Edge computing resource scheduling method and system for medical Internet of Things equipment

Abstract

The invention provides an edge computing resource scheduling method and system for medical Internet of Things equipment. The edge computing resource scheduling method of the medical Internet of Things equipment comprises the following steps: acquiring a cluster characteristic parameter corresponding to a medical Internet of Things equipment cluster through the medical Internet of Things equipment cluster in real time, and generating a medical task characteristic coefficient by utilizing the cluster characteristic parameter; according to the medical task characteristic coefficient, determining a load intensity degree of an edge node corresponding to a medical Internet of Things equipment cluster; and when the load intensity degree of the edge node is relatively large, dynamically adjusting the edge node corresponding to the medical Internet of Things equipment cluster. The system comprises modules corresponding to the steps of the method.

Description

Technical Field

[0001] The present invention relates to the technical field of edge computing resource scheduling, and particularly to the technical field of edge computing resource scheduling for medical Internet of Things devices. Background Art

[0002] In traditional medical Internet of Things systems, there is a lack of an effective evaluation method for the characteristics of medical tasks undertaken by a medical Internet of Things device cluster; usually, some simple indicators are used to understand the operating status of the device cluster, and it is difficult to comprehensively and accurately obtain the complexity and diversity of medical tasks. Under different medical tasks, there are various different requirements, and there is a lack of a method that can comprehensively consider various factors and quantitatively evaluate the characteristics of medical tasks, resulting in the inability to reasonably allocate resources according to the actual needs of medical tasks, etc. Summary of the Invention

[0003] The present invention provides a method and system for edge computing resource scheduling of medical Internet of Things devices to solve the above technical problems:

[0004] A method for edge computing resource scheduling of medical Internet of Things devices, the method for edge computing resource scheduling of medical Internet of Things devices comprising:

[0005] Real-time obtaining cluster characteristic parameters corresponding to a medical Internet of Things device cluster through the medical Internet of Things device cluster, and generating a medical task characteristic coefficient by using the cluster characteristic parameters;

[0006] Determining the load intensity degree of an edge node corresponding to the medical Internet of Things device cluster according to the medical task characteristic coefficient;

[0007] When the load intensity degree of the edge node is relatively large, dynamically adjusting the edge node corresponding to the medical Internet of Things device cluster.

[0008] Further, real-time obtaining cluster characteristic parameters corresponding to a medical Internet of Things device cluster through the medical Internet of Things device cluster, and generating a medical task characteristic coefficient by using the cluster characteristic parameters, comprising:

[0009] Real-time monitoring the operating parameters of each medical Internet of Things device included in each medical Internet of Things device cluster

[0010] Obtaining an operating characteristic vector corresponding to each medical Internet of Things device by using the operating parameters of each medical Internet of Things device;

[0011] Obtaining a medical task characteristic coefficient corresponding to the medical Internet of Things device cluster by using the norm corresponding to the operating characteristic vector of each medical Internet of Things device.

[0012] Further, determining the load intensity degree of an edge node corresponding to the medical Internet of Things device cluster according to the medical task characteristic coefficient, comprising:

[0013] Extract the operating parameters of the edge nodes when the medical task characteristic coefficients are generated by each medical Internet of Things device cluster each time;

[0014] Use the operating parameters of the edge nodes to obtain the resource linkage coefficient between the edge nodes and the medical Internet of Things device cluster;

[0015] Determine the load intensity degree of the edge nodes according to the relationship between the resource linkage coefficient and the preset linkage coefficient threshold.

[0016] Further, the resource linkage coefficient is obtained through the following formula:

[0017] R = (1 - P e )·(1 - P c )·(1 - P t )·σ

[0018] Wherein, R represents the resource linkage coefficient; P e , P c and P t represent the proportion of node operation energy consumption, CPU occupancy rate, and task processing time delay ratio; σ represents the medical task characteristic coefficient.

[0019] Further, determining the load intensity degree of the edge nodes according to the relationship between the resource linkage coefficient and the preset linkage coefficient threshold includes:

[0020] Compare the resource linkage coefficient with the preset linkage coefficient threshold;

[0021] When the resource linkage coefficient is lower than the preset linkage coefficient threshold, it is determined that the load intensity of the edge nodes corresponding to the medical Internet of Things device cluster is relatively large.

[0022] Further, when the load intensity degree of the edge nodes is relatively large, perform dynamic adjustment on the edge nodes corresponding to the medical Internet of Things device cluster, including:

[0023] When the load intensity degree of the edge nodes is relatively large, detect the resource fragmentation status of each edge node in the medical Internet of Things platform, and determine whether the resource fragmentation status reaches the screening condition;

[0024] Screen out the edge nodes whose resource fragmentation status reaches the screening condition as candidate edge nodes;

[0025] Determine the target node from the candidate edge nodes, and use the target node to replace the edge node corresponding to the medical Internet of Things device cluster.

[0026] Further, detecting the resource fragmentation status of each edge node in the medical Internet of Things platform includes:

[0027] Retrieve the medical density coefficient of each edge node and the effective parameters of aggregable resources of each edge node;

[0028] Extract the medical task feature coefficients of the medical IoT device clusters corresponding to each edge node;

[0029] Use the medical density coefficient, the effective parameters of aggregable resources of each edge node, and the medical task feature coefficients of the medical IoT device clusters corresponding to each edge node to obtain the node resource coefficients corresponding to each edge node.

[0030] Furthermore, determine whether the resource fragmentation state reaches the screening condition, including:

[0031] Compare the node resource coefficients corresponding to each edge node with a preset resource coefficient threshold;

[0032] When the node resource coefficient is not lower than the preset resource coefficient threshold, it indicates that the edge node reaches the screening condition.

[0033] An edge computing resource scheduling system for medical IoT devices, the edge computing resource scheduling system for medical IoT devices includes:

[0034] A medical task feature coefficient acquisition module, configured to obtain, in real time, the cluster feature parameters corresponding to the medical IoT device cluster through the medical IoT device cluster, and generate medical task feature coefficients by using the cluster feature parameters;

[0035] A load intensity determination module, configured to determine the load intensity level of the edge node corresponding to the medical IoT device cluster according to the medical task feature coefficients;

[0036] A node adjustment module, configured to dynamically adjust the edge node corresponding to the medical IoT device cluster when the load intensity level of the edge node is relatively high.

[0037] Advantages of the present invention:

[0038] A method and system for edge computing resource scheduling of medical Internet of Things devices proposed by the present invention can improve the accuracy of obtaining the task requirements of the medical Internet of Things device cluster and the load conditions of edge nodes by generating medical task characteristic coefficients. Dynamically adjusting the resource allocation of edge nodes according to the load intensity can avoid the idle and waste of resources and improve the overall utilization rate of edge computing resources. When the load intensity of the edge node is relatively large, timely dynamic adjustment can effectively avoid system crashes or performance degradation caused by resource overload. Through reasonable task migration and resource allocation, the loads of each edge node are kept relatively balanced. The optimized resource scheduling method can ensure that medical tasks are processed in a timely manner and reduce task processing delays. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 FIG. is a schematic diagram of a method for edge computing resource scheduling of a medical Internet of Things device;

[0040] Figure 2 FIG. is a schematic diagram of a system for edge computing resource scheduling of a medical Internet of Things device. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0041] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only for explaining and illustrating the present invention and are not used to limit the present invention.

[0042] A method for edge computing resource scheduling of a medical Internet of Things device, as Figure 1 shown, the method for edge computing resource scheduling of the medical Internet of Things device includes:

[0043] Obtaining in real time the cluster characteristic parameters corresponding to the medical Internet of Things device cluster through the medical Internet of Things device cluster, and generating medical task characteristic coefficients by using the cluster characteristic parameters;

[0044] Determining the load intensity of the edge node corresponding to the medical Internet of Things device cluster according to the medical task characteristic coefficients;

[0045] When the load intensity of the edge node is relatively large, dynamically adjusting the edge node corresponding to the medical Internet of Things device cluster.

[0046] The working principle of the above technical solution is as follows: A medical Internet of Things device cluster is formed by medical Internet of Things devices, and relevant data is obtained from the devices in real time through a data acquisition interface or protocol as cluster characteristic parameters.

[0047] Analyzing and processing the obtained cluster characteristic parameters to generate medical task characteristic coefficients. The characteristics of the medical tasks borne by the current medical Internet of Things device cluster are reflected by the medical task characteristic coefficients.

[0048] Determine the load intensity degree of the current edge node according to data such as characteristic coefficients.

[0049] When it is determined that the load intensity degree of the edge node is relatively large, perform dynamic adjustment on the edge node.

[0050] The effects of the above technical solutions are as follows: By generating medical task characteristic coefficients, the accuracy of obtaining the task requirements of the medical Internet of Things device cluster and the load conditions of edge nodes can be improved. Dynamically adjusting the resource allocation of edge nodes according to the load intensity degree avoids the idle and waste of resources, and improves the overall utilization rate of edge computing resources. When the load intensity degree of the edge node is relatively large, timely dynamic adjustment can effectively avoid system crashes or performance degradation problems caused by resource overload. Through reasonable task migration and resource allocation, the loads of each edge node are kept relatively balanced. The optimized resource scheduling method can ensure that medical tasks are processed in a timely manner and reduce task processing delays.

[0051] In one embodiment of the present invention, real-time obtain the cluster characteristic parameters corresponding to the medical Internet of Things device cluster through the medical Internet of Things device cluster, and generate medical task characteristic coefficients by using the cluster characteristic parameters, including:

[0052] Real-time monitor the operating parameters of each medical Internet of Things device included in each medical Internet of Things device cluster;

[0053] Obtain the corresponding operating characteristic vector of each medical Internet of Things device by using the operating parameters of each medical Internet of Things device;

[0054] Among them, the structure of the operating characteristic vector corresponding to each medical Internet of Things device is as follows:

[0055] K = [a 01 , a 02 , a 03 , a 04

[0056] Among them, K represents the operating characteristic vector corresponding to each medical Internet of Things device; a 01 , a 02 , a 03 and a 04 respectively represent the operating delay duration ratio, the amount of data generated during operation, the medical risk coefficient, and the privacy sensitivity coefficient;

[0057] Obtain the medical task characteristic coefficient corresponding to the medical Internet of Things device cluster by using the norm corresponding to the operating characteristic vector of each medical Internet of Things device. Among them, the medical task characteristic coefficient is obtained through the following formula:

[0058] ​

[0059] Among them, σ represents the medical task feature coefficient; F p represents the average value of the norms corresponding to the operation feature vectors of the medical IoT devices included in the medical IoT device cluster; F b represents the standard deviation value of the norms corresponding to the operation feature vectors of the medical IoT devices included in the medical IoT device cluster; F max and F min represent the maximum and minimum values of the norms corresponding to the operation feature vectors of the medical IoT devices included in the medical IoT device cluster.

[0060] The working principle of the above technical solution is as follows: The operation parameters of each medical IoT device included in each medical IoT device cluster are monitored in real time; among them, the operation parameters include the operation delay duration ratio, the amount of data generated during operation, the medical risk coefficient, and the privacy sensitivity coefficient, and the privacy sensitivity coefficient is set in advance according to the privacy level of each medical IoT device, and the value range of the privacy sensitivity coefficient is 0-1;

[0061] The operation feature vector corresponding to each medical IoT device is obtained by using the operation parameters of each medical IoT device;

[0062] Among them, the structure of the operation feature vector corresponding to each medical IoT device is as follows:

[0063] K = [a 01 , a 02 , a 03 , a 04

[0064] Among them, K represents the operation feature vector corresponding to each medical IoT device; a 01 , a 02 , a 03 and a 04 respectively represent the operation delay duration ratio, the amount of data generated during operation, the medical risk coefficient, and the privacy sensitivity coefficient. Among them, the privacy sensitivity coefficient is a parameter that quantifies the privacy leakage risk of medical data types and is used to dynamically adjust the data encryption strength and access control policy. Its value range is 0-1. The larger the value, the higher the data sensitivity, and more stringent privacy protection measures are required. And the privacy sensitivity coefficient can be set according to the number of privacy fields of the physiological data generated by the medical IoT device in combination with the preset basic score of the privacy field and the preset basic privacy coefficient of the device. The setting formula is:

[0065] Privacy sensitivity coefficient = device basic privacy coefficient × (1 + basic score of privacy field × number of privacy fields)

[0066] ​For example, the data obtained by a blood glucose meter is in the form of blood glucose data + timestamp + device ID, which can locate the patient. Among them, the blood glucose data is the basic data, and the timestamp and device ID for patient location are privacy fields (i.e., the number of privacy fields is 2). At the same time, the basic score of the preset privacy field is 0.2, and the preset basic privacy coefficient of the device is 0.3. Then, the privacy sensitivity coefficient of the medical IoT device is a 04 = 0.3×(1 + 0.2×2) = 0.42;

[0067] At the same time, the medical risk coefficient a 03 is obtained through the following formula:

[0068]

[0069] where a 03 represents the medical risk coefficient; n represents the number of times of biological signals collected by each medical IoT device; m represents the types of biological signals collected by each medical IoT device; Δy ij represents the deviation value of the j-th biological signal detected by the medical IoT device for the i-th time compared to its standard biological signal; y j represents the standard biological signal value corresponding to the j-th biological signal; w represents the department risk factor of the department where the medical IoT device is located, which is a weight parameter quantifying the clinical emergency degree of different medical departments and is used to dynamically adjust the resource scheduling priority. The department risk factor reflects the positive correlation between the risk of patient's condition deterioration and the timeliness of rescue. The higher the w value, the more stringent the real-time requirement for computing resources in this department; and the specific value of the department risk factor can be set according to the actual situation of the hospital department where the factor acts.

[0070] The medical task feature coefficient corresponding to the medical IoT device cluster is obtained by using the norm corresponding to the operation feature vector of each medical IoT device. Among them, the norm corresponding to the operation feature vector of each medical IoT device is the cluster feature parameter corresponding to the medical IoT device cluster, and the medical task feature coefficient is obtained through the following formula:

[0071]

[0072] where σ represents the medical task feature coefficient; F p represents the average value of the norms corresponding to the operation feature vectors of the medical IoT devices included in the medical IoT device cluster; F b represents the standard deviation of the norms corresponding to the operation feature vectors of the medical IoT devices included in the medical IoT device cluster; F max and F min represent the maximum and minimum values of the norms corresponding to the operation feature vectors of the medical IoT devices included in the medical IoT device cluster.

[0073] When the medical task characteristic coefficient is larger, it indicates that the dispersion degree of the corresponding norm of the device operation characteristic vectors in the medical Internet of Things device cluster is relatively large, that is, the device operation characteristics are quite different, and the imbalance of the operation states among devices is relatively significant. Furthermore, it increases the operation load of the device cluster. That is, the larger the medical task characteristic coefficient, the greater the operation load of the device cluster, belonging to a high-load operation scenario.

[0074] The effects of the above technical solutions are as follows: The operation state of the device can be analyzed comprehensively and from multiple angles through the operation parameters. The medical risk coefficient improves the accuracy of abnormal recognition of the data collected by the device. Integrating multiple operation parameters into one vector enhances the intuitiveness of the operation characteristics, and the operation characteristic vectors of different devices can be compared with each other to obtain the differences between devices. Taking the corresponding norm of the operation characteristic vector of each device as the cluster characteristic parameter of the medical Internet of Things device cluster can understand the operation status of the device cluster as a whole. The medical task characteristic coefficient comprehensively considers the average level, dispersion degree, and extreme value situation of the device operation characteristics in the device cluster, and can accurately reflect the operation load of the device cluster. At the same time, when the medical task characteristic coefficient is larger, it indicates that the dispersion degree of the corresponding norm of the device operation characteristic vectors in the device cluster is relatively large, the device operation characteristics are quite different, and the imbalance of the operation states among devices is relatively significant, thus increasing the operation load of the device cluster. Improve the timeliness of identifying the high-load operation scenario, and avoid the device cluster from malfunctioning or experiencing performance degradation due to overload. In the high-load operation scenario, reasonably increase the computing resources, storage resources, or network bandwidth to ensure that the device cluster can operate stably and process medical tasks in a timely manner. By optimizing the resource allocation, reduce the resource competition and imbalance among devices, and improve the overall performance and efficiency of the device cluster. Greatly improve the response speed and processing ability of the medical Internet of Things system, and ensure the timeliness and accuracy of remote medical diagnosis. Introduce the privacy sensitivity coefficient and reflect it in the device operation characteristic vector, so that the privacy protection of medical data can be fully considered in the process of resource scheduling and management of the device cluster.

[0075] On the other hand, by monitoring the operating parameters of medical Internet of Things devices in real time, changes in the device status can be captured in a timely manner. Using these parameters to generate an operating feature vector, and then obtaining the medical task feature coefficient, which can accurately reflect the resource demand characteristics of the current medical task, providing an accurate basis for subsequent resource allocation, scheduling and other decisions, and improving the timeliness and adaptability of the system's response to medical tasks. The operating feature vector covers multi-dimensional information such as the operating delay duration ratio, the amount of data generated during operation, the medical risk coefficient, and the privacy sensitivity coefficient, comprehensively depicting the device operating conditions and task characteristics. Based on the generated medical task feature coefficient, the system can more comprehensively evaluate the resource requirements of medical tasks, avoid one-sidedness caused by evaluating based on a single indicator, improve the scientific nature of resource management decisions, and optimize the overall performance of the system. Accurately obtaining the medical task feature coefficient helps the system to more reasonably allocate resources such as computing and storage, avoiding over-allocation or under-allocation of resources. For example, for tasks with a high medical risk coefficient, a large amount of data, and high real-time requirements (low operating delay duration ratio), high-quality resources can be preferentially allocated to improve the quality and efficiency of medical services, reduce the degree of resource fragmentation, and improve performance indicators such as resource utilization rate.

[0076] At the same time, the standard deviation value F of the norm corresponding to the operating feature vector is introduced into the formula of the above technical solution b , which reflects the dispersion degree of the device operating feature vector, effectively avoiding task execution obstacles caused by device differences, and ensuring the stability and reliability of system operation. And by using F max and F min to participate in the formula calculation, the maximum and minimum values of the norm of the device operating feature vector are considered. It can effectively prevent excessive interference of the extreme operating state of individual devices on the calculation of the medical task feature coefficient, making the calculation result more robust. At the same time, by considering the extreme values, the system can make resource reservation or adjustment strategies in advance for extreme situations, improve the system's ability to handle abnormal situations, and ensure the stability of performance indicators. Based on the average value F p of the norm corresponding to the operating feature vector, it reflects the overall operating feature level of the medical Internet of Things device cluster. On this basis, combining other parameters to calculate the medical task feature coefficient can balance the differences of individual devices, reflect the resource demand characteristics of medical tasks from the overall level, provide a reasonable quantitative reference for resource allocation, improve the accuracy of resource allocation, and then optimize the system performance.

[0077] In one embodiment of the present invention, determining the load intensity degree of the edge node corresponding to the medical Internet of Things device cluster according to the medical task feature coefficient includes:

[0078] Extracting the operating parameters of the edge node when the medical Internet of Things device cluster generates the medical task feature coefficient each time; wherein, the operating parameters include the proportion of node operating energy consumption, CPU occupancy rate, and task processing time delay ratio;

[0079] Obtain the resource linkage coefficient between the edge node and the medical Internet of Things device cluster by using the operating parameters of the edge node;

[0080] Determine the load intensity degree of the edge node according to the relationship between the resource linkage coefficient and the preset linkage coefficient threshold. Specifically, compare the resource linkage coefficient with the preset linkage coefficient threshold; when the resource linkage coefficient is lower than the preset linkage coefficient threshold, it is determined that the load intensity of the edge node corresponding to the medical Internet of Things device cluster is large.

[0081] Among them, the resource linkage coefficient is obtained through the following formula:

[0082] R = (1 - P e )·(1 - P c )·(1 - P t )·σ

[0083] Among them, R represents the resource linkage coefficient; P e , P c and P t represent the proportion of node operation energy consumption, CPU occupancy rate, and task processing time delay ratio; σ represents the medical task characteristic coefficient.

[0084] The working principle of the above technical solution is: extract the operating parameters of the edge node when the medical Internet of Things device cluster generates the medical task characteristic coefficient each time; among them, the operating parameters include the proportion of node operation energy consumption, CPU occupancy rate, and task processing time delay ratio; obtain the resource linkage coefficient between the edge node and the medical Internet of Things device cluster by using the operating parameters of the edge node; determine the load intensity degree of the edge node according to the relationship between the resource linkage coefficient and the preset linkage coefficient threshold. Specifically, compare the resource linkage coefficient with the preset linkage coefficient threshold; when the resource linkage coefficient is lower than the preset linkage coefficient threshold, it is determined that the load intensity of the edge node corresponding to the medical Internet of Things device cluster is large. Among them, the resource linkage coefficient is obtained through the following formula:

[0085] R = (1 - P e )·(1 - P c )·(1 - P t )·σ

[0086] Among them, R represents the resource linkage coefficient; P e , P c and P t represent the proportion of node operation energy consumption, CPU occupancy rate, and task processing time delay ratio; σ represents the medical task characteristic coefficient. Specifically, the operating efficiency of the above resource linkage coefficient (1 - P e )·(1 - Pc )·(1 - P t ) The larger the value, the lower the node energy consumption, CPU, and latency, and the higher the resource utilization efficiency. The larger the task characteristic coefficient σ value, the heavier the task load or the stronger the imbalance. At the same time, when the task load is large (σ is high) and the node can still maintain high efficiency (high operating efficiency), then R increases significantly, reflecting the strong collaboration between the edge node and the cluster. A high R value indicates that under high load or unbalanced tasks (σ is high), the edge node has low energy consumption, CPU, and latency, indicating extremely high collaboration efficiency and the system remains stable under pressure. When the R value is low, there may be two situations: a low-load scenario (σ is small), where the node has high efficiency (high operating efficiency), and the collaboration efficiency meets the operating requirements for the edge node to control the medical IoT device cluster. A high-load scenario (σ is high), where the node has low efficiency (low operating efficiency), indicating poor collaboration efficiency.

[0087] The effects of the above technical solution are as follows: Deeply obtain the resource consumption and task processing situation data of the edge node during operation. Reflect the operating state of the edge node from different perspectives, improving the comprehensiveness of understanding the operating state. Improve the accuracy of evaluating the collaboration efficiency, accurately reflecting the resource utilization efficiency and collaboration ability of the edge node under different task loads. Efficiently and quickly identify the edge nodes running at high load through a simple determination method, improving the system's identification efficiency. Improve the utilization sufficiency of the high-efficiency resource processing ability of the edge node, further optimize resource allocation, allocate more tasks to this edge node, and improve the processing ability and efficiency of the entire medical IoT system.

[0088] At the same time, by extracting operating parameters such as the proportion of node operating energy consumption, CPU occupancy rate, and task processing time delay ratio of the edge node, and combining with the medical task characteristic coefficient to calculate the resource linkage coefficient, and then determining the load intensity level, it can comprehensively and accurately evaluate the edge node load from multiple dimensions. Avoid the one-sidedness brought by evaluating the load based on a single indicator, helping the system accurately grasp the working state of the edge node, and providing a reliable basis for subsequent resource scheduling and system optimization. After accurately determining the edge node load intensity level, the system can reasonably allocate resources according to the load situation. For edge nodes with a large load intensity, timely take resource allocation measures, such as transferring some tasks to nodes with a lighter load, or increasing resource supply, so as to improve the resource utilization rate of the entire medical IoT system, reduce resource waste, ensure the efficient and stable operation of the system, and optimize the system performance indicators. Timely understanding of the edge node load intensity can discover potential system bottlenecks and performance risks in advance. Intervening in nodes with excessive load can effectively prevent system failures, task processing delays, etc. caused by node overload, ensure the continuity of medical services, and improve system performance indicators such as stability and reliability. And the proportion of node operating energy consumption P e 、CPU occupancy rate P c, the task processing time delay ratio P t and the medical task characteristic coefficient σ. The resource usage situation of the edge nodes and the characteristics of medical tasks are comprehensively considered, so that the calculated resource linkage coefficient can comprehensively reflect the resource association relationship between the edge nodes and the medical Internet of Things device cluster, providing more comprehensive and accurate data support for load intensity determination and improving the scientific nature of performance evaluation. Using the form of continuous multiplication of (1 - P e )(1 - P c )(1 - P t ) reflects that the operating parameters are not simply linearly related, but rather interact and act together. This calculation method is more in line with the complex coupling relationship of various factors in the actual system, can more truly reflect the internal connection between the resource status of the edge nodes and medical tasks, make the calculation result of the resource linkage coefficient closer to the actual operation situation, and then improve the accuracy of load intensity determination based on this coefficient and optimize the relevant indicators of system performance.

[0089] In an embodiment of the present invention, when the load intensity of the edge node is relatively large, dynamic adjustment is performed on the edge node corresponding to the medical Internet of Things device cluster, including:

[0090] When the load intensity of the edge node is relatively large, detect the resource fragmentation status of each edge node in the medical Internet of Things platform, and determine whether the resource fragmentation status reaches the screening condition; wherein, determining whether the resource fragmentation status reaches the screening condition includes: comparing the node resource coefficient corresponding to each edge node with a preset resource coefficient threshold; when the node resource coefficient is not lower than the preset resource coefficient threshold, it indicates that the edge node reaches the screening condition;

[0091] Screen out the edge nodes whose resource fragmentation status reaches the screening condition as candidate edge nodes;

[0092] Determine a target node from the candidate edge nodes, and use the target node to replace the edge node corresponding to the medical Internet of Things device cluster. Specifically, any one of the candidate edge nodes can be determined as the target node, or the candidate edge node corresponding to the maximum value of the node resource coefficient can be selected as the target node.

[0093] Or, obtain the target node according to the target node determination strategy, wherein the target node determination strategy is as follows:

[0094] Set the time window corresponding to each candidate edge node by using the node resource coefficient of each candidate edge node;

[0095] Wherein, the time window corresponding to each candidate edge node is obtained by the following formula:

[0096]

[0097] Among them, T represents the time length of the time window corresponding to each candidate edge node; T0 represents the preset initial window time length; S represents the node resource coefficient;

[0098] Collect the medical risk coefficients generated during the historical operation of each candidate edge node using the time window corresponding to each candidate edge node, and obtain the medical risk coefficient corresponding to each time window of each candidate edge node;

[0099] Obtain the risk parameter corresponding to each candidate edge node using the medical risk coefficient corresponding to each time window of each candidate edge node;

[0100] Among them, the risk parameter corresponding to each candidate edge node is obtained through the following formula

[0101]

[0102] Among them, G represents the risk parameter corresponding to each candidate edge node; g represents the number of data collections; B yi represents the available bandwidth rate of the candidate edge node corresponding to the i-th data collection; P di represents the delay jitter rate of the candidate edge node corresponding to the i-th data collection; D xi represents the packet loss rate of the candidate edge node corresponding to the i-th data collection; f i represents the medical risk coefficient of the candidate edge node corresponding to the i-th data collection;

[0103] Select the candidate edge node corresponding to the minimum risk parameter as the target node.

[0104] The working principle of the above technical solution is as follows: Determine the load intensity degree of the edge node through the determination method. When it is determined that the load intensity degree of a certain edge node is large, trigger the detection of the resource fragmentation state of each edge node in the medical Internet of Things platform. It is difficult to effectively integrate and utilize fragmented resource blocks, resulting in low resource utilization rate. It is necessary to detect the resource fragmentation state; evaluate the resource fragmentation state of each detected edge node. Screen out the edge nodes whose resource fragmentation state reaches the screening conditions as candidate edge nodes. Determine the target node from the candidate edge nodes. Use the determined target node to replace the edge node with a relatively large load intensity corresponding to the medical Internet of Things device cluster. During the replacement process, it is necessary to ensure the smooth migration of data and the seamless switching of services to ensure the normal operation of the medical Internet of Things system.

[0105] The effects of the above technical solution are as follows: By detecting the resource fragmentation status and screening out suitable candidate edge nodes, the resource utilization rate of edge nodes is improved. It realizes that the resource blocks that were originally idle or difficult to utilize due to resource fragmentation can be efficiently utilized. Replacing the edge nodes with high load and possible resource bottlenecks avoids task processing delays or failures caused by insufficient resources of a single node, and improves the resource utilization efficiency of the entire medical Internet of Things system. When the load intensity of edge nodes is relatively high, timely replacement can avoid system failures caused by node overload. The selection of target nodes undertakes the task processing requirements of the medical Internet of Things device cluster. It reduces performance fluctuations caused by resource fragmentation and improves the service stability of edge nodes. It ensures the overall stability of the medical Internet of Things system and improves the reliability of medical data transmission and processing.

[0106] On the other hand, when the load intensity of edge nodes is relatively high, by detecting the resource fragmentation status to screen candidate edge nodes and determining target nodes for replacement, the load can be effectively transferred from overloaded nodes to relatively idle nodes. This helps to balance the loads of edge nodes within the medical Internet of Things platform, avoids performance degradation of some nodes due to overload, improves the task processing capacity and efficiency of the entire system, and optimizes the performance index of system resource utilization efficiency. Candidate edge nodes are screened based on the node resource coefficient and resource fragmentation status, and nodes with relatively sufficient resources and low fragmentation degree are given priority. Using these nodes to replace high-load nodes can make the system resources be more reasonably configured, reduce resource waste, improve resource utilization rate, and thus improve the performance of the system in terms of resource usage. The target node determination strategy is adopted, and risk parameters are calculated by comprehensively considering factors such as the medical risk coefficient, available bandwidth rate, delay jitter rate, packet loss rate, etc. of candidate edge nodes, and the node with the smallest risk parameter is selected as the target node. This can effectively reduce the risks brought by node replacement, ensure the continuity and stability of medical services, reduce medical task interruptions or errors caused by node failures, network anomalies, etc., and improve performance indicators such as the reliability and stability of the system. This solution dynamically adjusts nodes according to the real-time load intensity, resource fragmentation status, and risk parameters of edge nodes. It enables the medical Internet of Things system to quickly respond to changes during operation, timely optimize resource allocation and node deployment, improve the adaptability of the system to different service scenarios and load changes, and ensure that the system performance is always at a relatively optimal level.

[0107] Meanwhile, the time window formula in the above technical solution Among them, the time window length T is related to the node resource coefficient S. The node resource coefficient reflects the resource status of the edge node. Different resource statuses result in dynamic changes in the time window length. When resources are sufficient (S is small), the time window is long, allowing for the collection of more historical data to comprehensively evaluate risks. When resources are scarce (S is large), the time window is short, focusing on recent data to quickly respond to resource changes, closely associating risk assessment with resource status, providing time-dimensional data that better suits the actual resource situation for target node determination, and enhancing the rationality and accuracy of target node determination. At the same time, the above formula uses a sine function to construct a non-linear mapping relationship, which can more delicately describe the relationship between the resource coefficient and the time window length. Compared with a linear relationship, this non-linear method can avoid the simple and crude impact of resource coefficient changes on the time window. According to the changes in the resource coefficient in different intervals, it can more accurately adjust the time window length, making the time window setting more in line with actual business requirements and resource characteristics, thereby improving the accuracy of risk assessment and target node determination based on time window data and optimizing system performance. And, the risk parameter formula incorporates multi-dimensional indicators such as bandwidth availability rate, delay jitter rate, packet loss rate, and medical risk coefficient. It comprehensively considers network transmission quality and medical service risk factors, avoiding the one-sidedness of evaluating risks based on a single indicator, and can more accurately reflect the risk status of candidate edge nodes during actual operation, providing a comprehensive and reliable risk assessment basis for target node determination, and enhancing the scientific nature and system stability of target node determination. At the same time, the formula weights and integrates each dimension indicator through a specific calculation method to reflect the comprehensive impact of different indicators on risks. For example, the higher the bandwidth availability rate, the lower the delay jitter rate and packet loss rate, the smaller the risk parameter may be. Combining with the medical risk coefficient balances network and medical service risks. This weighting method makes the calculation of the risk parameter more reasonable, can screen out the candidate edge node with the lowest comprehensive risk as the target node, reduce the probability of risk problems after node replacement, and ensure the business continuity and performance stability of the medical Internet of Things system.

[0108] In one embodiment of the present invention, detecting the resource fragmentation status of each edge node in the medical Internet of Things platform includes:

[0109] Retrieving the medical density coefficient of each edge node and the effective parameter of aggregable resources of each edge node;

[0110] Extracting the medical task characteristic coefficient of the medical Internet of Things device cluster corresponding to each edge node;

[0111] Obtaining the node resource coefficient corresponding to each edge node by using the medical density coefficient, the effective parameter of aggregable resources of each edge node, and the medical task characteristic coefficient of the medical Internet of Things device cluster corresponding to each edge node.

[0112] Moreover, determining whether the resource fragmentation state reaches the screening condition includes:

[0113] Comparing the node resource coefficient corresponding to each edge node with a preset resource coefficient threshold;

[0114] When the node resource coefficient is not lower than the preset resource coefficient threshold, it indicates that the edge node reaches the screening condition.

[0115] The working principle of the above technical solution is as follows: Detect the resource fragmentation state of each edge node in the medical Internet of Things platform, including:

[0116] Retrieving the medical density coefficient of each edge node and the effective parameter of the aggregable resources of each edge node;

[0117] Extracting the medical task feature coefficient of the medical Internet of Things device cluster corresponding to each edge node;

[0118] Using the medical density coefficient of each edge node, the effective parameter of the aggregable resources, and the medical task feature coefficient of the medical Internet of Things device cluster corresponding to each edge node to obtain the node resource coefficient corresponding to each edge node.

[0119] Among them, the node resource coefficient is used to evaluate the available strength of the edge node; moreover, the node resource coefficient is obtained through the following formula:

[0120]

[0121] Among them, S represents the node resource coefficient; M represents the medical density coefficient; U represents the effective parameter of the aggregable resources of each edge node;

[0122] Among them, the medical density coefficient is obtained through the following formula:

[0123]

[0124] Among them, C represents the number of medical tasks corresponding to the edge node; f represents the medical risk coefficient corresponding to the edge node; Y represents the number of executable task personnel of all medical devices corresponding to the edge node;

[0125] At the same time, the effective parameter of the aggregable resources of each edge node is obtained through the following formula:

[0126]

[0127] Among them, r represents the number of idle resource blocks corresponding to the edge node; size i represents the i-th idle resource block corresponding to the edge node; Z represents the total amount of resources corresponding to the medical Internet of Things.

[0128] The effects of the above technical solutions are as follows: By parameters such as the medical density coefficient, the effective parameters of aggregable resources, and the medical task characteristic coefficient, the resource status and available intensity of the edge nodes are evaluated, reflecting the capabilities of the edge nodes in actual operation. By calculating the node resource coefficient, the resource status of the edge nodes is quantified into a specific value, improving the compatibility of comparisons between different edge nodes. Based on the comparison between the node resource coefficient and the preset threshold, edge nodes that meet the requirements can be screened out. Edge nodes with certain advantages in terms of resource utilization and task processing capabilities are obtained, and are more suitable as candidate nodes for replacing edge nodes with high loads or severe resource fragmentation. Through the above screening and replacement strategies, the resource utilization rate of the entire medical Internet of Things system is improved. Idle and waste of resources are avoided. Selecting edge nodes with higher node resource coefficients for task processing can ensure that medical tasks are processed in a timely and effective manner. The above nodes have sufficient resources and capabilities to handle various medical tasks, reducing the risk of task processing delays or failures caused by insufficient resources. In high-load medical scenarios, by screening and replacing edge nodes, the resource allocation of the system can be adjusted in a timely manner to ensure the stable operation of the system. System failures caused by overloading of a single edge node are avoided.

[0129] At the same time, by comprehensively considering the medical density coefficient, the effective parameters of aggregatable resources, and the medical task characteristic coefficient to calculate the node resource coefficient, the resource status of the edge node can be comprehensively and accurately evaluated. The fusion analysis of multi-dimensional data avoids the limitations of single indicator evaluation, makes the detection of resource fragmentation status more accurate, and helps to timely discover potential unreasonable resource allocation problems. After accurately obtaining the node resource coefficient, it can be judged whether the edge node meets the screening conditions based on the comparison results with the preset resource coefficient threshold. This provides clear guidance for the resource optimization configuration of the medical Internet of Things platform. For edge nodes with severe resource fragmentation (high node resource coefficient), targeted resource adjustments can be made, such as reallocating tasks and supplementing resources, so as to improve the resource utilization of the entire platform and optimize system performance. Timely identification of resource fragmentation status and taking measures can effectively avoid the performance degradation of edge nodes caused by improper resource allocation, such as task processing delays, slow system response, etc. It ensures the stable operation of the medical Internet of Things platform, improves the timeliness and reliability of medical services, and thus improves the performance of the system in key performance indicators such as business processing capabilities and response time. In addition, the node resource coefficient formula involved in the above technical solution combines the medical density coefficient M and the effective parameter U of the aggregable resources, and takes into account the characteristics of the medical task (reflected by σ in the exponential term). This combination method comprehensively covers the business load of the edge node (medical density coefficient), the available resources (effective parameter of the aggregable resources) and the impact of the characteristics of the medical task on the resources, so that the calculated node resource coefficient can comprehensively reflect the available intensity of the resources of the edge node, and provide a comprehensive basis for the resource status assessment. The formula adopts nonlinear functions such as logarithmic functions and exponential functions to more realistically depict the complex interaction between the parameters. For example, the effective parameter U of the aggregable resources is in the logarithmic function, and the medical task characteristic coefficient σ is in the exponential function. This nonlinear expression can reflect that the resource factors in the actual scenario are not simply linearly related, so that the calculation result of the node resource coefficient is more in line with the actual resource status, and the accuracy of the resource fragmentation assessment is improved, thereby providing reliable data support for optimizing the system performance indicators. The calculation formula of the medical density coefficient M includes business-related parameters such as the number of medical tasks, medical risk coefficients, and the number of personnel who can perform tasks at the edge node, reflecting the relationship between business load and resources. The formula of the effective parameter U of the aggregatable resource takes into account resource indicators such as the number of idle resource blocks. In this way, the node resource coefficient formula can effectively reflect the relationship between edge node business demand and resource supply, which helps the system to reasonably allocate resources according to business conditions and improve performance indicators such as resource utilization efficiency.

[0130] The embodiment of the present invention proposes an edge computing resource scheduling system for medical Internet of Things devices, such as Figure 2 As shown, the edge computing resource scheduling system of the medical Internet of Things device includes:

[0131] A medical task feature coefficient acquisition module, which is used to obtain the cluster feature parameters corresponding to the medical Internet of Things device cluster in real time through the medical Internet of Things device cluster, and generate medical task feature coefficients by using the cluster feature parameters;

[0132] A load intensity determination module, which is used to determine the load intensity degree of the edge node corresponding to the medical Internet of Things device cluster according to the medical task feature coefficient;

[0133] A node adjustment module, which is used to dynamically adjust the edge node corresponding to the medical Internet of Things device cluster when the load intensity degree of the edge node is relatively large.

[0134] The working principle of the above technical solution is as follows: A medical Internet of Things device cluster is composed of multiple interconnected medical Internet of Things devices. Relevant data is obtained from the devices in real time through a data acquisition interface or protocol as cluster feature parameters. The obtained cluster feature parameters are analyzed and processed to generate medical task feature coefficients. The characteristics of the medical tasks undertaken by the current medical Internet of Things device cluster are reflected through the medical task feature coefficients. The load intensity degree of the current edge node is determined according to data such as the feature coefficient. When it is determined that the load intensity degree of the edge node is relatively large, the edge node is dynamically adjusted.

[0135] The effects of the above technical solution are as follows: By obtaining the cluster feature parameters in real time and generating medical task feature coefficients, the accuracy of obtaining the task requirements of the medical Internet of Things device cluster and the load conditions of the edge nodes can be improved. Dynamically adjusting the resource allocation of the edge node according to the load intensity degree can avoid the idle and waste of resources and improve the overall utilization rate of the edge computing resources. When the load intensity degree of the edge node is relatively large, timely dynamic adjustment can effectively avoid problems such as system crashes or performance degradation caused by resource overload. Through reasonable task migration and resource allocation, the loads of each edge node are kept relatively balanced, improving the stability and reliability of the entire medical Internet of Things system. The optimized resource scheduling method can ensure that medical tasks are processed in a timely manner and reduce task processing delays.

[0136] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these changes and modifications.

Claims

1. An edge computing resource scheduling method for a medical Internet of Things device, characterized in that, The edge computing resource scheduling method for the medical Internet of Things device includes: Obtaining the cluster characteristic parameters corresponding to the medical Internet of Things device cluster in real time through the medical Internet of Things device cluster, and generating a medical task characteristic coefficient by using the cluster characteristic parameters; Determining the load intensity degree of the edge node corresponding to the medical Internet of Things device cluster according to the medical task characteristic coefficient; When the load intensity degree of the edge node is large, dynamically adjusting the edge node corresponding to the medical Internet of Things device cluster.

2. The edge computing resource scheduling method for the medical Internet of Things device according to claim 1, wherein Obtaining the cluster characteristic parameters corresponding to the medical Internet of Things device cluster in real time through the medical Internet of Things device cluster, and generating a medical task characteristic coefficient by using the cluster characteristic parameters, including: Monitoring the operation parameters of each medical Internet of Things device included in each medical Internet of Things device cluster in real time; Obtaining the operation characteristic vector corresponding to each medical Internet of Things device by using the operation parameters of each medical Internet of Things device; Obtaining the medical task characteristic coefficient corresponding to the medical Internet of Things device cluster by using the norm corresponding to the operation characteristic vector of each medical Internet of Things device.

3. The edge computing resource scheduling method for the medical Internet of Things device according to claim 1, wherein, Determining the load intensity degree of the edge node corresponding to the medical Internet of Things device cluster according to the medical task characteristic coefficient, including: Extracting the operation parameters of the edge node when each medical Internet of Things device cluster generates the medical task characteristic coefficient; Obtaining the resource linkage coefficient between the edge node and the medical Internet of Things device cluster by using the operation parameters of the edge node; Determining the load intensity degree of the edge node according to the relationship between the resource linkage coefficient and the preset linkage coefficient threshold.

4. The edge computing resource scheduling method of the medical Internet of Things device according to claim 3, wherein, The resource linkage coefficient is obtained through the following formula: R = (1 - P e )·(1 - P c )·(1 - P t )·σ Among them, R represents the resource linkage coefficient; P e , P c and P t represent the proportion of node operation energy consumption, CPU occupancy rate, and task processing time delay ratio; σ represents the medical task characteristic coefficient.

5. The edge computing resource scheduling method for the medical Internet of Things device according to claim 3, wherein Determining the load intensity degree of the edge node according to the relationship between the resource linkage coefficient and the preset linkage coefficient threshold, including: Comparing the resource linkage coefficient with the preset linkage coefficient threshold; When the resource linkage coefficient is lower than the preset linkage coefficient threshold, it is determined that the load intensity of the edge node corresponding to the medical Internet of Things device cluster is large.

6. The edge computing resource scheduling method for the medical Internet of Things device according to claim 1, wherein, When the load intensity degree of the edge node is large, dynamically adjusting the edge node corresponding to the medical Internet of Things device cluster, including: When the load intensity degree of the edge node is large, detecting the resource fragmentation state of each edge node in the medical Internet of Things platform, and determining whether the resource fragmentation state reaches the screening condition; Screening out the edge nodes whose resource fragmentation state reaches the screening condition as candidate edge nodes; Determining a target node from the candidate edge nodes, and using the target node to replace the edge node corresponding to the medical Internet of Things device cluster.

7. The edge computing resource scheduling method of the medical Internet of Things device according to claim 6, characterized in that Detecting the resource fragmentation state of each edge node in the medical Internet of Things platform, including: Retrieving the medical density coefficient of each edge node and the effective parameter of the aggregable resource of each edge node; Extracting the medical task characteristic coefficient of the medical Internet of Things device cluster corresponding to each edge node; Obtaining the node resource coefficient corresponding to each edge node by using the medical density coefficient, the effective parameter of the aggregable resource of each edge node, and the medical task characteristic coefficient of the medical Internet of Things device cluster corresponding to each edge node.

8. The edge computing resource scheduling method for the medical Internet of Things device according to claim 6, characterized in that, Determining whether the resource fragmentation state reaches the screening condition, including: Compare the node resource coefficient corresponding to each edge node with a preset resource coefficient threshold; When the node resource coefficient is not lower than the preset resource coefficient threshold, it indicates that the edge node meets the screening conditions.

9. An edge computing resource scheduling system for a medical Internet of Things device, characterized in that, The edge computing resource scheduling system of the medical Internet of Things device includes: A medical task feature coefficient acquisition module, configured to obtain, in real time, cluster feature parameters corresponding to a medical Internet of Things device cluster through the medical Internet of Things device cluster, and generate a medical task feature coefficient by using the cluster feature parameters; A load intensity determination module, configured to determine the load intensity degree of an edge node corresponding to the medical Internet of Things device cluster according to the medical task feature coefficient; A node adjustment module, configured to dynamically adjust the edge node corresponding to the medical Internet of Things device cluster when the load intensity degree of the edge node is relatively large.

Images