Trusted interoperation system based on cloud edge optimization intelligent algorithm

Abstract

The invention relates to the technical field of Internet of Things, and particularly discloses a trusted interoperation system based on a cloud edge optimization intelligent algorithm, which comprises a trusted mechanism, a cloud computing platform, an edge computing platform and a cloud edge collaborative task management module, the trusted mechanism is used for generating a secret key and authenticating the accessed terminal equipment; the cloud computing platform is used for storing the encrypted Internet of Things data and responding to the request of the terminal equipment authenticated by the trusted mechanism; the edge computing platform is used for receiving and encrypting the Internet of Things data provided by the Internet of Things equipment platform, and transmitting the encrypted Internet of Things data to the cloud computing platform; the cloud edge collaborative task management module is used for identifying the task category of the Internet of Things equipment to obtain an unloading decision; and performing local calculation on the task according to the unloading decision, or unloading the task to an edge computing platform or an internet of things platform for calculation processing. The system aims to solve the problems that an existing system is low in task scheduling resource utilization rate, and secret leakage is prone to occurring in the data transmission process.

Description

Technical Field

[0001] The present invention relates to the technical field of Internet of Things, and in particular to a trusted interoperable system based on a cloud-edge optimization intelligent algorithm. Background Art

[0002] With the development of IoT technology, it is estimated that there will be more than 10 billion terminals and sensors in the world, and the number of IoT access nodes is showing an explosive growth trend. This phenomenon brings new challenges to the real-time performance and network bandwidth of network transmission, and also puts higher requirements on the storage, analysis and processing of data. The traditional centralized cloud computing model can no longer meet the needs of efficient transmission and processing of massive data in the era of the Internet of Everything. To solve the above problems, edge computing came into being. By extending cloud native capabilities to the edge, training AI models in the cloud, and sinking data processing, business applications and trained AI models to edge nodes for execution, it effectively solves the problems of real-time response, data privacy, and maintenance convenience when the Internet of Things is implemented.

[0003] Especially in the field of industrial Internet of Things, due to the massive number of digital devices connected to the industrial Internet of Things and the large amount of data generated, the demand for computing resources has shown an explosive growth, which has brought severe challenges to the limited computing resources of the terminal devices themselves. The cloud-edge collaborative computing mode that combines edge computing and cloud computing to handle computing-intensive tasks and delay-sensitive tasks generated by industrial Internet of Things devices will better meet the needs of system users. However, when computing tasks are offloaded, the computing, communication, storage and other resources in the cloud-edge collaborative system will change dynamically. How to optimize the system task scheduling, efficiently allocate resources, ensure the reliability of data transmission, and maximize resource utilization efficiency has become an urgent problem to be solved. In addition, how to achieve trust in device entities in an open environment and prevent data privacy leakage is also a key issue that needs to be solved. Summary of the invention

[0004] In view of the shortcomings of the existing technology, the technical problem solved by the present invention is to provide a trusted interoperable system based on a cloud-edge optimization intelligent algorithm to solve the problem that the existing industrial network Internet of Things cloud-edge collaborative system cannot perform task scheduling well and the data transmission process is prone to leakage.

[0005] In order to solve the above problems, the technical solution adopted by the present invention is: a trusted interoperable system based on cloud-edge optimization intelligent algorithm, the system includes a trusted institution, a cloud computing platform, an edge computing platform, and an Internet of Things device platform;

[0006] The trusted institution is used to generate keys and authenticate connected terminal devices; the cloud computing platform is used to store encrypted IoT data and respond to requests from terminal devices authenticated by the trusted institution; it is also used to receive tasks transmitted by the IoT device platform and execute them, and return the execution results to the IoT device platform;

[0007] The edge computing platform includes an encryption module, which is used to receive and encrypt IoT data provided by the IoT device platform and transmit the encrypted IoT data to the cloud computing platform; it is also used to receive and execute tasks transmitted by the IoT device platform and return the execution results to the IoT device platform;

[0008] The IoT device platform includes a cloud-edge collaborative task management module, which is used to identify the task category of the IoT device and obtain the offloading decision through a pre-trained scheduling model; and according to the offloading decision, the task is calculated locally, or offloaded to the edge computing platform or the IoT platform for calculation and processing.

[0009] The beneficial effects of this solution are: by introducing a trusted mechanism and encrypting data during data transmission, the security and integrity of data during transmission can be ensured, which helps prevent data from being stolen or tampered with, and protects user privacy and sensitive information. The cloud-edge collaborative task management module uses a pre-trained scheduling model to intelligently determine the execution location of the task according to the type of task, the amount of task data, and the real-time status of the system, thereby optimizing resource utilization and response time; it can effectively balance the load of edge computing and cloud computing and improve the performance of the overall system. The system can flexibly adjust the execution strategy of tasks according to different application scenarios and requirements. For example, for tasks that require a quick response, you can choose to execute them on the edge node; for tasks with large computational volume and low latency requirements, you can choose to offload them to the cloud center node for execution; so that the system can adapt to various different application requirements. By reasonably allocating tasks to edge nodes and cloud center nodes, you can reduce dependence on cloud computing resources, thereby reducing operating costs. At the same time, since edge nodes are usually closer to users, they can reduce network latency and bandwidth consumption, further reducing communication costs. When cloud connection is unavailable, the system can still execute tasks on edge nodes or local terminal devices, ensuring business continuity and reliability.

[0010] Furthermore, the cloud-edge collaborative task management module is also used to assign different category priorities S to different task categories, and assign dynamic priorities P to tasks of the same type, and schedule tasks according to the category priorities S and dynamic priorities P. The dynamic priorities are based on the task waiting time t iw and the estimated task calculation time t ic Determine, the dynamic priority calculation formula is as follows P = (tjc +t ic +t iw ) / t ic ; said t jc Estimate the calculation time for the current scheduled task.

[0011] When a new task arrives in the system queue, the task will use the remaining computing time of the task currently being executed by the system as the waiting time and the estimated computing time of the task itself to calculate the priority P, which is defined as the response ratio. When the system completes the task and reschedules the task, it selects the task with the highest priority P value in the queue for execution. At the same time, the priority P is recalculated based on the computing time of the selected task as the increased waiting time of other tasks in the queue, and this is repeated until the task is completed. The ratio of process waiting time to service time is taken into account, so that the CPU time slice can be allocated more fairly, and CPU resources can be effectively utilized to avoid starvation. Compared with the short task priority scheduling algorithm, this scheme is fairer and does not completely ignore the needs of long tasks.

[0012] Furthermore, the edge computing platform also includes a data cleaning module, which is used to clean IoT data and remove redundancy. By identifying and deleting duplicate or unnecessary data, the simplicity and accuracy of the data set are ensured, which helps to improve data encryption performance.

[0013] Further, the data encryption module uses any one of the Rijndael encryption algorithm, the Blowfish algorithm and the Twofish algorithm to encrypt data. The Rijndael algorithm supports variable data block length and key length, providing great flexibility. This flexibility enables it to adapt to different security requirements and application environments, and the Blowfish algorithm allows encryption using keys of different lengths, up to 448 bits, providing users with flexible options. In addition, it also supports multiple encryption modes, such as ECB, CBC, etc., further enhancing its application flexibility. The Twofish algorithm also supports multiple key length options (128, 192, 256 bits), and can be expanded to larger block lengths and more rounds to meet higher security requirements. This design makes the Twofish algorithm more flexible and changeable when facing different security challenges.

[0014] Furthermore, the scheduling model is a deep neural network model including an input layer, an LSTM layer, a fully connected layer, an advantage and value layer, and an output layer. The scheduling model input features include task size, computing queue status, transmission queue status, edge computing platform status, and historical load information of the edge computing platform, cloud computing platform status, and historical load information of the cloud computing platform. By using the LSTM-based neural network model to calculate the relevant time period from the generation of system tasks to the end of calculation, the load changes of edge nodes can be predicted more quickly, and reasonable and effective unloading decisions can be made in a timely manner to meet the differentiated needs of user services.

[0015] Furthermore, the task categories include command control, information collection, and mobile application, and the command control category has a higher priority than the information collection category, which is higher than the mobile application category. The control command category includes network scheduling commands, resource allocation commands, emergency commands, management control commands, and other command set data, which require low bandwidth, low latency, and high reliability communication support. Therefore, the command control task category has the highest priority. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic diagram of the system module of the present invention. DETAILED DESCRIPTION

[0017] The following is further described in detail through specific implementation methods:

[0018] Basically as attached Figure 1 As shown: A trusted interoperable system based on cloud-edge optimization intelligent algorithm, the system includes a trusted institution, a cloud computing platform, an edge computing platform, and an Internet of Things device platform;

[0019] The trusted institution is used to generate keys and authenticate connected terminal devices, such as user terminals, newly connected IoT terminals, and newly connected edge servers; the cloud computing platform is used to store encrypted IoT data and respond to requests from terminal devices authenticated by the trusted institution, such as data query requests initiated by user terminals; it is also used to receive tasks transmitted by the IoT device platform and execute them, and return the execution results to the IoT device platform;

[0020] The edge computing platform includes an encryption module, which is used to receive and encrypt the IoT data provided by the IoT device platform and transmit the encrypted IoT data to the cloud computing platform; it is also used to receive and execute tasks transmitted by the IoT device platform and return the execution results to the IoT device platform; the data encryption module uses any one of the Rijndael encryption algorithm, Blowfish algorithm and Twofish algorithm to encrypt data.

[0021] The IoT device platform includes a cloud-edge collaborative task management module, which is used to identify the task category of the IoT device and obtain the offloading decision through the pre-trained scheduling model; and the task is calculated locally or offloaded to the edge computing platform or IoT platform for calculation and processing according to the offloading decision. The scheduling model trained autonomously by the neural network determines whether each task needs to be offloaded, and whether it is offloaded to the edge node or the cloud center node for execution; if the offloading is not performed, the task is placed in the local terminal device execution queue and waits for the local terminal to execute.

[0022] The scheduling model is a deep neural network model including input layer, LSTM layer, fully connected layer, advantage and value layer and output layer. The scheduling model input features include task size, computing queue status, transmission queue status, edge computing platform status and historical load information of edge computing platform, cloud computing platform status and historical load information of cloud computing platform. By using the LSTM-based neural network model to calculate the relevant time period from the generation of system tasks to the end of calculation, the load changes of edge nodes can be predicted more quickly, and reasonable and effective unloading decisions can be made in time to meet the differentiated needs of user services.

[0023] The cloud-edge collaborative task management module is also used to assign different category priorities S to different task categories, assign dynamic priorities P to tasks of the same type, and schedule tasks based on the category priorities S and dynamic priorities P. The dynamic priorities are based on the task waiting time t iw and the estimated task calculation time t ic Determine, the dynamic priority calculation formula is as follows P = (t jc +t ic +t iw ) / t ic ; said t jc Estimate the calculation time for the currently scheduled task. When a new task arrives in the system queue, the task will use the remaining calculation time of the task currently being executed by the system as the waiting time, and the estimated calculation time of itself to calculate the priority P, which is defined as the response ratio. When the system completes the task and reschedules the task, the task with the highest priority P value in the queue is selected for execution. At the same time, the priority P is recalculated based on the calculation time of the selected task as the increased waiting time of other tasks in the queue, and this is continuously changed until the task is processed. The ratio of process waiting time to service time is taken into account, so that the CPU time slice can be allocated more fairly, and CPU resources can be effectively utilized to avoid starvation. Compared with the short task priority scheduling algorithm, this scheme is fairer and will not completely ignore the needs of long tasks.

[0024] Task categories include command control, information collection and mobile application. The priority of the command control category is higher than that of the information collection category, which is higher than that of the mobile application category. Mobile application categories include streaming data such as telephone and television, collaborative office, and video surveillance, which require high bandwidth and communication support with low requirements for latency and reliability; information collection categories include data collected in real time and needing to be processed, such as autonomous driving road condition information, drone inspections, and medical information, which require high bandwidth and communication support with relatively high requirements for latency and reliability; control command categories include instruction set data such as network scheduling commands, resource allocation commands, emergency commands, and management control commands, which require communication support with low bandwidth, low latency, and high reliability.

[0025] These tasks are queued, transmitted, and calculated together in the cloud-edge system, but the three types of tasks coexisting in the system have different sensitivities to delay, that is, different maximum delay constraints and priorities. For command control tasks, since they have high requirements for delay and reliability, and offloading to the cloud center for processing will result in a large transmission delay, the command control task category has the highest priority, and this type of task is generally calculated on local devices and edge nodes. For information collection tasks, the amount of data is relatively large, but it is urgent to perform calculations and obtain control and management instructions. Therefore, this type of task is mainly concentrated on edge computing; for mobile application tasks, the amount of data is the largest, but the real-time requirements are not high, so this type of task is mainly offloaded to the cloud computing center for execution.

[0026] The above is only an embodiment of the present invention, and the common knowledge such as the known specific structure and characteristics in the scheme is not described in detail here. It should be pointed out that for those skilled in the art, several deformations and improvements can be made without departing from the structure of the present invention, which should also be regarded as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A trusted interoperable system based on cloud-edge optimization intelligent algorithm, Features: The system includes a trusted institution, a cloud computing platform, an edge computing platform, and an Internet of Things device platform; The trusted institution is used to generate keys and authenticate connected terminal devices; the cloud computing platform is used to store encrypted IoT data and respond to requests from terminal devices authenticated by the trusted institution; it is also used to receive tasks transmitted by the IoT device platform and execute them, and return the execution results to the IoT device platform; The edge computing platform includes an encryption module, which is used to receive and encrypt IoT data provided by the IoT device platform and transmit the encrypted IoT data to the cloud computing platform; it is also used to receive and execute tasks transmitted by the IoT device platform and return the execution results to the IoT device platform; The IoT device platform includes a cloud-edge collaborative task management module, which is used to identify the task category of the IoT device and obtain the offloading decision through a pre-trained scheduling model; and according to the offloading decision, the task is calculated locally, or offloaded to the edge computing platform or the IoT platform for calculation and processing.

2. A trusted interoperable system based on cloud-edge optimization intelligent algorithm according to claim 1, Features: The cloud-edge collaborative task management module is also used to assign different category priorities S to different task categories, assign dynamic priorities P to tasks of the same category, and schedule tasks based on the category priorities S and dynamic priorities P. The dynamic priorities are based on the task waiting time t iw and the estimated task calculation time t ic Determine, the dynamic priority calculation formula is as follows P = (t jc +t ic +t iw ) / t ic ; said t jc Estimate the calculation time for the current scheduled task.

3. According to claim 1, a trusted interoperable system based on cloud-edge optimization intelligent algorithm, Features: The edge computing platform also includes a data cleaning module, which is used to clean the IoT data and remove redundancy.

4. According to claim 1, a trusted interoperable system based on cloud-edge optimization intelligent algorithm, Features: The data encryption module uses any one of the Rijndael encryption algorithm, the Blowfish algorithm and the Twofish algorithm to encrypt data.

5. According to claim 1, a trusted interoperable system based on cloud-edge optimization intelligent algorithm, Features: The scheduling model is a deep neural network model including an input layer, an LSTM layer, a fully connected layer, an advantage and value layer, and an output layer. The scheduling model input features include task size, computing queue status, transmission queue status, edge computing platform status and historical load information of the edge computing platform, cloud computing platform status and historical load information of the cloud computing platform.

6. A trusted interoperable system based on cloud-edge optimization intelligent algorithm according to claim 1, Features: The task categories include instruction control, information collection and mobile application. The instruction control category has a higher priority than the information collection category, which is higher than the mobile application category.

Images