Embedded edge-end collaborative information processing distributed system and method

Abstract

The invention discloses an embedded edge-end collaborative information processing distributed system and method.The embedded edge-end collaborative information processing distributed system comprises embedded edge equipment, a plurality of embedded end equipment and an inter-equipment network interconnection bus, the embedded edge equipment is used for achieving unified management of the distributed system by operating a cloud platform, and software maintenance upgrading and equipment state operation and maintenance are included; meanwhile, cooperative information processing of the embedded edge equipment and the plurality of embedded end equipment is realized through task unloading and resource scheduling; the plurality of embedded terminal devices are used for sensing, storing, computing and transmitting internal and external information of the distributed system and are data acquisition and decision execution devices at the tail end of the distributed system; the inter-device network interconnection bus is used for realizing information transmission and interconnection between the embedded edge device and the plurality of embedded end devices, and constructing a star-shaped or peer-to-peer communication switching type network topology architecture taking the embedded edge device as the center; the cloud platform is a container arrangement system, and each device adopts container-level virtualization to realize the unified management of the cloud platform on the application software and the running state of the device.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of collaborative computing, in particular to an embedded edge-end collaborative information processing distributed system and method. BACKGROUND

[0002] The current cloud-edge-end collaborative architecture has become the mainstream paradigm for processing distributed information and has shown great value in ground civilian scenarios such as the Internet of Things and smart cities. However, the technical route and application paradigm of existing edge computing and cloud-edge-end collaborative distributed systems are mainly built around ground Internet of Things scenarios. In the entire cloud-edge-end system of the ground Internet of Things, mutual collaboration between nodes can be completed by relying on the control surface, and the end device nodes are usually positioned as weak terminals with extremely limited resources, mainly responsible for data collection and simple execution, with limited computing and information processing capabilities; the edge nodes serve as the carrier of algorithm sinking, mainly responsible for data fusion processing, lightweight computing, and connection with the cloud center; the cloud center serves as a high-performance cluster, mainly responsible for global analysis, model training, and long-term strategy planning; and the ground system communicates through a standard Ethernet / IP protocol communication architecture.

[0003] However, the distributed cloud-edge-end system of embedded equipment has significant differences compared to the ground cloud-edge-end system, and the nodes in the system need to deeply consider the dual-collaboration mode of the control surface and the business surface. The end device nodes themselves have local information processing, real-time decision-making, and task execution capabilities, and are strong terminals. Moreover, the aerospace equipment system widely uses customized, non-standard communication protocols and interfaces for inter-node communication. Therefore, considering the characteristics of the dual-plane collaborative mode of the aerospace equipment cloud-edge-end system, non-standard communication, and strong capabilities of embedded end nodes, there are fundamental differences from the standard ground Internet of Things cloud-edge-end system. These differences make it difficult for the distributed collaborative information processing scheme suitable for ground Internet of Things scenarios to be applicable to the near-field collaborative scenario of the "embedded edge-embedded end" with a collaborative focus on the edge. Therefore, there is an urgent need to design an embedded edge-end collaborative information processing distributed system and method to fill the technical gap between edge devices and embedded end devices in the aerospace field and to extend it to other embedded edge-end collaborative scenarios in order to meet the new demands of distributed intelligence with strong real-time linkage and high autonomous execution and agile response.

[0004] At present, no similar technology or report has been found, and no similar domestic and foreign data has been collected. SUMMARY

[0005] In order to overcome the deficiencies in the prior art, the present application provides an embedded edge-end collaborative information processing distributed system and method to solve the current situation and technical problems of embedded edge-end information that cannot be collaboratively processed in a multi-device distributed scenario.

[0006] In order to achieve the above invention purposes, the technical solutions adopted to solve its technical problems are as follows: The application discloses an embedded edge-end cooperative information processing distributed system, which comprises an embedded edge device, a plurality of embedded end devices and an inter-networking bus. The embedded edge device is used for realizing unified management of the distributed system by running a cloud platform, including software maintenance and upgrade and device state operation and maintenance. The plurality of embedded end devices are used for sensing, storing, computing, processing and transmitting internal and external information of the distributed system, and are data acquisition and decision execution devices at the end of the distributed system. The inter-networking bus is used for realizing information transmission and interconnection between the embedded edge device and the plurality of embedded end devices, so as to build a star-type or exchange-type network topology architecture of peer-to-peer communication with the embedded edge device as the center. The cloud platform is a container orchestration system, and each device adopts container-level virtualization to realize unified management of application software and running state of the cloud platform.

[0007] Further, the embedded edge device comprises an edge device computing module, an edge device communication module, an edge device storage module and a switching network. The edge device computing module is used for realizing localized computing and processing of the distributed system task, and is completed by a processor device by running application software at a certain working frequency. The edge device communication module is used for realizing information transmission with other devices in the distributed system through the inter-networking bus, or realizing communication with devices outside the distributed system based on a physical link format and protocol. The edge device storage module is used for storing data received by the device or data generated by the device. The switching network is used for realizing interconnection of each computing module in the embedded edge device computing unit, and is composed of a switching chip to form a full switching topology. The distributed system task is a task of information processing generated by the edge device itself or a task of information processing received by the communication module from other devices in the interconnected network composed of the edge device and the end device; the processor device is implemented by a single CPU, a single DSP or a processor core of an SoC integrated circuit, and one or more processor devices exist in one edge device computing module, and the processor architecture and instruction set are not limited.

[0008] Further, the edge device computing module is constructed by using virtualization and layering technology to build software programs, including an operating system, middleware, a container, and an image repository, wherein: The middleware includes a container engine and a virtual network driver. The container includes a cloud platform software container and an application software container.

[0009] Further, the switching network uses a standard TCP / IP Ethernet protocol to implement the interconnection of various computing modules in the computing unit, and supports cloud platform standard management.

[0010] Further, the computing unit is constructed by using a plurality of edge device master computing modules and edge device slave computing modules through the switching network, and the computing unit, the communication module and the storage module use a physical communication link to implement information data transmission.

[0011] Further, the embedded end device includes an end device computing module, an end device communication module, an end device storage module, a sensing module, and an interconnected network, wherein: The end device computing module is used to implement localized computing and processing of the distributed system task, which is completed by the processor device by running application software at a certain working frequency; when a plurality of end device computing modules exist, in order to distinguish the cooperative control relationship, the end device computing module with the device overall control function is defined as an end device master computing module; when there is only one end device computing module, it is defaulted as an end device master computing module. The end device communication module is used to implement information transmission with other devices in the distributed system through an interconnection bus between devices, or to implement communication with devices outside the distributed system based on a physical link format and protocol. The end device storage module is used to store data received by the device or data generated by the device. The sensing module is used for the embedded end device to collect states and obtain information through images and detection means. The interconnected network is used to implement the interconnection of various modules in the embedded end device, and the interconnection mode can be switching or master-slave.

[0012] Further, the end device computing module adopts virtualization and layering technology to construct a software program, including an operating system, middleware, a container, and an image repository, wherein: The middleware includes a container engine and a virtual network driver. The container includes a cloud platform software container, a refactoring configuration software container, and an application software container.

[0013] Further, the refactoring configuration software container implements application software refactoring and upgrading of the end device slave computing module based on the image repository, and collects the running state of the end device slave computing module container and uniformly reports to the cloud platform.

[0014] Further, the interconnection network adopts a wired IP network or any bus communication form and is not limited by the cloud platform.

[0015] Further, the end device master computing module realizes centralized management and control of the computing modules in the device through the interconnection network, thereby realizing a unified cloud platform interface of the edge device, and other end device slave computing modules are not set.

[0016] Further, the interconnection bus between devices is interconnected based on a device communication module using a scene self-defined protocol bus mode, each device is converted through local virtual IP, a virtual network link is built to realize the connection of the cloud platform application interface; The virtual network link is used for mutual conversion of the self-defined protocol bus and the standard TCP / IP network protocol, and realizes north-south communication of data in the computing module processor.

[0017] Further, the cloud platform is used for realizing the unified management function of the distributed system composed of the edge device master computing modules, the edge device slave computing modules, and the end device master computing modules.

[0018] Further, the unified management function is used for device container life cycle management and device operation and maintenance, including cluster management, application orchestration and scheduling, resource configuration and optimization, network policy and security management, storage management, and device monitoring and operation and maintenance.

[0019] The application also discloses an embedded edge-end collaborative information processing method, which utilizes the above-mentioned embedded edge-end collaborative information processing distributed system to process information, and includes the following steps: Step 1: The edge device realizes software refactoring and upgrading of the computing modules in the edge device and the computing modules in the end device by pushing the image repository, thereby maintaining the state of each processing application software in the distributed system according to the application scene requirements. Step 2: The cloud platform enables unified management of all computing modules in edge devices and end devices, orchestrates and controls the containers of computing modules in edge devices and the containers of computing modules in end devices, and summarizes and reports the operating status of each container and the telemetry data of the device itself. Step 3: The computing module within the terminal device, based on the constraints of computing, communication, and storage resources within the distributed system, adopts heuristic or reinforcement learning methods to make information processing task offloading decisions and network resource scheduling according to the optimization goals of user service guarantee requirements, thereby achieving effective collaboration between edge devices within the distributed system.

[0020] By employing the above technical solutions, this invention has the following advantages and positive effects compared with the prior art: This invention discloses an embedded edge-device collaborative information processing distributed system and method, which can be used to achieve unified management of distributed systems at the control plane, including software maintenance and upgrades and device status operation and maintenance; at the same time, it can achieve collaborative information processing between edge devices and embedded devices at the business plane. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 This is a schematic diagram of the composition of an embedded edge-to-edge collaborative information processing distributed system according to the present invention; Figure 2 This is a schematic diagram of software layering for an edge-to-edge collaborative information processing method according to the present invention; Figure 3 This is a schematic diagram illustrating an application scenario of the edge-to-edge collaborative information processing system and method of the present invention. Detailed Implementation

[0023] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Example 1 like Figures 1-3 As shown, this embodiment discloses an embedded edge-device collaborative information processing distributed system, including an embedded edge device, several embedded terminal devices, and an inter-device network interconnection bus, wherein: The embedded edge device is used to achieve unified management of the distributed system through a cloud platform, including software maintenance and upgrades, and device status operation and maintenance. Simultaneously, it enables collaborative information processing between the embedded edge device and several embedded terminal devices through task offloading and resource scheduling. The cloud platform is a container orchestration system, and each device uses container-level virtualization to achieve unified management of its application software and operational status by the cloud platform.

[0025] In this embodiment, the cloud platform is used to implement unified management of the distributed system composed of the main computing modules of each edge device, the slave computing modules of each edge device, and the main computing modules of each end device. This unified management function is used for device container lifecycle management and device operation and maintenance, including cluster management, application orchestration and scheduling, resource configuration and optimization, network policy and security control, storage management, device monitoring and operation and maintenance, etc.

[0026] Several of the aforementioned embedded terminal devices are used for sensing, storing, computing, processing, and transmitting information within and outside the distributed system, and are data acquisition and decision execution devices at the end of the distributed system; The device interconnection bus is used to realize information transmission and interconnection between the embedded edge device and several embedded end devices, thereby constructing a star or peer-to-peer switching network topology centered on the embedded edge device.

[0027] See Figure 1 The embedded edge device includes an edge device computing module, an edge device communication module, an edge device storage module, and a switching network, wherein: The edge device computing module is used to implement localized computing processing of distributed system tasks. It is completed by the processor device running application software at a certain operating frequency, and its computing power is typically characterized by computing resources. When multiple edge device computing modules exist, to distinguish collaborative management relationships, the computing module with overall device management functions needs to be defined as the edge device master computing module; when there is only one edge device computing module, it is the default edge device master computing module. The distributed system tasks are information processing tasks generated by the device itself or received from other devices by the communication module within the interconnected network composed of edge devices and end devices. The processor device is implemented using a single-chip CPU, a single-chip DSP, or a processor core in a SoC integrated circuit. An edge device computing module may contain one or more processor devices, and the processor architecture and instruction set are unrestricted.

[0028] The edge device communication module is used for realizing information transmission with other devices inside the distributed system through an inter-network bus or realizing communication with devices outside the distributed system based on a physical link format and protocol; and is usually characterized by a communication resource representing a device channel transmission capability. The communication resource includes bandwidth, power, capacity, error rate, etc. of a communication link.

[0029] The edge device storage module is used for storing data received by the device or data generated by the device, and is usually characterized by a storage resource representing a size of a data storage capacity.

[0030] The switching network is used for realizing interconnection of various computing modules in the embedded edge device computing unit, and is composed of a switching chip to form a full switching topology; and can be used as a device independent module or integrated into other modules, such as a main computing module, as a functional component. The switching network uses a standard TCP / IP Ethernet protocol to realize interconnection of various computing modules inside the computing unit, and supports cloud platform standard management.

[0031] In this embodiment, a computing unit is constructed based on a plurality of edge device main computing modules and edge device slave computing modules through a switching network. The computing unit, the communication module and the storage module realize information data transmission through a physical communication link.

[0032] Further referring to Figure 1 , the embedded end device includes an end device computing module, an end device communication module, an end device storage module, a sensing module and an inter-network, wherein: The end device computing module is used for realizing localized computing processing of a distributed system task, and is completed by a processor device running application software at a certain working frequency; and is usually characterized by a computing resource representing a size of a computing capability. When there are a plurality of end device computing modules, an end device computing module with a device overall management function is defined as an end device main computing module to distinguish a cooperative management relationship. When there is only one end device computing module, it is defaulted as an end device main computing module. The end device main computing module realizes centralized management of various computing modules in the device through the inter-network, thereby realizing a unified cloud platform interface of the edge device, and other end device slave computing modules are not provided.

[0033] The end device communication module is used for realizing information transmission with other devices inside the distributed system through an inter-network bus or realizing communication with devices outside the distributed system based on a physical link format and protocol; and is usually characterized by a communication resource representing a device channel transmission capability. The communication resource includes bandwidth, power, capacity, error rate, etc. of a communication link.

[0034] The end device storage module is configured to store data received by the device or data generated by the device, and is usually represented by a storage resource in terms of the size of the data storage capacity thereof; The perception module is configured to collect states and acquire information by means of images and detection for the embedded end device. The interconnection network is configured to realize interconnection between the various modules of the embedded end device, and the interconnection mode can be switching or master-slave. The interconnection network adopts a wired IP network or any bus communication form and is not limited by the cloud platform.

[0035] Referring to Figure 1 The interconnection bus between the devices in the cloud platform is interconnected based on a scene self-defined protocol bus mode of the device communication module, and each device is converted into a virtual IP to build a virtual network link to realize the opening of the application interface with the cloud platform. The virtual network link is configured to convert the self-defined protocol bus and the standard TCP / IP network protocol to realize the north-south communication in the processor of the computing module.

[0036] Further, on the basis of the above-mentioned distributed system, referring to Figure 2 The embodiment also provides a software layered diagram of edge-end collaborative information processing, and the software layered architecture mainly includes: (1) Embedded edge device software program: the computing module (including the master computing module and the slave computing module) adopts virtualization and layering technology to build a software program, including an operating system, middleware, a container, and an image repository. The middleware includes a container engine, a virtual network driver, and other necessary drivers for the operation of the device; and the container includes a cloud platform software container and an application software container.

[0037] (2) Embedded end device software program: the computing module of the embedded end device adopts virtualization and layering technology to build a software program, including an operating system, middleware, a container, and an image repository. The middleware includes a container engine, a virtual network driver, and other necessary drivers for the operation of the device; and the container includes a cloud platform software container, a reconfiguration software container, and an application software container.

[0038] In the embodiment, the reconfiguration software container realizes the reconfiguration and upgrade of the application software of the slave computing module of the end device based on the image repository, collects the running state of the slave computing module container of the end device, and uniformly reports the cloud platform.

[0039] Further, referring to Figure 2 The embodiment also provides an embedded edge-end collaborative information processing method. In the method, the following are mainly realized: ① software reconfiguration and upgrade between the edge device and the end device, ② edge-end cluster management, and ③ collaborative information processing.

[0040] Further, referring to Figure 3 The embodiment also provides an application scenario diagram of the embedded edge-end collaborative information processing distributed system and method. The embedded edge device in the embedded edge-end collaborative information processing distributed system is mounted on an edge node in the application scenario diagram, and the embedded end device in the embedded edge-end collaborative information processing distributed system is mounted on an end node in the application scenario diagram. The collaboration between the edge node and the end node can include but is not limited to the collaboration between a vehicle edge node and a vehicle end node, the collaboration between a satellite edge node and a vehicle end node, and the collaboration between a satellite edge node and a flight end node. Through the edge-end information processing distributed system and method, the following functions can be realized: ① software reconstruction and upgrading, ② edge-end cluster management, and ③ collaborative information processing.

[0041] Embodiment two The embodiment discloses an embedded edge-end collaborative information processing method, which utilizes the above-mentioned embedded edge-end collaborative information processing distributed system to process information, and includes the following steps. Step 1: The edge device realizes software reconstruction and upgrading of the computing modules in the edge device and the computing modules (excluding slave computing modules) in the end device by pushing a mirror repository, so as to maintain the states of various processing application software in the distributed system according to application scenario requirements. Step 2: The cloud platform realizes unified management of all computing modules in the edge device and the end device, and arranges and controls the containers of the computing modules in the edge device and the containers of the computing modules in the end device. The running states of each container and the telemetry data of the device itself are collected and reported. Step 3: The computing modules in the end device make information processing task offloading decisions and network resource scheduling according to the optimization target of user service guarantee requirements by using a heuristic or reinforcement learning method based on the constraints of computing, communication and storage resources in the distributed system, so as to realize effective collaboration of the edge-end devices in the distributed system. The above-mentioned is only a preferred specific implementation manner of the present application, but the protection scope of the present application is not limited to this. Any change or replacement that can be easily thought of by those skilled in the art within the technical range disclosed by the present application should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. An embedded edge collaborative information processing distributed system, characterized in that, The embedded edge device, the plurality of embedded end devices, and an inter-device network interconnection bus are included, wherein: The embedded edge device is configured to implement unified management and control of a distributed system by running a cloud platform, including software maintenance and upgrade and device state operation and maintenance; meanwhile, the embedded edge device and the plurality of embedded end devices are configured to implement collaborative information processing through task offloading and resource scheduling. The plurality of embedded end devices are configured to sense, store, calculate, process, and transmit internal and external information of the distributed system, and are data acquisition and decision execution devices at the end of the distributed system. The inter-device network interconnection bus is configured to implement information transmission and interconnection between the embedded edge device and the plurality of embedded end devices, thereby constructing a star-type or exchange-type network topology architecture of peer-to-peer communication centered on the embedded edge device. The cloud platform is a container orchestration system, and each device adopts container-level virtualization to implement unified management and control of the cloud platform on application software and running states thereof.

2. The distributed system of claim 1, wherein, The embedded edge device includes an edge device computing module, an edge device communication module, an edge device storage module, and a switching network, wherein: The edge device computing module is configured to implement localized calculation and processing of distributed system tasks, which is completed by a processor device running application software at a certain working frequency; when there are a plurality of edge device computing modules, a computing module with device overall control function is defined as an edge device master computing module to distinguish the collaborative control relationship; when there is only one edge device computing module, it is by default an edge device master computing module. The edge device communication module is configured to implement information transmission with other devices inside the distributed system through the inter-device network interconnection bus, or to implement communication with devices outside the distributed system based on physical link format and protocol. The edge device storage module is configured to store data received by the device or data generated by the device. The switching network is configured to implement interconnection of each computing module in the embedded edge device computing unit, and is composed of a switching chip to form a full switching topology. The distributed system task is an information processing task generated by a device itself or received by a communication module from other devices in an interconnected network composed of edge devices and end devices; the processor device is implemented by a processor core in a single CPU, a single DSP, or an SoC integrated circuit; there is one or more processor devices in one edge device computing module, and the processor architecture and instruction set are not limited.

3. The distributed system of claim 2, wherein, The edge device computing module is constructed by using virtualization and layering technology to build software programs, including an operating system, middleware, a container, and an image repository, wherein: The middleware includes a container engine and a virtual network driver. The container includes a cloud platform software container and an application software container.

4. The distributed system of claim 2, wherein, The switching network implements interconnection of each computing module in the computing unit by using a standard TCP / IP Ethernet protocol, and supports cloud platform standard management and control.

5. The distributed system of claim 2, wherein, A computing unit is constructed by a plurality of edge device master computing modules and edge device slave computing modules through a switching network, and the computing unit, the communication module, and the storage module implement information data transmission by using a physical communication link.

6. The distributed system of claim 2, wherein, The embedded end device comprises an end device computing module, an end device communication module, an end device storage module, a sensing module and an interconnection network, wherein: The end device computing module is used for realizing localized computing processing of distributed system tasks, which is completed by a processor device by running application software at a certain working frequency; when there are multiple end device computing modules, in order to distinguish the cooperative control relationship, the end device computing module with the overall device control function is defined as an end device master computing module; when there is only one end device computing module, it is defaulted as an end device master computing module. The end device communication module is used for realizing information transmission with other devices inside the distributed system through an interconnection bus between devices, or realizing communication with devices outside the distributed system based on physical link format and protocol. The end device storage module is used for storing data received by the device or data generated by itself. The sensing module is used for state collection and information acquisition by the embedded end device through images and detection means. The interconnection network is used for realizing interconnection of each module inside the embedded end device, and the interconnection mode can be switching or master-slave.

7. The distributed system of claim 6, wherein the edge server is further configured to: The end device computing module adopts virtualization and layering technology to construct software programs, including an operating system, middleware, a container and an image repository, wherein: The middleware comprises a container engine and a virtual network driver. The container comprises a cloud platform software container, a reconfiguration software container and an application software container.

8. The distributed system of claim 7, wherein the edge server is further configured to: The reconfiguration software container realizes reconfiguration and upgrade of application software of the end device slave computing module based on the image repository, and collects the running state of the end device slave computing module container and uniformly reports to the cloud platform.

9. The distributed system of claim 6, wherein the edge server is further configured to: The interconnection network adopts a wired IP network or any bus communication form and is not limited by the cloud platform.

10. The distributed system of claim 6, wherein, The end device master computing module realizes centralized control of each computing module in the device through the interconnection network, thereby realizing a unified cloud platform interface of the edge device, and other end device slave computing modules are not set.

11. The distributed system of claim 1, wherein, The interconnection bus between devices adopts a scene self-defined protocol bus mode based on the device communication module, each device is converted through local virtual IP, a virtual network link is built to realize connection with the cloud platform application interface; The virtual network link is used for mutual conversion of the self-defined protocol bus and the standard TCP / IP network protocol, and realizes north-south communication of data in the computing module processor.

12. The distributed system of claim 4, wherein, The cloud platform is used for realizing unified management of the distributed system composed of each edge device master computing module, edge device slave computing module and end device master computing module.

13. The distributed system of claim 8, wherein the edge server is further configured to: The unified management function is used for device container life cycle management and device operation and maintenance, including cluster management, application orchestration and scheduling, resource configuration and optimization, network policy and security control, storage management, device monitoring and operation and maintenance.

14. An embedded edge collaborative information processing method, characterized by, The embedded edge-end cooperative information processing distributed system is used for information processing, comprising the following steps: Step 1: The edge device implements software reconfiguration and upgrade of the computing module in the edge device and the computing module in the terminal device by pushing the image repository, thereby maintaining the state of each processing application software in the distributed system according to the application scenario requirements; Step 2: The cloud platform realizes unified management of all computing modules in the edge device and the terminal device, and performs orchestration and control on the containers of the computing module in the edge device and the containers of the computing module in the terminal device, and the running state of each container and the device itself telemetry data are summarized and reported; Step 3: The computing module in the terminal device makes information processing task offloading decision and network resource scheduling based on the optimization target of user service guarantee demand by using heuristic or reinforcement learning method according to the computing, communication and storage resource constraints in the distributed system, thereby realizing effective coordination of edge and terminal devices in the distributed system.

Images