Data processing method in edge cloud cooperation system and edge cloud cooperation system

Abstract

The invention provides a data processing method in an edge cloud collaboration system and the edge cloud collaboration system. The method comprises the following steps: configuring a dynamic resource influence interface and a static resource influence interface on a node in the edge cloud cooperative system; the nodes in the edge cloud collaborative system utilize the dynamic resource influence interfaces of the nodes to perceive the application load and the resource use condition in real time, and adjust the computing resources, the storage resources and the network resources according to the perceived application load and the resource use condition; the nodes in the side cloud cooperative system utilize own static resources to influence the infrastructure resources of the interface configuration and management system; and the nodes in the edge cloud collaborative system obtain the memory use condition of the application process, and recover and release the memory according to the memory use condition. According to the invention, the resources of the cloud edge collaboration system can be better utilized.

Description

Technical Field

[0001] One or more embodiments of the present invention relate to network communication technology, and in particular, to a data processing method and an edge-cloud collaborative system. Background Art

[0002] As a new computing model, edge-cloud collaboration has become a new research hotspot in academia and industry. Edge computing is a new computing model that performs computing at the edge of the network. The edge-cloud collaboration system is a system that adopts the edge computing model. The downlink data in the edge-cloud collaboration system represents cloud service data, and the uplink data represents the Internet of Everything service data. Among them, the edge of edge computing refers to any computing and network resources between the path from the data source to the cloud service center.

[0003] In the edge-cloud collaborative system, user data no longer needs to be uploaded to the cloud service center. Instead, some data can be quickly processed by edge nodes deployed at the edge of the network, which greatly reduces the pressure on network bandwidth, reduces the pressure on cloud service centers, and significantly reduces the energy consumption of smart devices at the edge of the network. The edge-cloud collaborative system has become an indispensable key link in the industrial Internet and a core supporting technology for realizing the digital transformation of the industry.

[0004] The resources and tasks in the cloud-edge collaborative system are dynamically changing and large in scale. How to maximize the utilization of resources in the cloud-edge collaborative system is a problem that needs to be solved. Summary of the invention

[0005] One or more embodiments of the present invention describe a data processing method and an edge-cloud collaborative system in an edge-cloud collaborative system, which can better utilize the resources of the cloud-edge collaborative system.

[0006] According to a first aspect, a data processing method in an edge-cloud collaborative system is provided, the method comprising:

[0007] Configure dynamic resource impact interfaces and static resource impact interfaces on nodes in the edge-cloud collaborative system;

[0008] The nodes in the edge-cloud collaborative system use their own dynamic resource impact interfaces to perceive application load and resource usage in real time, and adjust computing resources, storage resources, and network resources based on the perceived application load and resource usage;

[0009] Nodes in the edge-cloud collaborative system use their own static resources to influence interface configuration and manage the infrastructure resources of the system;

[0010] The nodes in the edge-cloud collaborative system obtain the memory usage of the application process and recycle and release the memory based on the memory usage.

[0011] The nodes in the edge-cloud collaborative system use their own static resources to influence the interface configuration and manage the infrastructure resources of the system, including:

[0012] In the edge-cloud collaborative system, the storage resources of multiple nodes are integrated to form a unified storage resource pool. When an inference service runs on any of the multiple nodes, the data in the entire storage resource pool is accessed through the static resource impact interface on the node to achieve network storage sharing.

[0013] or,

[0014] In the edge-cloud collaborative system, when a node has insufficient resources, it borrows resources from other nodes through the static resource impact interface to meet its own needs; and when the node has excess resources, it contributes the excess resources to other nodes through the static resource impact interface;

[0015] or,

[0016] Assign a unique SHA-256 hash value to the resources in the edge-cloud collaborative system as the identifier of the resource. After receiving the update package for the resource through the static resource impact interface, calculate the SHA-256 hash value of the update package of the resource, and compare the calculated SHA-256 hash value of the resource update package with the SHA-256 hash value assigned to the resource. If they are the same, the resource will not be updated. If they are different, the resource will be updated.

[0017] The method further comprises:

[0018] The node dynamically increases or decreases the number of service instances based on the real-time load. When the system load increases, new service instances are started to handle additional reasoning tasks; when the load decreases, some service instances are shut down to release resources, thus realizing a serverless reasoning resource architecture.

[0019] and / or,

[0020] Package applications and their dependencies into independent containers to achieve application isolation, portability, and manageability;

[0021] and / or,

[0022] Break complex systems into a series of small, independent services, each of which performs a specific function.

[0023] Through real-time monitoring and analysis, the CPU utilization, memory usage, disk space utilization, and network connection status of the nodes are collected. The future resource requirements are predicted by using deep learning algorithms and the collected CPU utilization, memory usage, disk space utilization, and network connection status of the nodes.

[0024] The nodes in the edge-cloud collaborative system obtain the memory usage of the application process and recycle and release the memory according to the memory usage, including:

[0025] For each memory block, the nodes in the edge-cloud collaborative system perform the following operations: collecting the access frequency to the memory block, the last access time to the memory block, and the life cycle of the memory block;

[0026] Calculate the importance score for each memory block based on the collected access frequency, last access time, and life cycle of each memory block;

[0027] Predict future memory requirements;

[0028] According to the predicted memory demand and the importance score of each memory block, memory blocks are recycled starting from memory blocks with low importance scores, and memory blocks are recycled based on a dynamically adjusted memory recycling strategy.

[0029] The importance score is calculated for each memory block, including:

[0030] Use the following calculation formula to calculate:

[0031] S(b)=w_a\cdot A(b)+w_t\cdot T(b)+w_l\cdot L(b)]

[0032] Wherein: S(b) represents the importance score of memory block b; A(b) represents the access frequency to memory block b; T(b) represents the last access time to memory block b; L(b) represents the life cycle of memory block b, which is the length of time from the creation of memory block b to the current time; w_a, w_t, w_l are the weights of access frequency, last access time and life cycle respectively, and w_a+w_t+w_l=1.

[0033] The memory block recycling based on the dynamically adjusted memory recycling strategy includes:

[0034] Dynamically calculate the memory recycling threshold;

[0035] When the current memory usage is greater than the latest calculated memory recycling threshold, memory block recycling begins;

[0036] The memory recycling threshold is dynamically calculated using the following formula:

[0037] T_{new}=T_{base}+k\cdot(U-U_{target})

[0038] Among them: T_{new} represents the latest recycling threshold; T_{base} represents the pre-set basic recycling threshold; U represents the current memory usage; U_{target} represents the pre-set target memory usage; k is the pre-set adjustment coefficient, k reflects the responsiveness to memory pressure.

[0039] The forecast of future memory requirements includes:

[0040] Use a linear regression model to predict future memory requirements:

[0041] D(t)=\beta_0+\beta_1\cdot D(t-1)+\beta_2\cdot A(t-1)+\epsilon

[0042] in:

[0043] D(t) represents the memory requirement at time (t);

[0044] D(t-1) represents the memory requirement at the previous moment;

[0045] A(t-1) represents the memory access amount at the previous moment;

[0046] \beta_0,\beta_1,\beta_2 are all parameters of the linear regression model;

[0047] \epsilon is the error term set.

[0048] According to a second aspect, an edge-cloud collaboration system is provided, the system comprising: at least two edge nodes and a cloud service center node,

[0049] A dynamic resource impact interface and a static resource impact interface are configured on at least two edge nodes and at least one of the cloud service center nodes;

[0050] The node uses its own dynamic resource impact interface to perceive application load and resource usage in real time, and adjusts computing resources, storage resources, and network resources according to the perceived application load and resource usage;

[0051] The node uses its own static resources to influence interface configuration and manage the system's infrastructure resources;

[0052] At least two edge nodes and at least one of the cloud service center nodes obtain memory usage of the application process, and recycle and release memory according to the memory usage.

[0053] According to a third aspect, a computing device is provided, including a memory and a processor, wherein the memory stores executable code, and when the processor executes the executable code, the method described in any embodiment of the present invention is implemented.

[0054] It can be seen that each embodiment of the present invention has at least the following beneficial effects:

[0055] (1) By sensing application load and resource usage in real time, automatically adjusting resource allocation strategies, and achieving unified management and efficient use of infrastructure resources, the system ensures efficient operation and high-quality services.

[0056] (2) It reduces the complexity and operation and maintenance costs of the system and improves resource utilization and overall system performance.

[0057] (3) The resource response system based on the physical principle of the equipment is an efficient and reliable system resource management solution. It can not only achieve precise control and optimization of system resources, but also reduce system failure rate and maintenance costs, providing strong support for the digital transformation of enterprises.

[0058] (4) Effectively solves the memory leakage problem that may occur in intelligent reasoning services and improves the reliability and stability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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 creative work.

[0060] Figure 1 It is a schematic diagram of a system architecture applied in one embodiment of the present invention.

[0061] Figure 2 It is a flowchart of a data processing method in an edge-cloud collaborative system in one embodiment of the present invention.

[0062] Figure 3 It is a schematic diagram of classifying and acquiring static resources and dynamic resources based on a static resource impact interface and a dynamic resource impact interface respectively in an embodiment of the present invention.

[0063] Figure 4 It is a flowchart of a process of implementing a serverless processing based on a serverless reasoning resource architecture in one embodiment of the present invention. DETAILED DESCRIPTION

[0064] The solution provided by the present invention is described below in conjunction with the accompanying drawings.

[0065] First of all, it should be noted that the terms used in the embodiments of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The singular forms "a", "the" and "the" used in the embodiments of the present invention and the appended claims are also intended to include plural forms, unless the context clearly indicates otherwise.

[0066] It should be understood that the term "and / or" used in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0067] In order to facilitate understanding of the method provided by the present invention, the system architecture involved and applicable to the present invention is first described. Figure 1 As shown in , the system includes at least two edge nodes and a cloud service center node.

[0068] It should be understood that Figure 1 The number of edge nodes and cloud service center nodes in the example is only for reference. Any number can be selected and deployed according to the implementation requirements.

[0069] Figure 2 is a flow chart of a data processing method in an edge-cloud collaborative system in one embodiment of the present invention. Figure 1 and Figure 2 , the method comprising:

[0070] Step 201: configuring a dynamic resource impact interface and a static resource impact interface on a node in the edge-cloud collaborative system;

[0071] Step 203: The nodes in the edge-cloud collaborative system use their own dynamic resource impact interfaces to perceive application load and resource usage in real time, and adjust computing resources, storage resources, and network resources according to the perceived application load and resource usage;

[0072] Step 205: The nodes in the edge-cloud collaborative system use their own static resources to influence the interface configuration and manage the infrastructure resources of the system;

[0073] Step 207: The nodes in the edge-cloud collaborative system obtain the memory usage of the application process and recycle and release the memory according to the memory usage.

[0074] According to the above Figure 2It can be seen from the process shown that in the embodiment of the present invention, the dynamic resource impact interface and the static resource impact interface are configured on the nodes of the edge-cloud collaborative system. These two interfaces together constitute the basis of system resource management, ensuring that the system can be adaptively adjusted according to application requirements and environmental changes to achieve optimal operating efficiency and performance. In addition, memory is recycled and released based on memory usage, which improves memory usage efficiency.

[0075] In one embodiment of the present invention, in order to solve the common resource usage and fatigue management problems in the computing cloud platform, the following four technologies can be used:

[0076] 1. Build interfaces based on dynamic and static resources;

[0077] 2. Build a serverless reasoning resource architecture;

[0078] 3. Build a resource response system based on core physical principles;

[0079] 4. Memory self-healing management system based on memory leak.

[0080] First, let’s explain point 1: building interfaces based on dynamic and static resources.

[0081] In the prior art, the interface types of nodes are not differentiated, and all interfaces have unified functions. For example, the information transmitted on each interface often needs to be encrypted or a token is provided, which increases the complexity of processing.

[0082] See also Figure 3In the cloud-edge collaborative system, for example, when building an intelligent software reasoning service system, a core task is to design and implement dynamic resource impact interfaces and static resource impact interfaces that can flexibly adjust resource allocation. These two interfaces together constitute the basis of system resource management, ensuring that the system can make adaptive adjustments according to application requirements and environmental changes to achieve optimal operating efficiency and performance. First, for the dynamic resource impact interface. The main responsibility of this interface is to perceive application load and resource usage in real time, and automatically adjust computing resources, storage resources, and network resources accordingly. In the modern cloud computing environment, application loads often show a high degree of dynamics and uncertainty, which requires the system to have the ability to respond quickly and adaptively adjust. The dynamic resource impact interface can collect and analyze various performance indicators of the system in real time by integrating advanced monitoring technology and intelligent scheduling algorithms, such as CPU usage, memory usage, network bandwidth, etc., and dynamically adjust resource allocation strategies according to changes in these indicators. For example, when the load of an application suddenly increases, the interface can quickly identify this change and automatically increase the computing resources allocated to the application to ensure that it can run normally and meet user needs. This automated resource adjustment mechanism not only greatly improves the system's response speed and operating efficiency, but also significantly reduces the operation and maintenance costs. Next, for the static resource impact interface. Unlike the dynamic resource impact interface, the static resource impact interface is mainly responsible for configuring and managing the system's infrastructure resources, such as servers, storage devices, and network devices. These infrastructure resources are the basis for the operation of the system, and their performance and stability directly affect the efficiency and quality of the reasoning service. The static resource impact interface achieves unified management and efficient utilization of infrastructure resources by relying on network storage sharing and P2P resource sharing technologies. In terms of network storage sharing, the system can integrate the storage resources of each node to form a unified storage resource pool. In this way, no matter which node the reasoning service runs on, it can easily access the data in the entire storage resource pool, thereby greatly improving the availability and sharing of data. In terms of P2P resource sharing, the system further improves resource utilization and system scalability by realizing resource mutual assistance and sharing between nodes. Specifically, when a node has insufficient resources, it can borrow resources from other nodes to meet its own needs; when it has excess resources, it can contribute them to other nodes for use. This P2P-based resource sharing model not only reduces the overall cost of the system, but also improves resource utilization and overall system performance. In the static resource impact interface, resource matching and management is an important link. In order to ensure the accuracy and efficiency of resources, the system adopts a resource matching mechanism based on the SHA-256 algorithm. The SHA-256 algorithm is a secure hash algorithm that can map data of any length to a hash value of fixed length.In the system, each resource (such as servers, storage devices, network devices, etc.) is assigned a unique SHA-256 hash value as an identifier. When the system needs to find or access a resource, it only needs to search according to the hash value of the resource. This hash-based resource matching method can greatly improve the search efficiency and accuracy of the system, thereby ensuring rapid response and effective utilization of resources. In summary, dynamic and static resource impact interfaces are an important part of building an intelligent software reasoning service system based on cloud-edge collaboration. They ensure efficient operation and high-quality services of the system by sensing application load and resource usage in real time, automatically adjusting resource allocation strategies, and realizing unified management and efficient utilization of infrastructure resources.

[0083] It can be seen that in the embodiment of the present invention, in the above step 205, the nodes in the edge-cloud collaborative system use their own static resources to influence the interface configuration and the infrastructure resources of the management system, including:

[0084] In the edge-cloud collaborative system, the storage resources of multiple nodes are integrated to form a unified storage resource pool. When an inference service runs on any of the multiple nodes, the data in the entire storage resource pool is accessed through the static resource impact interface on the node to achieve network storage sharing.

[0085] or,

[0086] In the edge-cloud collaborative system, when a node has insufficient resources, it borrows resources from other nodes through the static resource impact interface to meet its own needs; and when the node has excess resources, it contributes the excess resources to other nodes through the static resource impact interface;

[0087] or,

[0088] Assign a unique SHA-256 hash value to the resources in the edge-cloud collaborative system as the identifier of the resource. After receiving the update package for the resource through the static resource impact interface, calculate the SHA-256 hash value of the update package of the resource, and compare the calculated SHA-256 hash value of the resource update package with the SHA-256 hash value assigned to the resource. If they are the same, the resource will not be updated. If they are different, the resource will be updated.

[0089] The second point is about building a serverless inference resource architecture.

[0090] See also Figure 4In the construction of a high-performance and resource-efficient intelligent software reasoning service system, the serverless reasoning resource architecture has become a key solution. This architecture aims to reduce the complexity of the system and improve resource utilization, thereby meeting the growing computing needs. In one embodiment of the present invention, the serverless reasoning resource architecture can realize on-demand allocation, dynamic, non-fixed, and highly concurrent processing threads and instances. By integrating containerization technology and microservice architecture, rapid deployment, flexible expansion, and seamless update of reasoning services are achieved, providing strong support for intelligent applications. First of all, the core of the serverless reasoning resource architecture lies in the application of its containerization technology. Containerization technology achieves the isolation, portability, and manageability of applications by packaging applications and their dependencies into independent containers. This means that each reasoning service can run independently in a container without being affected by other services, and it also facilitates the update and upgrade of a single service. In addition, containerization technology also makes it easier to dynamically allocate and recycle resources, thereby improving resource utilization. Secondly, the introduction of the microservice architecture further enhances the flexibility and scalability of the serverless reasoning resource architecture. The microservice architecture splits a complex system into a series of small, independent services, each of which focuses on completing a specific function. This architectural pattern makes the system easier to understand and maintain, while also improving the scalability and fault tolerance of the system. In the inference service, the microservice architecture allows us to quickly increase or decrease the number of service instances according to actual needs to meet different computing needs. In order to further improve the overall performance and response speed of the system, the serverless inference resource architecture also adopts an intelligent scheduling algorithm. This algorithm can perceive the system load in real time and dynamically adjust the allocation strategy of inference tasks according to the load. Through load balancing technology, the system can evenly distribute inference tasks to each service instance, avoiding the situation where a service instance is overloaded while other instances are idle. At the same time, the intelligent scheduling algorithm can also dynamically adjust the resource allocation strategy based on historical data and prediction models to cope with possible load changes in the future. In the serverless inference resource architecture, the dynamic scaling resource mechanism of the serverless interface is the key to achieving high concurrency and memory optimization. This mechanism allows the system to dynamically increase or decrease the number of service instances according to the real-time load situation. When the system load increases, the serverless interface can quickly start a new service instance to handle additional inference tasks; when the load decreases, some service instances can be shut down to release resources. This dynamic scaling mechanism not only ensures that the system can cope with high concurrency requirements, but also avoids the waste and leakage of memory resources. In addition, the serverless reasoning resource architecture does not need to save process data during the reasoning process, thereby maximizing resource utilization. In traditional reasoning services, the system usually needs to save a large amount of intermediate data and status information to support subsequent reasoning processes.However, in the serverless reasoning resource architecture, due to the stateless design concept, each service instance can independently process reasoning tasks without relying on other service instances or saving intermediate data, which allows the system to use computing resources more efficiently and also improves the scalability and fault tolerance of the system. In summary, the serverless reasoning resource architecture integrates advanced technologies such as containerization technology, microservice architecture, and intelligent scheduling algorithms to achieve rapid deployment, flexible expansion, and efficient operation of reasoning services. This architecture not only reduces the complexity and operation and maintenance costs of the system, but also improves resource utilization and the overall performance of the system.

[0091] Therefore, in an embodiment of the present invention, the data processing method based on the edge-cloud collaborative system may further include:

[0092] The node dynamically increases or decreases the number of service instances based on the real-time load. When the system load increases, new service instances are started to handle additional reasoning tasks; when the load decreases, some service instances are shut down to release resources, thus realizing a serverless reasoning resource architecture.

[0093] and / or,

[0094] Package applications and their dependencies into independent containers to achieve application isolation, portability, and manageability;

[0095] and / or,

[0096] Break complex systems into a series of small, independent services, each of which performs a specific function.

[0097] Now let’s talk about point 3: building a resource response system based on core physics processing.

[0098] Building a resource response system based on the physical principles of equipment is an important way to achieve precise control and optimization of system resources. Based on real-time monitoring and analysis of the system's operating status, the system combines physical principles and mathematical models to predict, schedule and configure system resources, thereby ensuring efficient and stable operation of the system. First, the system collects various operating data of the system, such as CPU utilization, memory usage, disk space utilization, network connection status, etc., through real-time monitoring and analysis technology. These data are the key to understanding the current status of the system and the basis for subsequent prediction and scheduling. Secondly, the system combines physical principles and mathematical models to process and analyze the collected data. Through the integration of deep learning algorithms and the physical structure of hardware, the system can better understand the working principles and failure modes of the system, thereby predicting future resource requirements. This prediction is not a simple trend analysis, but an in-depth insight based on the physical characteristics of the system, which can more accurately reflect the real needs of the system. Based on the prediction results, the system can schedule and configure resources in advance. For example, when it is predicted that a service is about to face high concurrent requests, the system can automatically increase the corresponding computing resources to ensure the stable operation of the service. Similarly, when the load on a certain node is too low, the system can transfer the resources on it to other places where they are needed, realizing global optimization of resources. This resource response system based on the physical principle of the device can not only improve the utilization and stability of system resources, but also reduce the system failure rate and maintenance cost. Through real-time prediction and scheduling, the system can timely discover and solve potential problems and avoid the occurrence of failures. At the same time, since the system can understand the physical characteristics of the system more accurately, when a failure occurs, the system can locate the problem more quickly and take corresponding repair measures. In addition, the integration of deep learning algorithms and hardware physical structures enables the system to better solve failure problems. Deep learning algorithms can learn the failure mode of the system from a large amount of data and predict future failure trends. The physical structure of the hardware provides the physical basis for solving failures. By combining the two, the system can achieve accurate prediction and rapid repair of failures, further improving the reliability and availability of the system. In short, the resource response system based on the physical principle of the device is an efficient and reliable system resource management solution. It can not only achieve accurate control and optimization of system resources, but also reduce the system failure rate and maintenance cost, providing strong support for the digital transformation of enterprises.

[0099] Therefore, based on the above point 3, it can be concluded that in an embodiment of the present invention, the CPU utilization, memory usage, disk space utilization, and network connection status of the nodes can be collected through real-time monitoring and analysis, and the deep learning algorithm and the collected CPU utilization, memory usage, disk space utilization, and network connection status of the nodes can be used to predict future resource requirements.

[0100] The following explains point 4: memory self-healing management system based on memory leak.

[0101] In order to solve the memory leakage problem that may occur in intelligent reasoning services, the present invention proposes an innovative memory self-healing management system based on memory leakage. The system uses a serverless interface to count resource errors, thereby realizing real-time monitoring and automatic processing of memory leakage. In intelligent reasoning services, memory leakage is a common problem, which may lead to system performance degradation, prolonged response time and even service crash. Traditional memory management methods often require manual intervention, which is not only inefficient, but also difficult to detect potential memory leakage problems. In order to overcome these limitations, the memory self-healing management system of this article came into being. The system first counts resource errors through a serverless interface to realize real-time monitoring of the memory usage of the application process. This serverless interface design enables the system to obtain the memory usage data of the application process in real time and analyze it accurately. Once the system detects abnormal memory usage, such as excessive memory usage or insufficient memory release, the early warning mechanism of memory leakage will be immediately triggered. After discovering the potential memory leakage problem, the memory self-healing management system will automatically recycle and release the memory. Through intelligent algorithms and strategies, the system can accurately determine which memory blocks can be recycled and which need to be retained. Then, the system automatically releases those memory blocks that are no longer needed to free up system resources and prevent further deterioration of memory leaks. In addition to automatically reclaiming and releasing memory, the system also provides memory leak analysis and diagnostic tools. These tools can help developers quickly locate and solve memory leaks. By deeply analyzing memory usage data and code logic, developers can find out the root cause of memory leaks and take corresponding repair measures. This ability to quickly locate and solve problems greatly improves the reliability and stability of the system. In addition, the memory self-healing management system is also scalable and customizable. It can be flexibly configured and adjusted according to different application scenarios and requirements. For example, developers can set parameters such as memory usage thresholds and recycling strategies according to actual conditions to meet the special needs of the system. This flexibility and customizability enables the memory self-healing management system to better adapt to various intelligent reasoning service scenarios. In short, the memory self-healing management system based on memory leaks is an efficient, reliable and flexible system resource management solution. It effectively solves the memory leak problem that may occur in intelligent reasoning services by monitoring the memory usage of application processes in real time, automatically reclaiming and releasing memory, and providing memory leak analysis and diagnostic tools, thereby improving the reliability and stability of the system.

[0102] Based on the above point 4, it can be obtained that in one embodiment of the present invention, in the above step 207, the node in the edge-cloud collaborative system obtains the memory usage of the application process and recycles and releases the memory according to the memory usage, including:

[0103] Step 2071: The nodes in the edge-cloud collaborative system execute, for each memory block, the following: collecting the access frequency to the memory block, the last access time to the memory block, and the life cycle of the memory block;

[0104] Step 2073: Calculate an importance score for each memory block based on the collected access frequency, last access time, and life cycle of each memory block;

[0105] Step 2075: predict future memory requirements;

[0106] Step 2077: According to the predicted memory demand and the importance score of each memory block, memory blocks are recycled starting from memory blocks with low importance scores, and memory blocks are recycled based on a dynamically adjusted memory recycling strategy.

[0107] In one embodiment of the present invention, in step 2073, calculating the importance score for each memory block includes:

[0108] Use the following calculation formula to calculate:

[0109] S(b)=w_a\cdot A(b)+w_t\cdot T(b)+w_l\cdot L(b)]

[0110] Wherein: S(b) represents the importance score of memory block b; A(b) represents the access frequency to memory block b; T(b) represents the last access time to memory block b; L(b) represents the life cycle of memory block b, which is the length of time from the creation of memory block b to the current time; w_a, w_t, w_l are the weights of access frequency, last access time and life cycle respectively, and w_a+w_t+w_l=1.

[0111] In one embodiment of the present invention, in the above step 2077, reclaiming the memory block based on the dynamically adjusted memory reclaiming strategy includes:

[0112] Dynamically calculate the memory recycling threshold;

[0113] When the current memory usage is greater than the latest calculated memory recycling threshold, memory block recycling begins;

[0114] The memory recycling threshold is dynamically calculated using the following formula:

[0115] T_{new}=T_{base}+k\cdot(U-U_{target})

[0116] Among them: T_{new} represents the latest recycling threshold; T_{base} represents the pre-set basic recycling threshold; U represents the current memory usage; U_{target} represents the pre-set target memory usage; k is the pre-set adjustment coefficient, k reflects the responsiveness to memory pressure.

[0117] In one embodiment of the present invention, predicting future memory requirements in step 2075 includes:

[0118] Use a linear regression model to predict future memory requirements:

[0119] D(t)=\beta_0+\beta_1\cdot D(t-1)+\beta_2\cdot A(t-1)+\epsilon

[0120] in:

[0121] D(t) represents the memory requirement at time (t);

[0122] D(t-1) represents the memory requirement at the previous moment;

[0123] A(t-1) represents the memory access amount at the previous moment;

[0124] \beta_0,\beta_1,\beta_2 are all parameters of the linear regression model;

[0125] \epsilon is the error term set.

[0126] In the embodiment of the present invention, by implementing the above-mentioned calculation formulas, an intelligent memory recycling system can be constructed to effectively improve memory utilization, reduce memory fragmentation, and thus improve overall system performance and response speed.

[0127] See also Figure 1 , an embodiment of the present invention provides an edge-cloud collaboration system, the system comprising: at least two edge nodes and a cloud service center node,

[0128] A dynamic resource impact interface and a static resource impact interface are configured on at least two edge nodes and at least one of the cloud service center nodes;

[0129] The node uses its own dynamic resource impact interface to perceive application load and resource usage in real time, and adjusts computing resources, storage resources, and network resources according to the perceived application load and resource usage;

[0130] The node uses its own static resources to influence interface configuration and manage the system's infrastructure resources;

[0131] At least two edge nodes and at least one of the cloud service center nodes obtain memory usage of the application process, and recycle and release memory according to the memory usage.

[0132] It can be seen that each embodiment of the present invention has at least the following beneficial effects:

[0133] (1) By sensing application load and resource usage in real time, automatically adjusting resource allocation strategies, and achieving unified management and efficient use of infrastructure resources, the system ensures efficient operation and high-quality services.

[0134] (2) It reduces the complexity and operation and maintenance costs of the system and improves resource utilization and overall system performance.

[0135] (3) The resource response system based on the physical principle of the equipment is an efficient and reliable system resource management solution. It can not only achieve precise control and optimization of system resources, but also reduce system failure rate and maintenance costs, providing strong support for the digital transformation of enterprises.

[0136] (4) Effectively solves the memory leakage problem that may occur in intelligent reasoning services and improves the reliability and stability of the system.

[0137] An embodiment of the present invention provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed in a computer, the computer is caused to execute a method in any one of the embodiments in the specification.

[0138] An embodiment of the present invention provides a computing device, including a memory and a processor, wherein the memory stores executable code, and when the processor executes the executable code, the method in any embodiment of the specification is implemented.

[0139] It is to be understood that the structure illustrated in the embodiment of the present invention does not constitute a specific limitation on the device of the embodiment of the present invention. In other embodiments of the specification, the above-mentioned device may include more or fewer components than those shown in the figure, or combine some components, or split some components, or arrange the components differently. The components shown in the figure may be implemented in hardware, software, or a combination of software and hardware.

[0140] The various embodiments of the present invention are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0141] Those skilled in the art should be aware that in one or more of the above examples, the functions described in the present invention can be implemented by hardware, software, widgets, or any combination thereof. When implemented by software, these functions can be stored in a computer-readable medium or transmitted as one or more instructions or codes on a computer-readable medium.

[0142] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made on the basis of the technical solution of the present invention should be included in the scope of protection of the present invention.

Claims

1. A data processing method in an edge-cloud collaborative system, characterized in that: The method includes: Configure dynamic resource impact interfaces and static resource impact interfaces on nodes in the edge-cloud collaborative system; The nodes in the edge-cloud collaborative system use their own dynamic resource impact interfaces to perceive application load and resource usage in real time, and adjust computing resources, storage resources, and network resources based on the perceived application load and resource usage; Nodes in the edge-cloud collaborative system use their own static resources to influence interface configuration and manage the infrastructure resources of the system; The nodes in the edge-cloud collaborative system obtain the memory usage of the application process and recycle and release the memory based on the memory usage.

2. The method according to claim 1, characterized in that The nodes in the edge-cloud collaborative system use their own static resources to influence the interface configuration and manage the infrastructure resources of the system, including: In the edge-cloud collaborative system, the storage resources of multiple nodes are integrated to form a unified storage resource pool. When an inference service runs on any of the multiple nodes, the data in the entire storage resource pool is accessed through the static resource impact interface on the node to achieve network storage sharing. or, In the edge-cloud collaborative system, when a node has insufficient resources, it borrows resources from other nodes through the static resource impact interface to meet its own needs; and when the node has excess resources, it contributes the excess resources to other nodes through the static resource impact interface; or, Assign a unique SHA-256 hash value to the resources in the edge-cloud collaborative system as the identifier of the resource. After receiving the update package for the resource through the static resource impact interface, calculate the SHA-256 hash value of the update package of the resource, and compare the calculated SHA-256 hash value of the resource update package with the SHA-256 hash value assigned to the resource. If they are the same, the resource will not be updated. If they are different, the resource will be updated.

3. The method according to claim 1, characterized in that The method further comprises: The node dynamically increases or decreases the number of service instances based on the real-time load. When the system load increases, new service instances are started to handle additional reasoning tasks; when the load decreases, some service instances are shut down to release resources, thus realizing a serverless reasoning resource architecture. and / or, Package applications and their dependencies into independent containers to achieve application isolation, portability, and manageability; and / or, Break complex systems into a series of small, independent services, each of which performs a specific function.

4. The method according to claim 1, characterized in that: Through real-time monitoring and analysis, the CPU utilization, memory usage, disk space utilization, and network connection status of the nodes are collected. The future resource requirements are predicted by using deep learning algorithms and the collected CPU utilization, memory usage, disk space utilization, and network connection status of the nodes.

5. The method according to claim 1, characterized in that The nodes in the edge-cloud collaborative system obtain the memory usage of the application process and recycle and release the memory according to the memory usage, including: For each memory block, the nodes in the edge-cloud collaborative system perform the following operations: collecting the access frequency to the memory block, the last access time to the memory block, and the life cycle of the memory block; Calculate the importance score for each memory block based on the collected access frequency, last access time, and life cycle of each memory block; Predict future memory requirements; According to the predicted memory demand and the importance score of each memory block, memory blocks are recycled starting from memory blocks with low importance scores, and memory blocks are recycled based on a dynamically adjusted memory recycling strategy.

6. The method according to claim 5, characterized in that The importance score is calculated for each memory block, including: Use the following calculation formula to calculate: S(b)=w_a\cdot A(b)+w_t\cdot T(b)+w_l\cdot L(b)] Wherein: S(b) represents the importance score of memory block b; A(b) represents the access frequency to memory block b; T(b) represents the last access time to memory block b; L(b) represents the life cycle of memory block b, which is the length of time from the creation of memory block b to the current time; w_a, w_t, w_l are the weights of access frequency, last access time and life cycle respectively, and w_a+w_t+w_l=1.

7. The method according to claim 5, characterized in that The memory block recycling based on the dynamically adjusted memory recycling strategy includes: Dynamically calculate the memory recycling threshold; When the current memory usage is greater than the latest calculated memory recycling threshold, memory block recycling begins; The memory recycling threshold is dynamically calculated using the following formula: T_{new}=T_{base}+k\cdot(U-U_{target}) Among them: T_{new} represents the latest recycling threshold; T_{base} represents the pre-set basic recycling threshold; U represents the current memory usage; U_{target} represents the pre-set target memory usage; k is the pre-set adjustment coefficient, k reflects the responsiveness to memory pressure.

8. The method according to claim 5, characterized in that The forecast of future memory requirements includes: Use a linear regression model to predict future memory requirements: D(t)=\beta_0+\beta_1\cdot D(t-1)+\beta_2\cdot A(t-1)+\epsilon in: D(t) represents the memory requirement at time (t); D(t-1) represents the memory requirement at the previous moment; A(t-1) represents the memory access amount at the previous moment; \beta_0,\beta_1,\beta_2 are all parameters of the linear regression model; \epsilon is the error term set.

9. Edge-cloud collaborative system, characterized by: The system includes: at least two edge nodes and a cloud service center node, A dynamic resource impact interface and a static resource impact interface are configured on at least two edge nodes and at least one of the cloud service center nodes; The node uses its own dynamic resource impact interface to perceive application load and resource usage in real time, and adjusts computing resources, storage resources, and network resources according to the perceived application load and resource usage; The node uses its own static resources to influence interface configuration and manage the system's infrastructure resources; At least two edge nodes and at least one of the cloud service center nodes obtain memory usage of the application process, and recycle and release the memory according to the memory usage.

10. A computing device, comprising a memory and a processor, wherein the memory stores executable codes, and when the processor executes the executable codes, the method according to any one of claims 1 to 8 is implemented.

Images