A node switching and service migration method, device, equipment and storage medium

Abstract

The present invention discloses a node switching and service migration method, apparatus, equipment and storage medium. According to the currently monitored node computing power resource utilization rate, energy consumption and security risk level of a big data center, switching trigger monitoring is performed on the current node; when the switching is triggered, the service instance to be migrated is determined according to the load correlation between the current node and its virtual resources; before the node switching is performed, the target node is pre-migrated; during the node switching process, a switching mode decision is made according to the classification tag information of the target node to determine the target switching mode; according to the image pre-migration result, the service instance and the image file are synchronously migrated to the target node according to the target switching mode, and the service migration is optimized based on a batch strategy, and the service migration is divided into a pre-migration stage and a post-migration stage, so that during the service migration process, the service interruption time can be reduced, the continuity of the stateful service can be ensured, and the user experience can be improved.

Description

Technical Field

[0001] The present invention relates to the field of service migration technology, and in particular to a node switching and service migration method, apparatus, device and storage medium. Background Art

[0002] With the rapid deployment of computing networks, edge computing technology is gaining widespread adoption. Edge computing nodes can leverage computing resources from big data centers, providing close-range computing services to end users. At the same time, the overall demand and scale of computing resources are also growing to support the massive volume of data processing and computational requests, which in turn leads to significant energy consumption and security risks. Therefore, finding a more energy-efficient and secure way to use computing resources is a critical issue that needs to be addressed. Existing solutions primarily monitor computing resource usage and trigger a switch to the edge computing node connected to the end user. These solutions also migrate and synchronize application service data in virtual machines and containers to the new node, effectively scheduling and allocating computing resources to improve overall resource utilization. Existing service migration solutions primarily consider when edge computing nodes should be switched, for whom the switch should be performed, and how application services can be seamlessly migrated.

[0003] With the increasing popularity of interactive services, users often generate large amounts of state data while running applications. Therefore, the issue of synchronous and seamless migration is becoming increasingly important. However, existing service migration solutions often incur short downtimes during the migration process, resulting in service interruptions. Furthermore, while users can continue to use the original service after switching to the new node, they may lose their previous state, impacting their user experience. Summary of the Invention

[0004] In response to the problems existing in the existing technology, an embodiment of the present invention provides a node switching and service migration method, device, equipment and storage medium, which can reduce service interruption time during the migration process, ensure the continuity of stateful services, and improve user experience.

[0005] In a first aspect, an embodiment of the present invention provides a node switching and service migration method, including:

[0006] According to the currently monitored node computing resource utilization, energy consumption of the big data center, and security risk level, the current node is triggered to monitor the switch;

[0007] When it is detected that the current node triggers a switch, the service instance to be migrated in the virtual resource is determined according to the load correlation between the current node and its virtual resource;

[0008] Before performing node switching, pre-migrating the virtual resource image to the target node; wherein the target node is determined based on the energy consumption and security risk level of the big data center;

[0009] During the node switching process, a switching mode decision is made based on the classification tag information of the target node to determine the target switching mode;

[0010] According to the image pre-migration result, the service instance and the image file are synchronously migrated to the target node according to the target switching mode.

[0011] As an improvement to the above solution, the target switching mode is a composite switching mode;

[0012] Among them, in the composite switching mode, transition switching is performed for the target node whose classification mark information is a positive node, and intermittent switching is performed for the target node whose classification mark information is a negative node; wherein, the classification mark information is determined based on the delay of the target node in performing transition switching in the past.

[0013] As an improvement to the above solution, the mirror pre-migration of the virtual resources on the target node includes:

[0014] According to the user access information of the virtual resource, a prediction is made through a pre-established prediction model to determine the target virtual resource;

[0015] Pre-migrate the image file corresponding to the target virtual resource to the target node.

[0016] As an improvement to the above solution, the step of synchronously migrating the service instance and the image file corresponding to the service to the target node according to the determined switching mode based on the image pre-migration result includes:

[0017] According to the image pre-migration result, determining whether the target node has deployed an image file corresponding to the service instance;

[0018] If so, synchronously migrate the corresponding service instance to the target node according to the switching mode;

[0019] If not, the corresponding service instance and its image file are synchronously migrated to the target node according to the switching mode.

[0020] As an improvement to the above solution, the switching trigger monitoring of the current node based on the currently monitored node computing resource utilization, energy consumption and security risk level of the big data center includes:

[0021] Calculate the comprehensive performance index of the big data center associated with the current node based on the average energy consumption and security risk level of the big data center within a set time period;

[0022] Determine a dynamic switching threshold based on the comprehensive performance indicator, the maximum node computing resource utilization rate among the node computing resource utilization rates of the current node monitored in real time, and a preset upper limit of the center energy consumption tolerance;

[0023] According to the dynamic switching threshold and the node computing resource utilization rate, switching trigger monitoring is performed on the current node.

[0024] As an improvement to the above solution, the determination of the target node includes the following steps:

[0025] Calculate the comprehensive performance index of the big data center associated with the current node based on the average energy consumption and security risk level of the big data center within a set time period;

[0026] Determine the performance status of the corresponding big data center based on the comprehensive performance indicators of the big data center associated with the current node;

[0027] When the performance state of the corresponding big data center is the first performance state, determining a target node in the corresponding big data center;

[0028] When the performance state of the corresponding big data center is the second performance state, switch to other data centers in the first performance state, and determine the target node in the other big data centers.

[0029] As an improvement to the above solution, determining the service instance to be migrated from the virtual resource according to the load correlation between the current node and its virtual resource includes:

[0030] Calculating the load correlation between the current node and its virtual resource based on the historical load record information of the virtual resource and the historical load record information of the current node;

[0031] determining a service instance to be migrated from the virtual resources according to the load correlation;

[0032] Calculate the load evaluation index of the corresponding service instance based on the load value of the computing resources occupied by the service corresponding to the service instance to be migrated;

[0033] The migration order of the service instances to be migrated is determined according to the load evaluation index.

[0034] In a second aspect, an embodiment of the present invention provides a node switching and service migration device, including:

[0035] The switching monitoring module is used to trigger the switching of the current node based on the currently monitored node computing resource utilization, energy consumption and security risk level of the big data center;

[0036] A switching decision module is used to determine the service instance to be migrated in the virtual resource based on the load correlation between the current node and its virtual resource when monitoring the current node triggering the switching;

[0037] An image pre-migration module is used to pre-migrate the image of the virtual resources to the target node before performing node switching; wherein the target node is determined based on the energy consumption and security risk level of the big data center;

[0038] A mode decision module is used to make a switching mode decision according to the classification tag information of the target node during the node switching process to determine the target switching mode;

[0039] The service migration module is used to synchronously migrate the service instance and the image file to the target node according to the image pre-migration result and the target switching mode.

[0040] In the third aspect, an embodiment of the present invention provides a node switching and service migration device, comprising: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements the node switching and service migration method as described in any one of the first aspects.

[0041] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, which stores a computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the node switching and service migration method as described in any one of the first aspects.

[0042] Compared with the prior art, a node switching and service migration method, apparatus, device and storage medium of an embodiment of the present invention, the medium, performs switching trigger monitoring on the current node through the currently monitored node computing resource utilization rate, energy consumption and security risk level of the big data center to determine whether the current node triggers switching; when it is monitored that the current node triggers switching, first determine the service instance to be migrated in the virtual resource according to the load correlation between the current node and its virtual resource, and determine the target node according to the energy consumption and security risk level of the big data center; then, before performing the node switching, perform mirror pre-migration of the virtual resource on the target node; thereafter, during the node switching process, make a switching mode decision according to the classification tag information of the target node, and determine the target node Switch mode; according to the result of the above-mentioned image pre-migration, the service instance and the image file are synchronously migrated to the target node according to the target switching mode; on the one hand, the present invention deploys relevant image files in advance on the target node in the pre-migration stage (i.e. before the node switching), effectively reducing the amount of data required to transmit during service migration when the node switches, and reducing the service interruption duration; on the other hand, in the post-migration stage (i.e. during the node switching process), the classification tag information of the target node is used to make a mode decision, and the best target switching mode is selected to migrate user state data such as service instances and image files, so as to achieve a dynamic balance between the computing power resource utilization and the service interruption duration during the switching process, effectively ensuring the continuity of the user's stateful service, and achieving synchronous and seamless migration, thereby improving the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solution of the present invention, the following is a brief introduction to the drawings used in the implementation methods. Obviously, the drawings described below are only some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0044] Figure 1 This is a flow chart of a node switching and service migration method provided by an embodiment of the present invention;

[0045] Figure 2 This is a schematic diagram of the architecture of the node switching and service migration process provided by an embodiment of the present invention;

[0046] Figure 3 is a flowchart of a node switching trigger monitoring process provided by an embodiment of the present invention;

[0047] Figure 4 This is a flowchart of a process for confirming a node migration service instance provided by an embodiment of the present invention;

[0048] Figure 5 is a flowchart of a target node determination process provided by an embodiment of the present invention;

[0049] Figure 6 This is a schematic diagram of application service migration across large data centers provided by an embodiment of the present invention;

[0050] Figure 7 This is a schematic diagram of a stateful service migration process based on mode decision-making provided by an embodiment of the present invention;

[0051] Figure 8 This is a schematic diagram of the target node classification and marking process provided by an embodiment of the present invention;

[0052] Figure 9 This is a schematic diagram of the node switching interaction process provided by an embodiment of the present invention;

[0053] Figure 10 This is a structural block diagram of a node switching and service migration device provided by an embodiment of the present invention;

[0054] Figure 11 This is a structural block diagram of a node switching and service migration device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0055] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0056] The following is an explanation of some terminology concepts involved in the embodiments of the present invention.

[0057] Mobile Edge Computing (MEC): Edge computing refers to providing the nearest service close to the source of objects or data using an open platform that integrates network, computing, storage, and application core capabilities. It is a technology that deeply integrates mobile access networks with Internet services by sinking computing and processing capabilities to the edge closest to the service.

[0058] Edge computing nodes: Edge computing nodes are a logical abstraction of the fundamental common capabilities of various edge-side product forms, such as edge gateways, edge controllers, and edge servers. These product forms provide common capabilities such as real-time edge data analysis, local data storage, and real-time network connectivity. In the embodiments of the present invention, edge computing nodes are also referred to as "MEC nodes" or "nodes."

[0059] Big data center: refers to a large data center close to the edge of the network or close to end users and devices, used to provide functions such as caching content, cloud computing resources and analysis.

[0060] Service: An application that provides specific functions to users. In edge computing scenarios, it usually runs on edge servers based on virtual machines or containers and generates user service instances.

[0061] Virtual Machine (VM): is a computing virtualization technology. A virtual machine comes with an operating system and is a hardware-level virtualization isolation device. It is mainly suitable for deploying applications or modules that have high requirements for system isolation and security.

[0062] Container: A computing virtualization technology that shares the operating system kernel with the host and can provide independent user domains, file systems, and run spaces for applications or modules through the isolation capabilities provided by the operating system kernel. It has higher mobility and deployment density.

[0063] Service instance: Service data generated based on the application image file based on the user's service request. The service instance corresponding to each user is different.

[0064] Service migration: The process of switching a user's service instance from the originally connected edge computing node to the newly connected edge computing node.

[0065] Service continuity: After a node switch occurs, it can quickly recover and connect to the previous user operation status without affecting the smoothness of the user experience, ensuring that the user is unaware of the switching process.

[0066] Transitional switching: Data from running application service instances (including memory data and execution status) is copied and transferred to the target node. During the transfer process, the running application service generates some status data updates, necessitating repeated transmission of updated data. When a predefined metric (such as the number of iterations) reaches a certain threshold, the application service on the source node is stopped, the remaining data is sent to the target node, and service is restored.

[0067] Intermittent switching method: directly stop the running application service instance in the source node, then send the application service instance data to the target node to restore the service, and send the remaining status data to the target node through on-demand access and other methods.

[0068] See also Figure 1 , Figure 1 1 is a flow chart of a node switching and service migration method provided by an embodiment of the present invention. The node switching and service migration method specifically includes:

[0069] S1: Based on the currently monitored node computing resource utilization, energy consumption of the big data center, and security risk level, trigger the monitoring of the current node switching;

[0070] S2: When it is detected that the current node triggers a switch, the service instance to be migrated in the virtual resource is determined based on the load correlation between the current node and its virtual resource;

[0071] S3: Before performing node switching, pre-migrate the virtual resource image to the target node; wherein the target node is determined based on the energy consumption and security risk level of the big data center;

[0072] S4: During the node switching process, a switching mode decision is made according to the classification tag information of the target node to determine the target switching mode;

[0073] S5: According to the image pre-migration result, the service instance and the image file are synchronously migrated to the target node according to the target switching mode.

[0074] It should be noted that the node switching and service migration method described in the embodiment of the present invention is executed by the current node; wherein the current node is the above-mentioned edge computing node, it is understandable that the execution subject of the method of the embodiment of the present invention (i.e., the current node) can be a computing service device with edge service migration, network communication and program running functions, such as a mobile phone, tablet computer, personal computer, etc., and can also be other electronic devices that implement the same or similar functions. The virtual resources described in the embodiment of the present invention include but are not limited to virtual machines, containers, and other applications or modules built using computing virtualization technology.

[0075] For example, see Figure 2 The node switching and service migration method described in the embodiment of the present invention mainly includes three parts and is executed by three modules: a switching monitoring module, a switching decision module, and a switching execution module. The details are as follows:

[0076] The switching monitoring module is used to monitor node resources and big data center performance; specifically, it monitors the node computing resource utilization rate of the current node itself, the energy consumption and security risk level of the big data center associated with the current node in real time; and regularly collects the node computing resource utilization rate of the current node itself, the energy consumption and security risk level of the big data center associated with the current node to make a switching trigger judgment on the current node, and judge whether the current node triggers switching at the current moment.

[0077] Among them, the node computing power resource utilization rate is determined based on the various computing power resource operation data of the current node. The computing power resources of the current node include but are not limited to: the computing power that can be provided by various processors such as the central processing unit (CPU), graphics processing unit (GPU), tensor processing unit (TPU), field-programmable gate array (FPGA), application-specific integrated circuit (ASIC), digital signal processor (DSP), I / O, etc. In the embodiment of the present invention, the computing power resource utilization rate of the corresponding computing power resources is used as the computing power resource operation data. Assume that the current node has a total of n types of computing power resources, and the corresponding various types of computing power resource utilization rates are R1, R2, ..., R n ; Among them, n can be selected and set according to actual needs and is not specifically limited in the embodiments of the invention; then the node computing power resource utilization rate of the current node is: R = {R1, R2, ..., R n}.

[0078] The energy consumption of a big data center is equal to the sum of the energy consumption of all physical nodes in the corresponding big data center. The energy consumption of each physical node is related to the number of processors, number of cores, memory size, and other hardware and software indicators. A node energy consumption measurement model can be established by quantifying related factors, and the energy consumption of node c at time t is recorded as e c (t). Let C be the set of all nodes in the big data center. Then the energy consumption per unit time of the big data center in the set time period from t1 to t2 is calculated as:

[0079]

[0080] The security risk level of a big data center is equal to the average security risk level of all physical nodes in the corresponding big data center. The security risk level of each physical node is related to its open interface, storage resources, and sensitivity. A node risk measurement model can be established by quantifying related factors, and the security risk level of node c at time t is s c (t). Assume that the number of elements in the set of all nodes in the big data center is crad(C), then the formula for calculating the average security risk level of the big data center in the set time period from t1 to t2 is:

[0081]

[0082] Alternatively, the risk level of all physical nodes in the corresponding big data center may be weighted averaged based on the importance of different physical nodes to obtain the security risk level of the big data center.

[0083] The embodiment of the present invention dynamically monitors the node computing resource utilization rate of the node, the energy consumption and security risk level of the big data center, and then conducts a quantitative analysis of the energy consumption and security risk level of the big data center, establishes a node energy consumption measurement model and a node risk measurement model, comprehensively considers the energy consumption and security of the big data center, and triggers switching monitoring of the corresponding node based on the computing resource operation data of the node and the energy consumption and security risk level of the big data center to determine whether the corresponding node triggers switching at the current moment; if so, the node switching is triggered and the next step is executed to find a new target node for connection and migrate the corresponding service; otherwise, the computing resource operation data of the corresponding node, the energy consumption and security risk level of the big data center are continued to be monitored. By dynamically monitoring the node computing resource utilization rate of the node and obtaining the energy consumption and security risk level of the big data center in real time, the node switching monitoring is performed, and the resources in the computing network are reasonably scheduled and utilized, so that the node switching can be triggered in time, effectively solving the problem of "when to execute the switch".

[0084] For example, a fixed threshold is set for the node computing resource utilization rate, the energy consumption of the big data center, and the security risk level of the big data center, respectively. The method judges whether any component of the node computing resource utilization rate is greater than the corresponding fixed threshold, whether the energy consumption of the big data center is greater than the corresponding fixed threshold, and whether the security risk level of the big data center is greater than the corresponding fixed threshold. If all three conditions are met, the switching is triggered, otherwise the switching is not triggered. Alternatively, a dynamic threshold is determined based on the node computing resource utilization rate, the energy consumption of the big data center, and the security risk level of the big data center using specific mathematical operations, and then judges whether any component of the node computing resource utilization rate is greater than the dynamic threshold. If so, the switching is triggered, otherwise the switching is not triggered.

[0085] Further, see Figure 3 Step S1: According to the currently monitored node computing resource utilization, energy consumption and security risk level of the big data center, the current node is triggered to monitor the switch, including:

[0086] S11: Calculate the comprehensive performance index of the big data center associated with the current node based on the average energy consumption and security risk level of the big data center within a set time period;

[0087] S12: Determine a dynamic switching threshold based on the comprehensive performance indicator, the maximum node computing resource utilization rate among the node computing resource utilization rates of the current nodes monitored in real time, and a preset upper limit of the center energy consumption tolerance;

[0088] S13: Perform switching trigger monitoring on the current node according to the dynamic switching threshold and the node computing resource utilization rate.

[0089] For example, the energy consumption and security risk level of the big data center are quantitatively analyzed by the above formulas (1) and (2), and the energy consumption E per unit time and the average security risk level S of the big data center in the set time period from t1 to t2 are calculated. Then, the energy consumption and the average security risk level of the big data center in the set time period are weighted and summed to obtain the comprehensive performance index of the big data center. For example, if the energy consumption E and the average security risk level S are assigned weight factors k1 and k2 respectively, the calculation formula of the comprehensive performance index P of the big data center is:

[0090] P = k1E + k2S (3);

[0091] Among them, st0 <k1<1,0<k2<1,k1+k2=1。

[0092] The larger the value of the comprehensive performance index P of the big data center, the worse the comprehensive performance of the big data center.

[0093] Set the upper limit of the central energy consumption tolerance of the corresponding big data center to P M , remember R i max is the maximum node computing resource utilization of the type i computing resources owned by the current node, that is, the node computing resource utilization of the current node R = {R1, R2, ..., R n}, the maximum node computing power resource utilization, the adaptive dynamic switching threshold components R i ′(1≤i≤n)The specific calculation formula is:

[0094] R i ′=(1-P / P M )*R i max (4);

[0095] Then, the dynamic switching threshold R′={R1′, R2′, ..., R n ′}, by taking the node computing resource utilization of the current node R={R1,R2,...,R n} and the dynamic switching threshold R′={R1′,R2′,...,R n When any component of the node computing resource utilization R is greater than the corresponding component of R', the current node is determined to trigger switching.

[0096] The embodiment of the present invention monitors node resources and big data center performance, establishes a big data center comprehensive performance calculation method based on energy consumption and security risk measurement, and determines the dynamic threshold for triggering node switching based on the comprehensive performance. The switching threshold can be adaptively and dynamically adjusted according to the comprehensive performance changes of the big data center. Compared with the existing technology that uses static thresholds for node switching judgment, the embodiment of the present invention uses the above-mentioned adaptive dynamic switching threshold, which can make the triggering of node switching more reasonable, and help reduce the energy consumption and security risks of the big data center system.

[0097] A switching decision module is used to determine the service instances to be migrated from the virtual resources and their migration order based on the load correlation between the current node and its virtual resources when the current node triggers a switch; and is also used to determine the migration target node corresponding to the above-determined service to be migrated based on the energy consumption and security risk level of the big data center;

[0098] The services (i.e., computing tasks) carried by the nodes are encapsulated in virtual resources such as virtual machines or containers. The higher the correlation between the load of virtual resources such as virtual machines or containers and the node load, the greater the impact of the corresponding virtual resources on the overload of edge node resources. The embodiment of the present invention screens services based on the load correlation between nodes and virtual resources, and determines the service instances corresponding to the services that need to be migrated. This not only reduces the migration cost, but also allows the node to return to normal load status more quickly, thereby optimizing overall resource utilization.

[0099] Further, see Figure 4 Step S2: Determine the service instance to be migrated from the virtual resource based on the load correlation between the current node and its virtual resource, including:

[0100] S21: Calculating the load correlation between the current node and its virtual resource based on the historical load record information of the virtual resource and the historical load record information of the current node;

[0101] S22: Determine a service instance to be migrated from the virtual resources according to the load correlation;

[0102] S23: Calculate the load evaluation index of the corresponding service instance based on the load value of the computing resources occupied by the service corresponding to the service instance to be migrated;

[0103] S24: Determine the migration order of the service instances to be migrated according to the load evaluation index.

[0104] For example, first calculate the load correlation r between virtual resources such as virtual machines or containers and nodes based on the historical load records of virtual resources and nodes. uv, where u represents the historical load record of virtual resources, and v represents the historical load record of nodes. The historical load record of virtual resources such as virtual machines or containers is recorded as {u1, u2, ..., u k}, the historical load record of the node is {v1, v2, ..., v k}, where v k Indicates the load record when the node runs the service for the kth time; u k It indicates the load record of the corresponding virtual resource when the node runs the corresponding service for the kth time. It should be noted that the embodiment of the present invention does not specifically limit the calculation method of the load correlation between virtual resources and nodes. Users can customize the selection according to their needs. For example, the correlation coefficient algorithm can be calculated using the Pearson correlation coefficient method, the Selpiman correlation coefficient method, and the Chebyshev correlation coefficient method. According to the obtained load correlation r between virtual resources and nodes uv , we can filter out the service instances corresponding to the services in the virtual resources such as virtual machines and containers that are strongly correlated, and optimally migrate out the service instances to obtain the service instance set. For example, the absolute value of the correlation between the virtual resource and the node load ‖r uv ‖ is compared with the preset strong correlation threshold value. When the absolute value of the load correlation ‖r uv If ‖ is greater than the strong correlation threshold, it is determined that the node is strongly correlated with the virtual resource when running the service corresponding to the corresponding load, and the service instance corresponding to the corresponding service is selected as the service instance to be migrated.

[0105] For each service instance to be migrated out in the filtered service instance set, further migration sorting is required. First, the load evaluation index of each service instance to be migrated out in the service instance set is calculated, and L is recorded. j is the load evaluation index of the jth service instance in the service set, L ij is the load value of the i-th type of computing resources occupied by the service corresponding to the j-th service instance, α i is the impact factor of the i-th type of computing resources, then the load evaluation index of the j-th service instance is:

[0106]

[0107] It is understandable that if the load evaluation index of a service instance is higher, the computing power resources occupied by the corresponding service will be higher, and it needs to be migrated out first. Therefore, the service instances in the set of service instances to be migrated are arranged in descending order according to the load evaluation index to determine the migration order. Then, in subsequent service migrations, they are migrated out in sequence according to the migration order until the original node returns to normal load level.

[0108] When selecting a target node, it's important to consider whether to search across large data centers and which node within the large data center to select as the target node. This embodiment of the present invention first determines whether to search across large data centers based on the energy consumption and security risk level of the large data center associated with the current node, and then determines the target node through a two-round screening process.

[0109] Further, if Figure 5 As shown, the determination of the target node includes the following steps:

[0110] S100: Calculating a comprehensive performance index of the big data center associated with the current node based on the average energy consumption and security risk level of the big data center within a set time period;

[0111] S200: Determine the performance status of the corresponding big data center based on the comprehensive performance indicators of the big data center associated with the current node;

[0112] S300: When the performance state of the corresponding big data center is a first performance state, determining a target node in the corresponding big data center; wherein the first performance state is an excellent performance state or a medium performance state;

[0113] S400: When the performance state of the corresponding big data center is a second performance state, switching to another data center in the first performance state, and determining a target node in the other big data center; wherein the second performance state is a low performance state;

[0114] For example, regarding the question of whether to search for the target node across large data centers, we determine whether it is necessary to search for the target node across large data centers based on the energy consumption and comprehensive security performance of the large data center associated with the current node. The calculation of the comprehensive performance index of the large data center associated with the current node in step S100 is referred to step S11 above and will not be repeated here. First, based on the pre-set upper limit of the central energy consumption tolerance of the data center as P M Divide into multiple performance evaluation intervals, where different performance evaluation intervals correspond to a performance state, and then determine the performance evaluation interval to which the comprehensive performance index P of the big data center calculated in step S100 belongs, so as to determine the performance state of the big data center. Set the comprehensive performance index P to be less than P M / 3, the comprehensive performance status of the big data center is the best performance status, and the comprehensive performance index P is P M / 3 and 2P M / 3, the comprehensive performance status of the big data center is medium performance, and the comprehensive performance index P is greater than 2P M / 3, the comprehensive performance status of the big data center is low performance. Based on the above settings, when the performance status of the corresponding big data center is in the excellent performance state or the medium performance state, the target node is determined in the corresponding big data center; when the performance status of the corresponding big data center is low performance, switch to other data centers in the excellent performance state or the medium performance state, and determine the target node in the other big data center to cut in and perform service migration, such as Figure 6 The diagram shows a cross-data center migration diagram of an application service (i.e., a service described in an embodiment of the present invention). When a large data center has high energy consumption or low security, the embodiment of the present invention searches for target nodes across large data centers, effectively optimizing the matching of computing resources.

[0115] After determining whether to perform node switching and service migration across large data centers, it is necessary to further confirm the target node; first, perform a first round of preliminary screening to filter out nodes whose network, computing, storage and other capabilities do not meet the basic requirements. On this basis, a second round of fine screening is performed to comprehensively consider factors such as the target node's energy consumption, security risk level, and user service quality, and select the optimal target node for switching and service migration. The embodiment of the present invention determines the strongly correlated service instances to be migrated based on the load correlation between the current node and its virtual resources, and determines the ranking of the service instances to be migrated based on the load evaluation index, which can quickly restore the node to a normal load state, reduce migration costs, optimize overall resource utilization, and effectively achieve performance improvement in all aspects.

[0116] A switching execution module, wherein the switching execution module includes an image pre-migration module, a mode decision module, and a service migration module;

[0117] The image pre-migration module is used to perform image pre-migration of the virtual resources on the target node before performing node switching.

[0118] The mode decision module is used to make a switching mode decision based on the classification tag information of the target node and determine the target switching mode. Each node is provided with a target node marker, which is used to mark and update its own classification tag in real time, such as marking it as a positive node or a negative node. For example, it can be determined based on the delay when the corresponding node switched nodes in the past, or it can also be based on the current network, computing, storage and other capabilities, energy consumption, security risk level, etc. of the corresponding node. According to the classification tag information returned by the target node, the current node adopts a targeted target switching mode for it in the subsequent node switching process. For example, a switching mode decider is set, and the best target switching mode is selected among the transitional switching mode, intermittent switching mode and compound switching mode in combination with user demand preferences. This can effectively ensure the continuity and stability of stateful services, protect user status data from being lost, and achieve that users are basically unaware after the node switch. For example Figure 7 As shown, the switching mode decision maker can decide a switching mode from the transition switching mode, the intermittent switching mode, and the compound switching mode according to the classification tag information of the target node and the user's demand preference, and perform stateful service migration to complete the switching; wherein, the transition switching mode is to perform transition switching for target nodes whose classification tag information is a positive node or a negative node; the intermittent switching mode is to perform intermittent switching for target nodes whose classification tag information is a positive node or a negative node; the compound switching mode is to perform transition switching for target nodes whose classification tag information is a positive node, and to perform intermittent switching for target nodes whose classification tag information is a negative node.

[0119] It should be noted that the user's demand preference selection can be determined based on the mode selection operation of the user on the corresponding user interface when the current node triggers the switch, or based on the user's pre-set preference switching mode stored in the current node, or selecting the switching mode that has been used most times in the previous switching modes of the current node as the target mode.

[0120] The service migration module is used to synchronously migrate the service instance and the image file to the target node according to the image pre-migration result and the target switching mode.

[0121] The present invention adopts an application service migration optimization method based on a batch strategy, which divides service migration into a pre-migration stage and a post-migration stage. Through this batch migration method; on the one hand, in the pre-migration stage (i.e., before the node switching), relevant image files are deployed in advance on the target node, which effectively reduces the amount of data required to transmit during service migration when the node switches, reduces the service interruption duration, and speeds up the service migration speed; on the other hand, in the post-migration stage (i.e., during the node switching process), the classification tag information of the target node is used to make a mode decision, and the best target switching mode is selected to migrate user status data such as service instances and image files, which can effectively ensure the stable connection of user usage status after the node switching, realize the dynamic balance of computing power resource utilization and service interruption duration during the switching process, effectively ensure the continuity of user stateful services, truly achieve synchronous and seamless migration, improve user experience, and at the same time realize reasonable and optimized allocation of computing power resources, and reduce energy consumption and security risks of large data center systems.

[0122] Furthermore, the target switching mode is a composite switching mode;

[0123] Among them, in the composite switching mode, transition switching is performed for the target node whose classification mark information is a positive node, and intermittent switching is performed for the target node whose classification mark information is a negative node; wherein, the classification mark information is determined based on the delay of the target node in performing transition switching in the past.

[0124] For example, see Figure 8 , the classification labeling of the target node includes the following steps:

[0125] Measurement phase: Measure the delay from the moment the target node establishes a connection to the moment the current node disconnects during the transitional switching interaction between the nodes.

[0126] Marking phase: Target nodes are marked as either positive or negative based on the measured latency. For example, if the latency exceeds a preset upper limit, the target node is marked as negative; otherwise, it is marked as positive. The upper limit can be set based on actual conditions and is not specified here.

[0127] Maintenance phase: Maintain and dynamically update the classification tag information of the target node.

[0128] Based on the classification tag information of the target node, a targeted switching mode can be adopted in the subsequent execution of node switching. When migrating user service instances, the use of transitional switching will cause a waste of computing resources due to the simultaneous occupation of two nodes, while intermittent switching will result in a relatively long interruption time. In an embodiment of the present invention, a composite switching mode is preferably selected, and the target nodes are first classified and marked. In the post-migration stage, transitional switching is performed for target nodes with positive classification tag information, and intermittent switching is performed for target nodes with negative classification tag information, thereby performing intermittent switching on negative nodes that generate a large waste of computing resources to release extra computing resources and reduce resource consumption, while performing transitional switching on positive nodes to achieve a dynamic balance between effective resource utilization and control of interruption time.

[0129] Specifically, performing mirror pre-migration of the virtual resources on the target node includes:

[0130] According to the user access information of the virtual resource, a prediction is made through a pre-established prediction model to determine the target virtual resource;

[0131] Pre-migrate the image file corresponding to the target virtual resource to the target node.

[0132] For example, for the virtual resources corresponding to the service instances that need to be migrated, their user access information (including but not limited to user access volume, number of user accesses, cumulative duration of user accesses, date of user accesses, etc.) is used as the input of a pre-built prediction model. The prediction model is used to make predictions to determine the target virtual resources in the corresponding virtual resources that need to pre-migrate the image files in advance, and the image files of the target virtual resources such as virtual machines and containers with large user access volume are deployed in advance to the relevant target nodes. In actual edge computing scenarios, user activities usually show certain regularities over time and space, which affects the demand for application services provided by the nodes. For example, users in rush hour will have an increased demand for travel taxis and map navigation applications, and users in industrial parks will have a greater demand for applications such as data analysis and image processing. The embodiment of the present invention can effectively reduce the time required for subsequent migration by deploying the image files of virtual resources such as virtual machines and containers with large user access volume to the relevant target nodes in advance. Among them, the prediction model is obtained by training a machine learning model constructed based on heuristic algorithms, meta-heuristic algorithms and machine learning algorithms. The network structure and training method of the machine learning model belong to the prior art and will not be repeated here.

[0133] Specifically, the step of synchronously migrating the service instance and the image file corresponding to the service to the target node according to the image pre-migration result and the determined switching mode includes:

[0134] According to the image pre-migration result, determining whether the target node has deployed an image file corresponding to the service instance;

[0135] If so, synchronously migrate the corresponding service instance to the target node according to the switching mode;

[0136] If not, the corresponding service instance and its image file are synchronously migrated to the target node according to the switching mode.

[0137] For example, in the post-migration phase, a composite switching mode is used for node switching, and at the same time, a determination is made as to whether the target node has an image file. If the target node has deployed the image file corresponding to the service instance in the pre-migration phase, only the user's service instance needs to be migrated, otherwise both need to be migrated.

[0138] See also Figure 9 ,The node switching interaction process includes:

[0139] The current node sends a node switching notification to the user;

[0140] The current node pre-migrates the image to the target node, and deploys the image files of virtual resources such as virtual machines or containers in advance on the target node based on the prediction results output by the prediction model;

[0141] The current node sends a mode switching confirmation notification to the user;

[0142] After receiving the switch mode confirmation notification, the user confirms the switch mode operation, returns the switch mode confirmation information to the current node, and connects to the switch mode decision maker;

[0143] The current node queries the classification tag information of the target node;

[0144] The current node obtains the classification label information of the target node and connects to the switching mode decision maker;

[0145] The current node performs service migration to the target node according to the target switching mode determined;

[0146] The target node provides services to the user, and the user resumes the running state before the switch;

[0147] The current node releases computing resources.

[0148] Compared with the prior art, the embodiments of the present invention have the following advantages:

[0149] (1) An application service migration optimization method based on a batch strategy is adopted, and the service migration is divided into a pre-migration stage and a post-migration stage. Through this batch migration method, on the one hand, in the pre-migration stage (i.e., before the node switch), the relevant image files are deployed in advance on the target node, which effectively reduces the amount of data required for service migration during the node switch, reduces the service interruption time, and speeds up the service migration speed; on the other hand, in the post-migration stage (i.e., during the node switch), the classification tag information of the target node is used to make a mode decision, and the best target switching mode is selected to migrate user state data such as service instances and image files, which can effectively ensure the stable connection of user usage status after the node switch, achieve a dynamic balance between the computing power resource utilization and the service interruption time during the switching process, effectively ensure the continuity of user stateful services, improve user experience, and realize the reasonable optimization and allocation of computing power resources, reducing the energy consumption and security risks of the big data center system. The embodiment of the present invention can effectively solve the problems of when the node performs the switch, for whom the switch is performed, and how the application service is synchronized and seamlessly migrated in the current computing power network scenario, and provide a high-efficiency, energy-saving, safe and reliable technical solution for node switching and service migration.

[0150] (2) Based on the comprehensive performance changes of the big data center, the dynamic switching threshold used to trigger node switching is adaptively and dynamically adjusted. Compared with the prior art that uses a static threshold to make node switching judgments, the embodiment of the present invention uses the above-mentioned adaptive dynamic switching threshold, which can make the triggering of node switching more reasonable and help reduce the energy consumption and security risks of the big data center system.

[0151] (3) The embodiment of the present invention determines the set of strongly correlated service instances to be migrated and the method for sorting them based on the load correlation between the current node to be migrated and its virtual resources such as virtual machines or containers and the load value of the computing resources occupied by the services corresponding to the service instances to be migrated. This can quickly restore the nodes to a normal load state, reduce migration costs, optimize overall resource utilization, and effectively achieve performance improvements in all aspects. At the same time, it fully considers the problem of finding target nodes across big data centers when the energy consumption of big data centers is high or the security performance is low, and can effectively achieve optimized matching of computing resources.

[0152] (4) The embodiment of the present invention adopts a composite node switching execution method based on node classification tag information. By adopting a transitional switching method for positive nodes that generate little waste of computing resources, and an intermittent switching method for negative nodes that generate a large waste of computing resources, a dynamic balance between computing resource utilization and service interruption duration is achieved during the switching process.

[0153] See also Figure 10 , Figure 10 The embodiment of the present invention provides a structural block diagram of a node switching and service migration device, wherein the node switching and service migration device includes:

[0154] Switching monitoring module 1 is used to trigger switching monitoring of the current node based on the currently monitored node computing resource utilization, energy consumption and security risk level of the big data center;

[0155] Switching decision module 2, for determining the service instance to be migrated in the virtual resource according to the load correlation between the current node and its virtual resource when monitoring the current node triggering the switch;

[0156] The image pre-migration module 3 is used to pre-migrate the image of the virtual resources to the target node before the node switching; wherein the target node is determined based on the energy consumption and security risk level of the big data center;

[0157] Mode decision module 4, used for making a switching mode decision according to the classification tag information of the target node during the node switching process, and determining the target switching mode;

[0158] The service migration module 5 is configured to synchronously migrate the service instance and the image file to the target node according to the image pre-migration result and the target switching mode.

[0159] In an optional embodiment, the target switching mode is a composite switching mode;

[0160] Among them, in the composite switching mode, transition switching is performed for the target node whose classification mark information is a positive node, and intermittent switching is performed for the target node whose classification mark information is a negative node; wherein, the classification mark information is determined based on the delay of the target node in performing transition switching in the past.

[0161] In an optional embodiment, the image pre-migration module 2 includes:

[0162] a target virtual resource determining unit, configured to determine the target virtual resource by predicting the user access information of the virtual resource using a pre-established prediction model;

[0163] The image file migration unit is used to pre-migrate the image file corresponding to the target virtual resource to the target node.

[0164] In an optional embodiment, the service migration module 4 includes:

[0165] An image file determination unit, configured to determine, based on the image pre-migration result, whether the target node has deployed therein an image file corresponding to the service instance;

[0166] a first synchronous migration unit, configured to, if yes, synchronously migrate the corresponding service instance to the target node according to the switching mode;

[0167] The second synchronization unit is configured to, if not, synchronously migrate the corresponding service instance and its mirror file to the target node according to the switching mode.

[0168] In an optional embodiment, the handover monitoring module 1 includes:

[0169] A performance index calculation unit, configured to calculate a comprehensive performance index of the big data center associated with the current node based on the average energy consumption and security risk level of the big data center within a set time period;

[0170] A dynamic switching threshold determination unit is configured to determine a dynamic switching threshold based on the comprehensive performance indicator, the maximum node computing resource utilization rate among the node computing resource utilization rates of the current node monitored in real time, and a preset upper limit of the center energy consumption tolerance;

[0171] The node switching trigger monitoring unit is used to perform switching trigger monitoring on the current node according to the dynamic switching threshold and the node computing power resource utilization rate.

[0172] In an optional embodiment, the device further includes:

[0173] A center performance status determination unit is configured to determine the performance status of the corresponding big data center based on the comprehensive performance index of the big data center associated with the current node; wherein the comprehensive performance index of the big data center is calculated based on the average energy consumption and security risk level of the big data center associated with the current node within a set time period;

[0174] a first target node determining unit, configured to determine a target node in the corresponding big data center when the performance state of the corresponding big data center is a first performance state;

[0175] The second target node determination unit is used to switch to other data centers in the first performance state when the performance state of the corresponding big data center is the second performance state, and determine the target node in the other big data centers.

[0176] In an optional embodiment, the handover decision module 2 includes:

[0177] A load correlation calculation unit, configured to calculate the load correlation between the current node and its virtual resource based on the historical load record information of the virtual resource and the historical load record information of the current node;

[0178] a service instance determining unit, configured to determine a service instance to be migrated from the virtual resources according to the load correlation;

[0179] A load evaluation index calculation unit, configured to calculate a load evaluation index of a corresponding service instance based on a load value of computing resources occupied by a service corresponding to the service instance to be migrated;

[0180] The service instance sorting unit is used to determine the migration order of the service instances to be migrated according to the load evaluation index.

[0181] It should be noted that the working process of each module in the node switching and service migration device described in the embodiment of the present invention can refer to the working process of the node switching and service migration method described in the above embodiment, and the technical effect achieved is also the same as the node switching and service migration device described in the above embodiment, which will not be repeated here.

[0182] See also Figure 11 , Figure 11is a block diagram of the structure of a node switching and service migration device provided in an embodiment of the present invention. The node switching and service migration device includes a processor 21, a memory 22, and a computer program stored in the memory 22 and executable on the processor 21. When the processor 21 executes the computer program, it implements the steps of each of the above-mentioned node switching and service migration method embodiments, such as steps S1 to S5.

[0183] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory 22 and executed by the processor 21 to implement the present invention. The one or more modules / units may be a series of computer program instruction segments capable of implementing specific functions, and the instruction segments are used to describe the execution process of the computer program in the node switching and service migration device.

[0184] The node switching and service migration device may include, but is not limited to, a processor 21 and a memory 22. Those skilled in the art will appreciate that the schematic diagram is merely an example of the node switching and service migration device and does not limit the node switching and service migration device. The device may include more or fewer components than shown in the diagram, or may combine certain components or different components. For example, the node switching and service migration device may also include input and output devices, network access devices, buses, and the like.

[0185] The processor 21 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor. The processor 21 is the control center of the node switching and service migration device, and utilizes various interfaces and lines to connect various parts of the entire node switching and service migration device.

[0186] The memory 22 can be used to store the computer programs and / or modules. The processor 21 implements the various functions of the node switching and service migration device by running or executing the computer programs and / or modules stored in the memory 22, and calling the data stored in the memory 22. The memory 22 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function (such as a sound playback function, an image playback function, etc.); the data storage area can store data created based on the use of the mobile phone (such as audio data, a phone book, etc.). In addition, the memory 22 can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage device.

[0187] Wherein, if the module / unit integrated with the node switching and service migration device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by the processor 21, it can implement the steps of the above-mentioned various method embodiments. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device that can carry the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0188] It should be noted that the device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed across multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment. In addition, in the drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art can understand and implement the present invention without inventive effort.

[0189] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, many improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A node switching and service migration method, characterized in that: include: According to the currently monitored node computing resource utilization, energy consumption of the big data center, and security risk level, the current node is triggered to monitor the switch; When it is detected that the current node triggers a switch, the service instance to be migrated in the virtual resource is determined according to the load correlation between the current node and its virtual resource; Before performing node switching, pre-migrating the virtual resource image to the target node; wherein the target node is determined based on the energy consumption and security risk level of the big data center; During the node switching process, a switching mode decision is made based on the classification tag information of the target node to determine the target switching mode; the classification tag information includes a classification tag marked as a positive node or a negative node; the switching mode includes a transition switching mode, an intermittent switching mode, and a compound switching mode, wherein the transition switching mode is to perform transition switching on all target nodes whose classification tag information is a positive node or a negative node; the intermittent switching mode is to perform intermittent switching on all target nodes whose classification tag information is a positive node or a negative node; the compound switching mode is to perform transition switching on target nodes whose classification tag information is a positive node, and to perform intermittent switching on target nodes whose classification tag information is a negative node; when the delay of the node switching exceeds the preset delay upper limit, the corresponding target node is marked as a negative node, otherwise it is marked as a positive node; According to the image pre-migration result, the service instance and the image file are synchronously migrated to the target node according to the target switching mode.

2. The node switching and service migration method according to claim 1, wherein: The target switching mode is a composite switching mode; Among them, in the composite switching mode, transition switching is performed for the target node whose classification mark information is a positive node, and intermittent switching is performed for the target node whose classification mark information is a negative node; wherein, the classification mark information is determined based on the delay of the target node in performing transition switching in the past.

3. The node switching and service migration method according to claim 1, wherein: The mirror pre-migration of the virtual resources on the target node includes: According to the user access information of the virtual resource, a prediction is made through a pre-established prediction model to determine the target virtual resource; Pre-migrate the image file corresponding to the target virtual resource to the target node.

4. The node switching and service migration method according to claim 1, wherein: The step of synchronously migrating the service instance and the image file corresponding to the service to the target node according to the image pre-migration result and the determined switching mode includes: According to the image pre-migration result, determining whether the target node has deployed an image file corresponding to the service instance; If so, synchronously migrate the corresponding service instance to the target node according to the switching mode; If not, the corresponding service instance and its image file are synchronously migrated to the target node according to the switching mode.

5. The node switching and service migration method according to claim 1, wherein: The switching trigger monitoring of the current node according to the currently monitored node computing resource utilization, energy consumption and security risk level of the big data center includes: Calculate the comprehensive performance index of the big data center associated with the current node based on the average energy consumption and security risk level of the big data center within a set time period; Determine a dynamic switching threshold based on the comprehensive performance indicator, the maximum node computing resource utilization rate among the node computing resource utilization rates of the current node monitored in real time, and a preset upper limit of the center energy consumption tolerance; According to the dynamic switching threshold and the node computing resource utilization rate, switching trigger monitoring is performed on the current node.

6. The node switching and service migration method according to claim 1, wherein: The determination of the target node comprises the following steps: Calculate the comprehensive performance index of the big data center associated with the current node based on the average energy consumption and security risk level of the big data center within a set time period; Determine the performance status of the corresponding big data center based on the comprehensive performance indicators of the big data center associated with the current node; When the performance state of the corresponding big data center is the first performance state, determining a target node in the corresponding big data center; When the performance state of the corresponding big data center is the second performance state, switch to other data centers in the first performance state, and determine the target node in the other big data centers.

7. The node switching and service migration method according to claim 1, wherein: The determining, based on the load correlation between the current node and its virtual resource, the service instance to be migrated from the virtual resource includes: Calculating the load correlation between the current node and its virtual resource based on the historical load record information of the virtual resource and the historical load record information of the current node; determining a service instance to be migrated from the virtual resources according to the load correlation; Calculate the load evaluation index of the corresponding service instance based on the load value of the computing resources occupied by the service corresponding to the service instance to be migrated; The migration order of the service instances to be migrated is determined according to the load evaluation index.

8. A node switching and service migration device, characterized in that: include: The switching monitoring module is used to trigger the switching of the current node based on the currently monitored node computing resource utilization, energy consumption and security risk level of the big data center; A switching decision module is used to determine the service instance to be migrated in the virtual resource based on the load correlation between the current node and its virtual resource when monitoring the current node triggering the switching; An image pre-migration module is used to pre-migrate the image of the virtual resources to the target node before performing node switching; wherein the target node is determined based on the energy consumption and security risk level of the big data center; A mode decision module is used to make a switching mode decision and determine a target switching mode according to the classification tag information of the target node during the node switching process; the classification tag information includes a classification tag marked as a positive node or a negative node; the switching mode includes a transition switching mode, an intermittent switching mode, and a compound switching mode, wherein the transition switching mode is to perform transition switching on all target nodes whose classification tag information is a positive node or a negative node; the intermittent switching mode is to perform intermittent switching on all target nodes whose classification tag information is a positive node or a negative node; the compound switching mode is to perform transition switching on target nodes whose classification tag information is a positive node, and to perform intermittent switching on target nodes whose classification tag information is a negative node; when the delay of the node switching exceeds the preset delay upper limit, the corresponding target node is marked as a negative node, otherwise it is marked as a positive node; The service migration module is used to synchronously migrate the service instance and the image file to the target node according to the image pre-migration result and the target switching mode.

9. A node switching and service migration device, characterized in that: include: A processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the node switching and service migration method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the node switching and service migration method according to any one of claims 1 to 7.

Images