and schedules the resources of the integrated space-ground super-intelligent fusion computing system at different granularities. While supporting traditional applications with weak computing requirements, it enables efficient processing of complex computing tasks on the integrated space-ground super-intelligent fusion computing system, breaking through the limitations of traditional space computing systems, such as the insufficient integration capabilities of small computing power, single satellites, and multiple satellites, as well as the single type of computing resources.

[0100] The Space Super Intelligent Fusion Computing System aims to build a comprehensive space computing system that integrates powerful computing capabilities, efficient storage, stable communications, reliable energy supply and heat dissipation, and a solid hardware foundation. By integrating innovative technologies across various subsystems, it aims to achieve a leap in space computing capabilities, enabling collaborative services with ground-based supercomputing to meet the needs of complex computing tasks across multiple fields.

[0101] Operation mode can refer to Figure 2 As shown, depending on actual needs, multiple Space Super Intelligent Fusion Computing Systems can be deployed in low Earth orbits, heliosynchronous orbits, and even deep space. A certain number of Space Super Intelligent Fusion Computing Systems in the same orbit form a computing system cluster, providing unified computing services. These Space Super Intelligent Fusion Computing Systems can exchange data and tasks with super, intelligent, and data centers (supercomputing centers, intelligent computing centers, and data centers) built in cities, oceans, and even deserts on the ground, collaboratively providing computing services to meet the needs of the Internet of Vehicles, artificial intelligence, and low-altitude economic space management. Furthermore, they can leverage the advantages of space communications to bring services to the ground via mobile phones. Space Super Intelligent Fusion Computing Systems in low Earth orbit can also directly observe and process remote sensing data, avoiding the inefficiencies associated with remote data transmission. Since the ground coverage rate of satellites in low-Earth orbit is low, dedicated communication satellites are deployed in high-Earth orbit to build a GEO (high-Earth orbit satellite) communication network. This is used for fast data communication between low-Earth orbit computing systems in different regions, and also for data transmission relay with ground communications. In addition, high-Earth orbit satellites can also collect deep space data, which can be transmitted back to low-Earth orbit computing system satellites for calculation and processing.

[0102] The present invention has the following technical effects:

[0103] The computing subsystem determines the computing resources in the current deployment space, receives the target job, and processes the target job based on the computing resources when the computing resources meet the computing requirements of the target job. The storage subsystem determines the read-write characteristics corresponding to the target job when executing the target job, and determines the read-write path based on the read-write characteristics. The data corresponding to the target job is stored based on the read-write path. The communication subsystem establishes a first communication channel between each subsystem and a second communication channel between the space super-intelligent fusion computing power system and the external system. The global management subsystem monitors the operating status of each subsystem, and when the operating status of at least one subsystem is a fault state, the rule engine determines the task instructions corresponding to the fault state, and executes each task instruction according to the intelligent agent corresponding to each task instruction to change the operating status to a normal state, realizing the integration of powerful computing power, efficient storage, stable communication, reliable energy supply and heat dissipation, and a solid hardware foundation in space, effectively improving the performance of space intelligent computing.

[0104] It should be noted that the terms used in the present invention are only for describing specific embodiments and are not intended to limit the scope of this application. As shown in the present specification, unless the context clearly indicates an exception, the words "one", "an", "a kind of" and / or "the" do not specifically refer to the singular and may also include the plural. The terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method or device comprising a series of elements includes not only those elements, but also other elements that are not explicitly listed, or also include elements inherent to such process, method or device. In the absence of further restrictions, the elements defined by the sentence "comprise a..." do not exclude the presence of other identical elements in the process, method or device comprising the elements.

[0105] It should also be noted that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside", etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, they cannot be understood as limitations on the present invention. Unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", etc. should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be a communication between the internal parts of two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0106] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention.

Claims

1. A space super-intelligent fusion computing system, characterized by: include: A computing subsystem, configured to determine computing resources within a current deployment space, receive a target job, and process the target job based on the computing resources if the computing resources meet computing requirements of the target job; a storage subsystem configured to, when executing the target job, determine read / write characteristics corresponding to the target job, determine a read / write path based on the read / write characteristics, and store data corresponding to the target job based on the read / write path; wherein the read / write characteristics are determined based on the read / write throughput, read / write burstiness, and source data operation volume of the target job; A communication subsystem, configured to establish a first communication channel between the subsystems and a second communication channel between the Space Super Intelligent Fusion Computing System and an external system; The global management subsystem is configured to monitor the operating status of each subsystem and, if at least one subsystem is in a faulty operating state, determine, based on a rule engine, the task instructions corresponding to the faulty state and execute the task instructions according to the agents corresponding to the task instructions to restore the operating state to a normal state. The space super-intelligent fusion computing system also includes an energy and thermal control management subsystem for: Determine the idle node retention threshold based on the moon shadow status; According to the target job, the estimated number of computing nodes is determined, and according to the estimated number of computing nodes and the idle node retention threshold, the retained node number is determined; according to the retained node number, the switch state of each computing node in the computing subsystem is controlled; for each computing node whose switch state is turned on, if the computing node does not receive a job signal notification, the computing node is controlled to operate at a preset minimum main frequency; if the computing node receives a job signal notification, the computing node is controlled to operate at a preset maximum main frequency; wherein, the computing node corresponds to a computing blade server.

2. The system according to claim 1, wherein: The currently deployed space has a system rack, and the system rack evenly distributes multiple space modules, and each subsystem is placed on at least one space module; the first number of space modules corresponding to the computing subsystem is determined according to the preset computing power requirement, the computing amount corresponding to a single computing blade server, and the maximum number of computing blade servers corresponding to each space module; the second number of space modules corresponding to the storage subsystem is determined according to the preset storage requirement, the storage capacity of the storage blade servers corresponding to each storage medium, and the maximum number of storage blade servers corresponding to each space module; the third number of space modules corresponding to the communication subsystem is determined according to the preset computing power requirement, the computing amount corresponding to a single computing blade server, the preset storage requirement, the storage capacity of the storage blade servers corresponding to each storage medium, and the maximum number of network blade servers corresponding to each space module; the fourth number of space modules corresponding to the global management subsystem is at least two; the sum of the first number, the second number, the third number, and the fourth number is less than or equal to the total number of space modules on the system rack.

3. The system according to claim 1, wherein: The second communication channel includes an inter-satellite communication channel and a satellite-to-ground communication channel. The communication subsystem is further configured to: Transmitting data corresponding to the target job based on an encrypted time-sensitive network; The encrypted time-sensitive network extends encryption information in a network identification number field in a data frame of an Ethernet link layer, and adds a security tag field and an integrity check value field to the data frame.

4. The system according to claim 1, wherein: The computing subsystem is further configured to: if the computing resources do not meet the computing requirements of the target job, send the computing resources and the computing requirements to the communication subsystem; The communication subsystem is further configured to: receive the computing resources and the computing requirements, and send the computing resources and the computing requirements to a ground scheduling server, so that the ground scheduling server determines a space cooperation system based on the computing resources and the computing requirements; Receive the space collaboration system corresponding to the computing demand fed back by the ground scheduling server, establish a second communication channel with the space collaboration system, and collaboratively process the target job based on the second communication channel, the space super-intelligent fusion computing power system in the current deployment space, and the space collaboration systems in other deployment spaces.

5. The system according to claim 1, wherein: The storage subsystem includes: an acceleration storage layer, a persistent storage layer, and a cold data storage layer; the number of storage blade servers in the persistent storage layer is less than the number of storage blade servers in the acceleration storage layer; The read / write path includes a first path and a second path; the acceleration storage layer and the persistent storage layer serve together as the first path, and the persistent storage layer serves as the second path; when the read / write path is the first path, a parallel file system is constructed based on the acceleration storage layer, data is stored based on the parallel file system, and after the target job is completed, the data in the parallel file system is copied to the persistent storage layer, and the parallel file system is destroyed; The cold data storage layer is configured to store the packaged data corresponding to the generated data in the persistent storage layer when the stabilization time of the stored data in the persistent storage layer reaches a first time threshold.

6. The system according to claim 5, characterized in that The storage subsystem further includes: a software image library; Among them, the software image library is used to store the software image corresponding to each application, so that when the application cannot run, the corresponding software image can be pulled from the software image library; when the update completion time of the software image library reaches the preset update period, the software image library obtains a new software image from the ground data server through the communication subsystem to update the software image library.

7. The system according to claim 1, wherein: The energy and thermal control management subsystem is connected to the energy system and cooling system of the current deployment space; the energy and thermal control management subsystem is used to connect the liquid inlet and outlet of the liquid cooling plate on each blade server in each subsystem with the cooling system of the current deployment space; the energy and thermal control management subsystem also includes a peripheral air cooling device to perform thermal management of each subsystem.

8. The system according to claim 1, wherein: The global management subsystem is further used to: Collect monitoring data from each subsystem at regular intervals, analyze the monitoring data based on a rule engine, and determine target processing instructions; Based on the commander agent, the target processing instructions pushed by the rule engine are received, and the target processing instructions are analyzed and disassembled to obtain task instructions, and each task instruction is sent to the corresponding functional agent respectively, so that each functional agent executes the corresponding task instruction.

9. The system according to claim 1, wherein: Also includes: System software subsystem, used to store and start the system software required by each subsystem; The application software subsystem is used to store and integrate application software corresponding to various technical fields.

Images