A method and system for elastic resource scheduling under a cloud computing platform

Abstract

This invention discloses a resource elastic scheduling method and system for a cloud computing platform, relating to the field of resource scheduling technology. The method includes the following steps: setting up a cluster of multiple virtual machines and establishing a shared space between adjacent virtual machines. This invention expands or shrinks the target virtual machine by using the shared space and demand load, ensuring the stability and continuity of task operation within the cluster and reducing the negative impacts of resource contention and service interruptions caused by changes in virtual machines within the cluster. Through the setting of a dependency symbiosis model, task data does not need to be directly migrated during data migration, reducing transmission interruptions, data loss, and inaccuracies during task data migration. Furthermore, migration rules facilitate load balancing between two virtual machines within the same shared space, ensuring the stability and continuity of task operation.

Description

Technical Field

[0001] This invention relates to the field of resource scheduling technology, and specifically to a resource elastic scheduling method and system under a cloud computing platform. Background Technology

[0002] A cloud computing platform is an internet-based computing service model that uses virtualization technology to centralize computing resources (such as servers, storage, networks, and applications) into a shared resource pool. Users can acquire and use these resources on demand without having to build and maintain their own physical infrastructure. Elastic scaling is a core feature of cloud computing platforms. It requires the platform to automatically and quickly adjust computing and storage resources based on dynamic changes in application load to achieve efficient resource utilization and ensure service level agreements (SLAs). Traditional resource elastic scheduling methods typically rely on adding or removing virtual machine instances at the cluster level to achieve horizontal scaling.

[0003] However, while the aforementioned traditional methods can overcome the limitations of single-machine resources through horizontal scaling, introducing or removing virtual machines not only introduces risks such as startup delays and service interruptions, but may also trigger short-term resource contention between virtual machines within the cluster due to resource reallocation, thus disrupting the overall stability of the cluster. Furthermore, during resource adjustments, especially when migrating task data between different virtual machines, it is usually necessary to migrate the entire task data directly. When the task data is large or the network fluctuates, this migration method is prone to transmission interruptions, data loss, or data inconsistencies, thereby affecting the continuity and accuracy of tasks. Moreover, if a virtual machine fails and task data is lost, the recovery process is time-consuming, making it difficult to ensure the stability and continuity of task operation. Summary of the Invention

[0004] The purpose of this invention is to provide a resource elastic scheduling method and system under a cloud computing platform to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a resource elastic scheduling method under a cloud computing platform, comprising the following steps:

[0006] Set up a cluster consisting of multiple virtual machines and establish a shared space between two adjacent virtual machines;

[0007] Monitor the load demand of tasks in the cluster and scale up or down the virtual machines according to the load demand.

[0008] A dependency symbiosis model is established based on the shared space. Data migration is carried out between corresponding virtual machines through the dependency symbiosis model until the preset conditions are met.

[0009] In a preferred embodiment, the step of setting up a cluster of multiple virtual machines and establishing a shared space between two adjacent virtual machines includes:

[0010] In a cloud computing platform, a cluster of multiple virtual machines is created based on the task.

[0011] Configure the topology of multiple virtual machines in the cluster, and establish a shared space between every two adjacent virtual machines in the cluster according to the topology; the same shared space belongs to the two virtual machines connected to it.

[0012] In a preferred embodiment, the step of monitoring the load demand of tasks in the cluster and scaling up or down the virtual machines based on the load demand includes:

[0013] Monitor the load requirements of tasks in the cluster in real time using monitoring tools;

[0014] Determine the target virtual machine and its remaining load based on the task's required load;

[0015] The remaining load of the target virtual machine is compared with the required load, and the target virtual machine is scaled up or down based on the comparison results.

[0016] In a preferred embodiment, the step of comparing the remaining load of the target virtual machine with the required load, and scaling up or down the target virtual machine based on the comparison result, includes:

[0017] Obtain the difference between the remaining load and the required load of the target virtual machine to get the comparison result;

[0018] If the comparison result is zero, then there is no need to expand or shrink the target virtual machine.

[0019] If the comparison result is greater than zero, the target virtual machine is scaled down based on the comparison result;

[0020] If the comparison result is less than zero, the target virtual machine will be expanded according to the comparison result.

[0021] In a preferred embodiment, the steps of establishing a dependency symbiosis model based on a shared space include:

[0022] Multiple conversion points and a database are deployed within a shared space. Multiple storage areas are divided within the virtual machines connected to the shared space. The storage areas are storage nodes used to store task data, and the conversion points store the data encoding of the dynamic data corresponding to the storage areas.

[0023] The database stores multiple basic graphs, each with a corresponding connection node. The basic graphs communicate with the conversion points through the connection nodes. The basic graphs are arrays formed by multiple points. Each point is an encoding obtained by mapping the metadata of the initial task data in the corresponding storage area. The metadata corresponding to each point is different. There is a metadata-encoding mapping conversion table between the metadata and the points.

[0024] A monitoring point is deployed next to each storage area; the monitoring point is a communication node used to monitor task data within the storage area, and the conversion point is connected to the storage area one-to-one through the monitoring point.

[0025] A dependency symbiosis model is established based on a conversion point, a connection node, a basic graph, a monitoring point, and a storage area; wherein, the dependency symbiosis model is a migration framework for migrating task data in the storage area through the conversion point.

[0026] In a preferred embodiment, the step of obtaining the data code corresponding to the storage area stored within the conversion point includes:

[0027] A monitoring cycle is set, and the monitoring points capture task data in the storage area based on the monitoring cycle. The task data is transformed into task graphs through a mapping transformation table and array diagram.

[0028] The task graph is compared point by point with the base graph connecting the storage area to identify all changed points, point positions and modification types, and all changed points, point positions and modification types are treated as dynamic data.

[0029] Dynamic data is serialized into data encoding and stored in the conversion point connected to the storage area; wherein, the serialization conversion includes: converting the dynamic data into binary according to a preset structure and compressing it.

[0030] In a preferred embodiment, the step of migrating data between corresponding virtual machines using a dependency symbiosis model until a preset condition is met includes:

[0031] Preset migration rules are used to determine the data to be migrated when the migration rules are met. Based on the data to be migrated, the migration start point and migration end point are determined. The migration start point and migration end point are the storage area where the data to be migrated is located and the destination storage area where the data to be migrated needs to be migrated, respectively. The migration rule is: when the difference in demand load between two virtual machines connected in the same shared space is greater than a preset threshold, the virtual machine corresponding to the high demand load will be migrated to the virtual machine corresponding to the low demand load.

[0032] Based on the data to be migrated, determine the conversion point to connect the migration start point, and establish a communication connection between the conversion point and the migration end point;

[0033] The data encoding within the conversion point is merged with the basic graph of the connection, and after merging, it is mapped to the migration endpoint through a mapping conversion table. The migration endpoint stores the migrated data.

[0034] The data code generated by comparing the migrated data with the base graph is used as the first verification code, and the data code generated by comparing the data to be migrated with the base graph is used as the second verification code. The first verification code and the second verification code are checked for consistency. When the first verification code and the second verification code are consistent, the data to be migrated is deleted. At the same time, the conversion point is disconnected from the storage area where the data to be migrated is located.

[0035] After the data migration is completed, it is determined whether the two virtual machines that were migrated meet the migration rules. If not, the data migration continues until the preset conditions are met.

[0036] In a preferred embodiment

[0037] The present invention also provides a resource elastic scheduling system under a cloud computing platform, including a setting layer, a monitoring and adjustment layer and a data migration layer;

[0038] The configuration layer is used to configure a cluster of multiple virtual machines and establish a shared space between two adjacent virtual machines.

[0039] The monitoring and adjustment layer is used to monitor the demand load of tasks in the cluster and to scale up or down the virtual machines according to the demand load.

[0040] The data migration layer is used to establish a dependency symbiosis model based on the shared space, and to perform data migration between corresponding virtual machines through the dependency symbiosis model until the preset conditions are met.

[0041] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0042] This invention expands or shrinks the target virtual machine by sharing space and demand load, thereby ensuring the stability and continuity of task operation within the cluster. At the same time, it does not require adding or removing virtual machines, which can reduce the negative impacts of resource competition and service interruption caused by changes in virtual machines in the cluster.

[0043] This invention, by establishing a symbiotic model, eliminates the need for direct migration of task data during data migration, reducing transmission interruptions, data loss, and inaccuracies. Furthermore, by encoding and recording dynamic data between the task graph and the base graph corresponding to the task data, it enables rapid recovery of lost task data within the virtual machine through merging of dynamic data and the base graph, enhancing the security of task data within the virtual machine. Finally, by setting migration rules, it facilitates load balancing between two virtual machines within the same shared space, ensuring the stability and continuity of task execution, preventing overload or resource idleness in a single virtual machine, and improving the overall resource utilization of the cluster. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0045] Figure 1 This is a flowchart of the method of the present invention.

[0046] Figure 2 This is a combined view of the points and the basic graphics in the method of this invention.

[0047] Figure 3 This is a system block diagram of the present invention. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0049] Example 1, please refer to Figure 1 and Figure 2 As shown in this embodiment, a resource elastic scheduling method under a cloud computing platform includes the following steps:

[0050] Set up a cluster consisting of multiple virtual machines and establish a shared space between two adjacent virtual machines;

[0051] Monitor the load demand of tasks in the cluster and scale up or down the virtual machines according to the load demand.

[0052] A dependency symbiosis model is established based on the shared space. Data migration is carried out between corresponding virtual machines through the dependency symbiosis model until the preset conditions are met.

[0053] In this embodiment, the target virtual machine is expanded or shrunk by sharing space and demand load, thereby ensuring the stability and continuity of task operation within the cluster. This eliminates the need to add or remove virtual machines, reducing the negative impacts of resource contention and service interruptions caused by changes in virtual machines within the cluster. By setting up a dependency symbiosis model, task data does not need to be directly migrated during data migration, reducing transmission interruptions, data loss, and inaccuracies during migration. Furthermore, by recording dynamic data between the task graph and the base graph corresponding to the task data through data encoding, task data can be quickly recovered by merging dynamic data with the base graph when it is lost within the virtual machine, making the task data within the virtual machine more secure. By setting migration rules, load balancing between two virtual machines within the same shared space can be achieved, ensuring the stability and continuity of task operation, preventing overload or idle resources of a single virtual machine, and improving the overall resource utilization of the cluster.

[0054] In one embodiment, the steps of setting up a cluster of multiple virtual machines and establishing a shared space between two adjacent virtual machines include:

[0055] In a cloud computing platform, a cluster of multiple virtual machines is created based on the task.

[0056] Configure the topology of multiple virtual machines in the cluster, and establish a shared space between every two adjacent virtual machines in the cluster according to the topology. The shared space is a logical node that is connected to two adjacent virtual machines and has communication and storage functions. The same shared space belongs to the two virtual machines it is connected to.

[0057] It's important to note that in a cloud computing platform, multiple virtual machines (VMs) are created on a physical server using the platform's virtualization management tools (such as OpenStack and VMware vSphere). A cluster of VMs is then configured based on task requirements (e.g., storage space, RAM, performance-related needs). For example, a data analysis task would require VMs with large memory. A topology is defined within the cluster, based on business logic or data interaction frequency. For instance, VMs with frequent interactions are placed adjacent to each other to reduce communication latency. A shared space is established between every two adjacent VMs in the cluster, with its memory size configured according to the anticipated tasks. Each shared space belongs to the two VMs it connects to, meaning both adjacent VMs can access and use this shared space. After establishing the shared space, the entire cluster is configured and tested to ensure that VMs can communicate and interact normally through the shared space, and to verify that the shared space's performance meets the requirements of certain tasks. Furthermore, by establishing a shared space, the two virtual machines can be adjusted (expanded or reduced) when the workload of a task increases or decreases, so as to avoid cluster instability caused by adding or removing virtual machines and mitigate the negative impact of adding or removing virtual machines in the cluster (such as short-term resource contention caused by adding virtual machines and temporary resource overload triggered by removing virtual machines).

[0058] In one embodiment, the step of monitoring the demand load of tasks in the cluster and scaling up or down virtual machines based on the demand load includes:

[0059] Monitor the load requirements of tasks in the cluster in real time using monitoring tools;

[0060] Determine the target virtual machine and its remaining load based on the task's required load;

[0061] The remaining load of the target virtual machine is compared with the required load, and the target virtual machine is scaled up or down based on the comparison results.

[0062] In one embodiment, the step of comparing the remaining load of the target virtual machine with the required load, and scaling up or down the target virtual machine based on the comparison result, includes:

[0063] Obtain the difference between the remaining load and the required load of the target virtual machine to get the comparison result;

[0064] If the comparison result is zero, then there is no need to expand or shrink the target virtual machine.

[0065] If the comparison result is greater than zero, the target virtual machine is scaled down based on the comparison result;

[0066] If the comparison result is less than zero, the target virtual machine will be expanded according to the comparison result;

[0067] It should be noted that the cluster's required load for corresponding tasks is collected in real time using monitoring tools such as Zabbix and Prometheus. The target virtual machine is the virtual machine corresponding to the required load of the task (the same or related to the task). The remaining load of the target virtual machine is obtained by subtracting the currently used resources from the initial configured resources of the virtual machine. If the comparison result is less than zero, the target virtual machine is expanded by allocating some memory from the shared space to the target virtual machine. If the comparison result is greater than zero, the target virtual machine is shrunk by allocating some memory from the target virtual machine to the shared space. If the comparison result is equal to zero... If the target virtual machine does not require adjustment, then no adjustment is needed. For example, if virtual machine A (the target virtual machine) has a total memory capacity of 64GB and 30GB is already used, the remaining workload of virtual machine A is calculated to be 34GB of memory. Comparing the remaining workload of 34GB with the required workload (e.g., the task's required workload is 40GB of memory), a difference of -6GB is obtained. Since the comparison result is less than zero, virtual machine A is expanded. Virtual machine A expands its memory by adjusting the shared space, and the total memory capacity of the shared space decreases from the original 14GB to 8GB, allowing virtual machine A to obtain 6GB of memory through the adjustment of the shared space, thus achieving the expansion effect. Furthermore, by expanding or shrinking the target virtual machine through shared space and required workload (i.e., elastic scheduling), the stability and continuity of task operation within the cluster can be guaranteed, while eliminating the need to add or remove virtual machines, thereby reducing the negative impacts of resource contention and service interruptions caused by changes in virtual machines in the cluster.

[0068] In one embodiment, the steps for establishing a dependency symbiosis model based on a shared space include:

[0069] Multiple conversion points and a database are deployed within a shared space. Multiple storage areas are divided within the virtual machines connected to the shared space. The storage areas are storage nodes used to store task data, and the conversion points store the data encoding of the dynamic data corresponding to the storage areas. The conversion points are also communication nodes.

[0070] The database stores multiple basic graphs, each with a corresponding connection node. The basic graphs communicate with the conversion points through the connection nodes. The basic graphs are arrays formed by multiple points. Each point is an encoding obtained by mapping the metadata of the initial task data in the corresponding storage area. The metadata corresponding to each point is different. There is a metadata-encoding mapping conversion table between the metadata and the points.

[0071] A monitoring point is deployed next to each storage area; the monitoring point is a communication node used to monitor task data within the storage area, and the conversion point is connected to the storage area one-to-one through the monitoring point.

[0072] A dependency symbiosis model is established based on a conversion point, a connection node, a basic graph, a monitoring point, and a storage area; wherein, the dependency symbiosis model is a migration framework for migrating task data in the storage area through the conversion point.

[0073] In one embodiment, the step of obtaining the data encoding corresponding to the storage area within the conversion point includes:

[0074] A monitoring cycle is set, and the monitoring points capture task data in the storage area based on the monitoring cycle. The task data is transformed into task graphs through a mapping transformation table and array diagram.

[0075] The task graph is compared point by point with the base graph connecting the storage area to identify all changed points, point positions and modification types (add / delete / remove operation information, etc.), and all changed points, point positions and modification types are treated as dynamic data.

[0076] Dynamic data is serialized into data encoding and stored in the conversion point connected to the storage area; wherein, the serialization conversion includes: converting the dynamic data into binary according to a preset structure and compressing it.

[0077] In one embodiment, the step of migrating data between corresponding virtual machines using a dependency symbiosis model until a preset condition is met includes:

[0078] Preset migration rules are used to determine the data to be migrated when the migration rules are met. Based on the data to be migrated, the migration start point and migration end point are determined. The migration start point and migration end point are the storage area where the data to be migrated is located and the destination storage area where the data to be migrated needs to be migrated, respectively. The migration rule is: when the difference in demand load between two virtual machines connected in the same shared space is greater than a preset threshold, the virtual machine corresponding to the high demand load will be migrated to the virtual machine corresponding to the low demand load.

[0079] Based on the data to be migrated, determine the conversion point to connect the migration start point, and establish a communication connection between the conversion point and the migration end point;

[0080] The data encoding within the conversion point is merged with the basic graph of the connection, and after merging, it is mapped to the migration endpoint through a mapping conversion table. The migration endpoint stores the migrated data.

[0081] The data code generated by comparing the migrated data with the base graph is used as the first verification code, and the data code generated by comparing the data to be migrated with the base graph is used as the second verification code. The first verification code and the second verification code are checked for consistency. When the first verification code and the second verification code are consistent, the data to be migrated is deleted. At the same time, the conversion point is disconnected from the storage area where the data to be migrated is located.

[0082] After the data migration is completed, it is determined whether the two virtual machines that were migrated meet the migration rules. If not, the data migration continues until the preset conditions are met.

[0083] It should be noted that the data to be migrated consists of task data within the storage area of ​​the two virtual machines, which is most conducive to achieving load balancing between them. By establishing a transformation point and database in the shared space, and creating a dependency symbiotic model among the transformation point, connection nodes, basic graph, monitoring point, and storage area, the data encoding within the transformation point and the basic graph of the connection can be merged during data migration of task data in the storage area within the virtual machine (the data encoding undergoes serialization conversion during merging). After merging, the data is mapped to the migration endpoint through a mapping transformation table, thus obtaining the complete task data in the storage area. After the migration is completed, the consistency of the first and second verification codes is verified, ensuring the accuracy of the data migration and reducing the risk of data loss. The reduced data volume in the storage system allows for seamless migration of task data from one virtual machine to another without direct data migration, minimizing transmission interruptions, data loss, and inaccuracies during migration. Furthermore, by encoding and recording dynamic data between the task graph and the base graph, task data can be quickly recovered in case of loss within the virtual machine through merging of dynamic data and the base graph, enhancing the security of task data within the virtual machine. Finally, the configuration of migration rules facilitates load balancing between two virtual machines within the same shared space, ensuring the stability and continuity of task execution, preventing overload or resource idleness in a single virtual machine, and improving the overall resource utilization of the cluster.

[0084] Example 2, please refer to Figure 3 As shown in this embodiment, a resource elastic scheduling system under a cloud computing platform includes a setting layer, a monitoring and adjustment layer, and a data migration layer.

[0085] The configuration layer is used to configure a cluster of multiple virtual machines and establish a shared space between two adjacent virtual machines.

[0086] The monitoring and adjustment layer is used to monitor the demand load of tasks in the cluster and to scale up or down the virtual machines according to the demand load.

[0087] The data migration layer is used to establish a dependency symbiosis model based on the shared space, and to perform data migration between corresponding virtual machines through the dependency symbiosis model until the preset conditions are met.

[0088] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A resource elastic scheduling method under a cloud computing platform, characterized in that, Includes the following steps: Set up a cluster consisting of multiple virtual machines and establish a shared space between two adjacent virtual machines; Monitor the workload demand of tasks in the cluster, and expand or shrink the virtual machine according to the workload demand. Expanding means allocating part of the memory in the shared space to the target virtual machine, and shrinking means allocating part of the memory in the target virtual machine to the shared space. A dependency symbiosis model is established based on the shared space. Data migration is carried out between corresponding virtual machines through the dependency symbiosis model until the preset conditions are met. The steps for establishing a dependency symbiosis model based on a shared space include: Multiple conversion points and a database are deployed within a shared space. Multiple storage areas are divided within the virtual machines connected to the shared space. The storage areas are storage nodes used to store task data, and the conversion points store the data encoding of the dynamic data corresponding to the storage areas. The database stores multiple basic graphs, each with a corresponding connection node. The basic graphs communicate with the conversion points through the connection nodes. The basic graphs are arrays formed by multiple points. Each point is an encoding obtained by mapping the metadata of the initial task data in the corresponding storage area. The metadata corresponding to each point is different. There is a metadata-encoding mapping conversion table between the metadata and the points. A monitoring point is deployed next to each storage area; the monitoring point is a communication node used to monitor task data within the storage area, and the conversion point is connected to the storage area one-to-one through the monitoring point. A dependency symbiosis model is established based on a conversion point, a connection node, a basic graph, a monitoring point, and a storage area; wherein, the dependency symbiosis model is a migration framework for migrating task data in the storage area through the conversion point.

2. The resource elastic scheduling method under a cloud computing platform according to claim 1, characterized in that, The steps for setting up a cluster of multiple virtual machines and establishing a shared space between two adjacent virtual machines include: In a cloud computing platform, a cluster of multiple virtual machines is created based on the task. Configure the topology of multiple virtual machines in the cluster, and establish a shared space between every two adjacent virtual machines in the cluster according to the topology; the same shared space belongs to the two virtual machines connected to it.

3. The resource elastic scheduling method under a cloud computing platform according to claim 1, characterized in that, The steps for monitoring the workload demand of tasks in the cluster and scaling up or down virtual machines based on the workload demand include: Monitor the load requirements of tasks in the cluster in real time using monitoring tools; Determine the target virtual machine and its remaining load based on the task's required load; The remaining load of the target virtual machine is compared with the required load, and the target virtual machine is scaled up or down based on the comparison results.

4. The resource elastic scheduling method under a cloud computing platform according to claim 3, characterized in that, The steps of comparing the remaining load of the target virtual machine with the required load, and scaling up or down the target virtual machine based on the comparison results, include: Obtain the difference between the remaining load and the required load of the target virtual machine to get the comparison result; If the comparison result is zero, then there is no need to expand or shrink the target virtual machine. If the comparison result is greater than zero, the target virtual machine is scaled down based on the comparison result; If the comparison result is less than zero, the target virtual machine will be expanded according to the comparison result.

5. The resource elastic scheduling method under a cloud computing platform according to claim 1, characterized in that, The steps for obtaining the data encoding corresponding to the dynamic data stored in the conversion point include: A monitoring cycle is set, and the monitoring points capture task data in the storage area based on the monitoring cycle. The task data is transformed into task graphs through a mapping transformation table and array diagram. The task graph is compared point by point with the base graph connecting the storage area to identify all changed points, point positions and modification types, and all changed points, point positions and modification types are treated as dynamic data. Dynamic data is serialized into data encoding and stored in the conversion point connected to the storage area; wherein, the serialization conversion includes: converting the dynamic data into binary according to a preset structure and compressing it.

6. The resource elastic scheduling method under a cloud computing platform according to claim 5, characterized in that, The steps involved in migrating data between corresponding virtual machines using a dependency symbiosis model until preset conditions are met include: Preset migration rules are used to determine the data to be migrated when the migration rules are met. Based on the data to be migrated, the migration start point and migration end point are determined. The migration start point and migration end point are the storage area where the data to be migrated is located and the destination storage area where the data to be migrated needs to be migrated, respectively. The migration rule is: when the difference in demand load between two virtual machines connected in the same shared space is greater than a preset threshold, the virtual machine corresponding to the high demand load will be migrated to the virtual machine corresponding to the low demand load. Based on the data to be migrated, determine the conversion point to connect the migration start point, and establish a communication connection between the conversion point and the migration end point; The data encoding within the conversion point is merged with the basic graph of the connection, and after merging, it is mapped to the migration endpoint through a mapping conversion table. The migration endpoint stores the migrated data. The data code generated by comparing the migrated data with the base graph is used as the first verification code, and the data code generated by comparing the data to be migrated with the base graph is used as the second verification code. The first verification code and the second verification code are checked for consistency. When the first verification code and the second verification code are consistent, the data to be migrated is deleted. At the same time, the conversion point is disconnected from the storage area where the data to be migrated is located. After the data migration is completed, it is determined whether the two virtual machines that were migrated meet the migration rules. If not, the data migration continues until the preset conditions are met.

7. A resource elastic scheduling system under a cloud computing platform, used to implement the resource elastic scheduling method under a cloud computing platform as described in any one of claims 1-6, characterized in that, It includes a settings layer, a monitoring and adjustment layer, and a data migration layer; The configuration layer is used to configure a cluster of multiple virtual machines and establish a shared space between two adjacent virtual machines. The monitoring and adjustment layer is used to monitor the demand load of tasks in the cluster and to scale up or down the virtual machines according to the demand load. The data migration layer is used to establish a dependency symbiosis model based on the shared space, and to perform data migration between corresponding virtual machines through the dependency symbiosis model until the preset conditions are met.

Images