Cross-domain PaaS application elastic expansion system, method and device based on micro-service architecture and medium

Abstract

The invention discloses a cross-domain PaaS application elastic expansion system, method and device based on a micro-service architecture and a medium, and belongs to the technical field of micro-services, and the system comprises a service registration center, a load state evaluation module, a light load state resource optimization module, an overload state resource optimization module and an execution module. The service registration center is used for receiving the micro-service information and summarizing and storing the micro-service information to form a micro-service list; the load state evaluation module is used for monitoring the index data of the micro-service in the micro-service list, and evaluating the load state of the micro-service according to the index data to obtain a load state evaluation result of the micro-service; the light load state resource optimization module is used for generating a resource recovery decision; the overload state resource optimization module is used for generating a resource extension decision; and the execution module is used for executing the resource recovery decision and the resource extension decision. According to the invention, a cross-domain resource coordination allocation mechanism is combined with service priority arbitration, so that the problem of unreasonable resource allocation is solved.

Description

Technical Field

[0001] This invention relates to the field of microservices technology, and more specifically to a cross-domain PaaS application elastic scaling system, method, device, and medium based on a microservices architecture. Background Technology

[0002] With the rapid development of information technology, microservice architecture and Platform as a Service (PaaS) have been widely adopted in the cloud computing field. Microservice architecture improves the scalability, maintainability, and flexibility of the technology stack by breaking down monolithic applications into multiple independent and deployable service units. At the same time, PaaS platforms provide developers with a complete application development, deployment, and runtime environment, effectively reducing the complexity of infrastructure management. In multi-cloud and hybrid cloud deployment environments, cross-domain microservice deployment has become the norm, requiring unified management and coordinated operation of microservices across multiple domains in different geographical locations and with different cloud service providers. To cope with changes in business load, elastic scaling technology has emerged, automatically adjusting computing resources to ensure service quality and resource utilization efficiency.

[0003] However, existing technologies have many shortcomings in the management and elastic scaling of microservices in cross-domain PaaS applications. First, traditional service registration and discovery mechanisms are mainly designed for single-domain scenarios and lack unified management capabilities for multi-domain microservices in cross-domain environments, failing to effectively support cross-domain service discovery and instance information synchronization. Second, existing load assessment methods typically use simple threshold judgments and lack load status analysis capabilities, making it difficult to accurately identify the light-load, normal, and overload states of microservices. More importantly, existing resource allocation strategies fail to fully consider business priorities and cannot conduct reasonable arbitration during resource competition, potentially causing critical business operations to be affected by improper resource allocation. Furthermore, traditional solutions lack a systematic coordination mechanism for cross-domain resource allocation, making it impossible to achieve optimized resource configuration and migration between different domains. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by this invention is: how to address the issues of unreasonable allocation of microservice resources and inability to guarantee the continuity of critical business in cross-domain environments, by constructing a cross-domain service registration center, a load status assessment module, a differentiated resource optimization module, and an execution module, to achieve unified management, load assessment, resource optimization decisions based on business priorities, and coordinated allocation of cross-domain resources for cross-domain microservices, thereby improving resource utilization efficiency and service quality.

[0006] To address the aforementioned technical problems, this invention provides the following technical solution: a cross-domain PaaS application elastic scaling system based on a microservice architecture, comprising, The system comprises a service registry, a load status assessment module, a light-load state resource optimization module, an overload state resource optimization module, and an execution module. The service registry receives microservice information automatically registered by microservices within each domain upon startup and aggregates and stores this information to form a microservice list. The load status assessment module monitors the metric data of the microservices in the microservice list and assesses the load status of the microservices based on this data, obtaining a load status assessment result. This result includes overload, normal, and light-load states. The light-load state resource optimization module, when the load status assessment result is light-load, generates a resource reclamation decision based on business priority and available resources in the current domain. The overload state resource optimization module, when the load status assessment result is overload, generates a resource expansion decision based on business priority, available resources in the current domain, and available resources in other domains. The execution module executes the resource reclamation and resource expansion decisions.

[0007] As a preferred embodiment of the cross-domain PaaS application elastic scaling system based on microservice architecture described in this invention, the service registry center includes: a cross-domain service registration interface for receiving registration requests carrying microservice information automatically sent by microservices within each domain at startup; a cross-domain service information repository for aggregating and storing microservice information from different domains, classifying and indexing the microservice information according to domain and microservice type to form a microservice list; a cross-domain service discovery interface for providing instance information of a specified microservice in different domains based on query requests from the light-load state resource optimization module, the overload state resource optimization module, and the execution module; and a cross-domain heartbeat management unit for receiving instance health status and active time sent by registered microservices in each domain, and updating the registered microservices based on the instance health status and active time.

[0008] As a preferred embodiment of the cross-domain PaaS application elastic scaling system based on microservice architecture described in this invention, the load status assessment module includes: a data acquisition unit, used to locate and collect indicator data of microservices in each domain according to the microservice list provided by the service registry; a load assessment unit, used to analyze the preprocessed indicator data using machine learning algorithms to obtain the load status assessment result of the microservices; and a data transmission unit, used to send the indicator data and the load status assessment result to the light-load state resource optimization module and the overload state resource optimization module according to a preset data format and transmission protocol.

[0009] As a preferred embodiment of the cross-domain PaaS application elastic scaling system based on microservice architecture described in this invention, the light-load state resource optimization module includes: a data input unit for receiving the indicator data and load state evaluation results sent by the load state evaluation module; a business priority division unit for dividing business priorities into high, medium, and low according to the service level agreement; a resource availability calculation unit for obtaining the capacity data of the current domain's total resource pool and calculating the capacity data of the available resource pool in the current domain based on CPU utilization, memory usage, and the capacity data of the current domain's total resource pool; and a resource reclamation unit for calculating the number of instances that can be released and generating an instance shrinkage instruction when the business priority is low; and generating a resource quota reduction instruction to proportionally reduce the upper limit of CPU and memory allocation when the business priority is medium or high.

[0010] As a preferred embodiment of the cross-domain PaaS application elastic scaling system based on microservice architecture described in this invention, the overload state resource optimization module includes: a business priority division unit, used to divide business priorities into high, medium, and low according to the service level agreement; a current domain available resource detection unit, used to detect available computing, storage, and network resources within the current domain; a cross-domain resource detection unit, used to interact with the service registry center to obtain resource availability information of domains that can provide additional resource support; a resource demand assessment unit, used to calculate the total amount of new resources required to alleviate the current overload state based on the request volume and response time sent by the load state assessment module, and obtain the resource demand assessment result; a resource expansion decision-making unit, used to formulate a resource expansion decision according to a preset expansion strategy, combined with business priorities, resource demand assessment results, current domain available resource assessment results, and other domain available resource assessment results; and a decision output unit, used to encapsulate the resource expansion decision according to a preset data format and transmission protocol, and send it to the execution module.

[0011] As a preferred embodiment of the microservice-based cross-domain PaaS application elastic scaling system described in this invention, the execution module includes: a decision receiving unit for receiving the resource reclamation decision and the resource expansion decision; a resource reclamation decision execution unit for executing instance scaling down or resource quota reduction operations according to the type of the resource reclamation decision; a resource expansion decision execution unit for determining the source domain of resource allocation and the target domain of resource expansion based on the evaluation results of available resources in other domains, and executing non-cross-domain resource allocation operations and cross-domain resource allocation operations according to the type of the resource expansion decision; and a resource pool status update unit for triggering the update of the current domain available resource pool or related cross-domain resource availability information data after the operation is executed.

[0012] As a preferred embodiment of the cross-domain PaaS application elastic scaling system based on microservice architecture described in this invention, the resource scaling decision execution unit includes: selecting one or more domains with the highest resource sufficiency as the source domain for resource allocation based on the evaluation results of available resources in other domains, and determining the domain where the currently overloaded microservice is located as the target domain for resource scaling; when the resource scaling decision is to scale within the current domain and does not require cross-domain resource allocation, calling the resource management interface of the target domain, requesting and allocating the resource amount specified in the decision, creating a specified number of new microservice instances according to the microservice deployment template, and triggering the new instances to automatically register with the service registry; when the resource scaling decision is cross-domain resource allocation and specifies the source domain and target domain, sending resource release instructions and allocation instructions to the source domain, coordinating the release of the resource amount specified in the decision in the source domain, and after confirming the successful release of resources in the source domain, calling the resource management interface in the target domain, requesting and allocating the allocated resources and the local resources specified in the decision, creating a specified number of new microservice instances according to the microservice deployment template in the target domain, and triggering the new instances to automatically register with the service registry.

[0013] This invention provides a method for elastic scaling of cross-domain PaaS applications based on a microservice architecture.

[0014] To address the aforementioned technical problems, this invention provides the following technical solution: a cross-domain PaaS application elastic scaling method based on a microservice architecture, comprising: receiving microservice information automatically registered by microservices within each domain at startup, and summarizing and storing the microservice information to form a microservice list; monitoring the indicator data of the microservices in the microservice list, evaluating the load status of the microservices based on the indicator data, and obtaining a load status evaluation result for the microservices, wherein the load status evaluation result includes overload status, normal status, and light load status; when the load status evaluation result is a light load status, generating a resource reclamation decision by combining business priority and available resources in the current domain; when the load status evaluation result is an overload status, generating a resource expansion decision by combining business priority, available resources in the current domain, and available resources in other domains; and executing the resource reclamation decision and the resource expansion decision.

[0015] The present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, characterized in that the processor executes the computer program to implement the steps of the cross-domain PaaS application elastic scaling system based on microservice architecture.

[0016] The present invention provides a computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the steps of the cross-domain PaaS application elastic scaling system based on microservice architecture.

[0017] The beneficial effects of this invention are as follows: This invention utilizes a load status assessment module to subdivide microservice load status into three categories: overload, normal, and light load, providing accurate judgment criteria for differentiated resource optimization strategies under different load conditions. The light load state resource optimization module combines business priorities and available resources in the current domain to generate resource reclamation decisions, realizing the proactive reclamation and reallocation of light load resources, improving overall resource utilization efficiency while ensuring business continuity. The overload state resource optimization module can prioritize resource expansion within the current domain based on preset expansion strategies, business priorities, and resource demand assessment results. When resources are insufficient, resources are then allocated from other domains with high resource sufficiency, and can also prioritize business needs. The system employs a multi-level arbitration mechanism to prioritize resource expansion needs for high-priority services, effectively addressing issues of unreasonable resource allocation and inability to guarantee critical business operations. This improves resource utilization efficiency and enhances system reliability and service quality. The execution module ensures the atomicity and consistency of resource reclamation and expansion operations, translating abstract resource optimization decisions into concrete resource operation instructions. It achieves end-to-end management from microservice registration, load status assessment, resource optimization decisions to execution feedback. Through a cross-domain resource coordination and allocation mechanism combined with business priority arbitration, it enhances system reliability and service quality while improving resource utilization efficiency. This provides comprehensive elastic scaling capabilities for cross-domain PaaS applications and resolves the problem of unreasonable resource allocation. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is an architecture diagram of a cross-domain PaaS application elastic scaling system based on a microservice architecture, provided as an embodiment of the present invention. Detailed Implementation

[0020] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0021] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0022] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0023] Example 1, referring to Figure 1 This is one embodiment of the present invention, which provides a cross-domain PaaS application elastic scaling system based on a microservice architecture, including: The system includes a service registry, a load status assessment module, a light-load resource optimization module, an overload resource optimization module, and an execution module.

[0024] The service registry is used to receive microservice information that is automatically registered by microservices in each domain when they start up, and to summarize and store the microservice information to form a microservice list.

[0025] The load status assessment module is used to monitor the metric data of microservices in the microservice list, assess the load status of microservices based on the metric data, and obtain the load status assessment results of microservices. The load status assessment results include overload status, normal status, and light load status.

[0026] The Light Load State Resource Optimization Module is used to generate resource reclamation decisions when the load state assessment result is light load state, by combining business priorities and the available resources in the current domain.

[0027] The overload state resource optimization module is used to generate resource expansion decisions when the load state assessment result is overloaded, by combining business priorities, available resources in the current domain, and available resources in other domains.

[0028] The execution module is used to make resource reclamation and resource expansion decisions.

[0029] It should be noted that with the development of cloud computing technology, microservice architecture and PaaS applications have been widely adopted. However, in cross-domain scenarios, existing technologies still have many shortcomings in terms of unified management, collaborative scheduling, and elastic scaling of multi-domain microservices. Furthermore, during resource allocation, existing technologies often fail to adequately consider business priorities for reasonable arbitration. When resources are scarce, the service quality of critical businesses may be severely affected by unreasonable resource allocation.

[0030] Therefore, to address the aforementioned issues of unreasonable resource allocation and inability to guarantee the quality of critical business services, a unified registration and inventory management system for microservices within each domain is implemented through a service registry center, load status assessment module, light-load resource optimization module, overload resource optimization module, and execution module. This system monitors and utilizes machine learning algorithms in real time to assess the load status of microservices; it generates precise resource reclamation or expansion decisions based on business priorities and available resources in the current and other domains; and it automates the execution of these decisions through the execution module. When resources are insufficient, resources can be allocated from other domains with sufficient resources, and arbitration is conducted based on business priorities to prioritize the needs of high-priority businesses. This effectively solves the problem of unreasonable resource allocation, ensuring the stability of critical businesses while improving overall resource utilization efficiency.

[0031] Example 2, refer to Figure 1 As one embodiment of the present invention, based on the previous embodiment, a cross-domain PaaS application elastic scaling system based on a microservice architecture is provided, comprising: In this embodiment of the invention, the service registration center includes: The cross-domain service registration interface is used to receive registration requests carrying microservice information, which are automatically sent by microservices within each domain at startup.

[0032] Specifically, the registration request is initiated by the service discovery client built into the microservice.

[0033] Specifically, microservice information includes a unique service name, service version number, service address, port number, function description, initial health status, and the domain identifier.

[0034] A cross-domain service information repository is used to aggregate and store microservice information from different domains. It categorizes and indexes microservice information according to domain and microservice type to form a microservice list, and supports querying and locating microservices from different domains.

[0035] For example, the microservice list includes: Domain: E-commerce front-end domain, Microservice type: User service, Microservice instance: User authentication service, which is responsible for user login and registration functions, such as verifying the username and password entered by the user and generating user tokens; Domain: E-commerce back-end domain, Microservice type: Order service, Microservice instance: Order creation service, which is used to process user requests to submit orders, generate order numbers, and record product information and shipping addresses in the order.

[0036] The cross-domain service discovery interface is used to provide instance information of a specified microservice in different domains based on query requests from the light-load state resource optimization module, the overload state resource optimization module, and the execution module.

[0037] Specifically, the instance information of the microservice in different domains includes the instance address, port, and health status. The instance address and port jointly identify the specific location of the microservice instance on the network, enabling other microservices or clients to accurately establish connections and communicate with it. The health status reflects the current running status of the microservice instance, helps filter out unhealthy instances, and ensures that requests are only sent to normally running instances, thereby guaranteeing service reliability.

[0038] The cross-domain heartbeat management unit is used to receive instance health status and active time from registered microservices in each domain, and update the registered microservices based on the instance health status and active time.

[0039] Specifically, the registered microservices are updated based on the instance's health status and active time, including: If an instance of a registered microservice is found to be in an unhealthy state or its active time exceeds a preset threshold, the instance information of the registered microservice will be removed from the microservice list.

[0040] It should be noted that through the collaborative work of the cross-domain service registration interface, cross-domain service information repository, cross-domain service discovery interface, and cross-domain heartbeat management unit, unified registration, classification indexing, and health status management of multi-domain microservices are achieved. Among them, the dual classification index of domain identifier and microservice type enables microservices in different geographical locations or cloud environments to be uniformly discovered and located. The cross-domain heartbeat management unit, based on the judgment mechanism of health status and active time, promptly cleans up failed instances, ensures the accuracy of the microservice list, avoids the delay and error of service discovery in cross-domain environment, and improves the reliability and efficiency of communication between microservices.

[0041] In this embodiment of the invention, the load status assessment module includes: The data acquisition unit is used to locate and collect the indicator data of microservices in each domain based on the microservice list provided by the service registry.

[0042] Specifically, the metrics data for microservices include CPU utilization, memory usage, request volume, and response time. These metrics data carry a unique identifier and domain identifier for the microservice.

[0043] The load assessment unit is used to analyze preprocessed indicator data using machine learning algorithms to obtain the load status assessment results of microservices.

[0044] Specifically, machine learning algorithms are used to analyze the preprocessed indicator data, including: Based on historical metric data, machine learning algorithms such as decision trees, support vector machines, or neural networks are used for training to build a load status assessment model, enabling the load assessment unit to assess load status. The historical metric data includes CPU utilization, memory usage, request volume, and response time.

[0045] For example, when using a decision tree to construct a load status assessment model, the load status assessment model is built based on the decision tree algorithm to classify microservice instances into light load, normal, or overload states in real time. The load status assessment model is trained based on historical indicator data and correspondingly labeled with load status tags, including light load, normal, or overload. The historical indicator data is input into the decision tree to obtain the load status assessment model. The load status assessment model generates classification rules by learning split thresholds in the features (e.g., if the response time is >200ms and the CPU utilization is >85%, it is judged as overload). The model training uses the Gini coefficient or information gain as the split criteria and limits the tree depth to prevent overfitting. After training, the load status assessment model is integrated into the load assessment unit.

[0046] The data transmission unit is used to send the indicator data and load status evaluation results to the light load state resource optimization module and the overload state resource optimization module according to the preset data format and transmission protocol.

[0047] It should be noted that the data acquisition unit acquires multi-dimensional indicator data carrying unique identifiers and domain identifiers, the load assessment unit uses machine learning algorithms to build a load status assessment model, and the data transmission unit uniformly distributes the assessment results, thus realizing intelligent and accurate assessment of the load status of cross-domain microservices. Among them, the machine learning model trained with historical data can identify load change trends and abnormal patterns, accurately distinguishing between light load, normal, and overload states. The carrying of unique identifiers and domain identifiers ensures the traceability of assessment results, providing a reliable data foundation for subsequent resource optimization decisions and improving the accuracy and real-time performance of load assessment.

[0048] In this embodiment of the invention, the light-load state resource optimization module includes: The data input unit is used to receive indicator data and load status assessment results sent by the load status assessment module.

[0049] The business priority division unit is used to divide business priorities into high, medium, and low according to the service level agreement.

[0050] The resource availability calculation unit is used to obtain the capacity data of the total resource pool in the current domain, and calculate the capacity data of the available resource pool in the current domain based on the CPU utilization, memory usage and the capacity data of the total resource pool in the current domain.

[0051] Specifically, the capacity data of the available resource pool in the current domain includes the number of available CPU cores and the size of free memory.

[0052] Specifically, the capacity data of the available resource pool within the current domain includes the number of available CPU cores and the amount of available memory, which can be represented as follows: Available CPU cores = Total CPU cores × (1 - Current CPU utilization); Available memory size = total memory size × (1 - current memory usage).

[0053] The resource reclamation unit is used to calculate the number of instances that can be released and generate instance shrinkage instructions when the business priority is low; when the business priority is medium or high, it generates resource quota reduction instructions to proportionally reduce the upper limit of CPU and memory allocation.

[0054] Specifically, when the business priority is low, the number of instances that can be released is calculated, and an instance scaling-down instruction is generated, including: Based on request volume data and historical request volume data, the minimum number of instances required to maintain the service over a future period is predicted using a Long Short-Term Memory (LSTM) network.

[0055] Based on the total number of currently running instances and the predicted minimum number of instances, calculate the number of instances that can be safely released, and generate an instance shrinking instruction. The instance shrinking instruction contains a list of unique identifiers of the instances to be released.

[0056] Furthermore, the number of instances that can be safely released can be calculated as follows: The number of instances that can be safely released = the current number of instances - ceil(predicted minimum number of instances × (1 + safety factor)); Here, ceil is rounded up; the safety factor is set to 0.1~0.2 to prevent insufficient resources due to prediction errors.

[0057] Specifically, when the business priority is medium or high, a resource quota reduction instruction is generated to proportionally reduce the upper limit of CPU and memory allocation, including: Obtain the microservice metric data sent by the load status assessment module, as well as the current CPU and memory allocation limits allocated to the microservice instance.

[0058] The analysis is based on the CPU utilization and memory usage compared to preset light load thresholds. The reduction ratio of CPU and memory quotas is calculated based on the difference between these thresholds and a preset safety redundancy coefficient. The preset light load threshold is determined through statistical analysis of historical load data. The safety redundancy coefficient is used to avoid excessive resource reclamation and is typically set as follows: CPU redundancy coefficient is 0.1~0.2, and memory redundancy coefficient is 0.1~0.2.

[0059] A resource quota reduction instruction is generated based on the reduction ratio of CPU quota and memory quota. The resource quota reduction instruction is used to specify the reduction of the upper limit of CPU allocation and memory allocation of microservice instances according to the reduction ratio of CPU quota and memory quota.

[0060] Furthermore, the specific calculation of the reduction ratio can be expressed as follows: The reduction ratio = max(0, (current usage rate - light load threshold) / (1 - light load threshold)) × (1 - safety redundancy coefficient).

[0061] For example, if the CPU utilization is 15%, the light load threshold is 20%, and the safety factor is 0.1, then the reduction ratio = max(0, (0.15 - 0.2) / 0.8) × 0.9 = 0 (no reduction).

[0062] It should be noted that by prioritizing business operations and implementing differentiated resource reclamation strategies, refined resource reclamation based on business importance has been achieved. Specifically, by using a long short-term memory network to predict future service demand, it ensures that low-priority businesses do not affect service continuity when scaling down. For medium- and high-priority businesses, a resource quota reduction strategy is adopted to maintain a stable number of instances while reclamating redundant resources. The introduction of a security redundancy coefficient avoids performance risks caused by excessive reclamation, and the analysis of the gap between CPU utilization and memory usage and the light load threshold ensures the scientific nature of reclamation decisions, thereby improving resource utilization efficiency and ensuring the service quality of critical businesses.

[0063] In this embodiment of the invention, the overload state resource optimization module includes: The business priority division unit is used to divide business priorities into high, medium, and low according to the service level agreement.

[0064] The Current Domain Available Resources Detection Unit is used to detect the available computing, storage, and network resources within the current domain.

[0065] Specifically, the available computing, storage, and network resources within the current domain include the number of idle CPU cores, remaining memory capacity, available disk space, and idle network bandwidth.

[0066] The cross-domain resource detection unit is used to interact with the service registry to obtain information on the availability of resources in domains that can provide additional resource support.

[0067] Specifically, the resource availability information for domains that can provide additional resource support includes the number of idle CPU cores, remaining memory capacity, available disk space, and idle network bandwidth for each domain.

[0068] The resource demand assessment unit is used to calculate the total amount of new resources needed to alleviate the current overload state based on the number of requests and response times sent by the load status assessment module, and obtain the resource demand assessment results.

[0069] Specifically, the total amount of new resources includes the required number of CPU cores, memory capacity, disk space, and network bandwidth.

[0070] Furthermore, the total amount of new resources required to alleviate the current overload state is calculated, including: Based on the current request volume exceeding the processing capacity of the microservice instance group and the degree to which the response time exceeds the preset threshold, the total additional processing capacity is quantified.

[0071] Invoke the predefined microservice capacity model to obtain the baseline processing capacity and corresponding baseline resource requirements of a single microservice instance under the target response time constraint.

[0072] Based on the ratio of the total processing capacity to be added to the baseline processing capacity of a single instance, the required number of new instances is determined, and this number is multiplied by the baseline resource requirement of a single instance to obtain the total amount of new resources, which includes the total number of CPU cores and the total memory capacity required.

[0073] Furthermore, a predefined microservice capacity model is established, including: Based on historical performance stress test data and production environment runtime monitoring data, we analyze the relationship between resource utilization (such as CPU utilization and memory usage) and response time of microservice instances under different request loads (such as request volume), thereby establishing a mapping relationship with request load and target response time as input and required resource quotas (such as CPU cores and memory size) as output.

[0074] The resource expansion decision-making unit is used to make resource expansion decisions based on preset expansion strategies, combined with business priorities, resource demand assessment results, current domain available resource assessment results, and other domain available resource assessment results.

[0075] Specifically, the preset expansion strategies include: When the assessment results of available resources in the current domain meet or partially meet the total amount of new resources required, the decision to expand resources within the current domain should be prioritized.

[0076] When the current domain's available resource assessment results cannot meet the required total amount of new resources, one or more domains with the highest resource sufficiency are selected as the source domain for resource allocation based on the assessment results of available resources in other domains, and the domain where the microservice currently experiencing overload is located is determined as the target domain for resource expansion.

[0077] When multiple overloaded microservices compete for resources in the current domain and other domains, arbitration is conducted based on business priority information. Priority is given to satisfying the resource expansion needs of high-priority services, followed by medium-priority and low-priority resource expansion needs.

[0078] Specifically, resource expansion decisions include the source and target domains for resource allocation, the number of new microservice instances, and specific parameters for resource allocation, including the number of CPU cores, memory capacity, disk space, and network bandwidth.

[0079] Furthermore, the specific details of resource expansion decisions include: The source and target domains for resource allocation are used to specify from which source domains resources are allocated to which target domain.

[0080] The number of new microservice instances is used to calculate and determine the number of new microservice instances that need to be deployed in the target domain, based on the estimated total amount of new resources and the standard or minimum resource requirements of a single microservice instance. Specifically, it can be expressed as: Number of instances = ceil (estimated total new resources / standard resource quota per instance).

[0081] The specific parameters for resource allocation are used to allocate specific resource quotas to each newly added microservice instance, including the number of CPU cores, memory capacity, disk space, and network bandwidth.

[0082] The decision output unit is used to encapsulate the resource expansion decision according to the preset data format and transmission protocol, and send it to the execution module to trigger the corresponding resource expansion operation.

[0083] It should be noted that cross-domain resource allocation and optimization are achieved through multi-domain resource detection, demand assessment, and expansion decision-making based on business priorities. Specifically, by assessing the resource sufficiency of the current domain and other domains, resources in the current domain are prioritized. When resources are insufficient, the domain with the highest sufficiency is automatically selected for cross-domain allocation. Business priority arbitration ensures reasonable allocation when multiple overloaded microservices compete for resources, and high-priority services are given priority expansion permissions. The multi-level decision-making logic of preset expansion strategies avoids blind cross-domain allocation, reduces network latency and data transmission costs, and improves resource allocation efficiency and business response speed.

[0084] In this embodiment of the invention, the execution module includes: The decision receiving unit is used to receive resource recovery decisions and resource expansion decisions.

[0085] The resource recycling decision execution unit is used to execute instance scaling down or resource quota reduction operations based on the type of resource recycling decision.

[0086] Specifically, depending on the type of resource recycling decision, perform instance scaling down or resource quota reduction operations, including: If the resource reclamation decision is an instance shrinkage instruction, then based on the list or number of unique instance identifiers specified in the instruction, the resource management interface of the target domain is called to terminate the specified microservice instance and notify the service registry to update the microservice list.

[0087] If the resource reclamation decision is a resource quota reduction instruction, then according to the reduction ratio and microservice identifier specified in the instruction, the resource management interface of the target domain is called to proportionally reduce the CPU allocation limit and memory allocation limit of all running instances of the corresponding microservice.

[0088] The resource expansion decision execution unit is used to determine the source domain for resource allocation and the target domain for resource expansion based on the assessment results of available resources in other domains, and to execute resource allocation operations that do not require cross-domain allocation and cross-domain resource allocation operations according to the type of resource expansion decision.

[0089] Specifically, the resource expansion decision-making and execution unit includes: Based on the assessment results of available resources in other domains, select one or more domains with the highest resource sufficiency as the source domain for resource allocation, and determine the domain where the microservice currently experiencing overload resides as the target domain for resource expansion.

[0090] When the resource expansion decision is to expand within the current domain and does not require cross-domain resource allocation, the resource management interface of the target domain is called to request and allocate the amount of resources specified in the decision. Based on the microservice deployment template, a specified number of new microservice instances are created, triggering the new instances to automatically register with the service registry.

[0091] When the resource expansion decision is cross-domain resource allocation and a source domain and a target domain are specified, a resource release instruction and an allocation instruction are sent to the source domain. The source domain is coordinated to release the specified number of resources. After confirming the successful release of resources in the source domain, the resource management interface is called in the target domain to apply for and allocate the allocated resources and the local resources specified in the decision. According to the deployment template of microservices in the target domain, a specified number of new microservice instances are created, and the new instances are automatically registered with the service registry.

[0092] Furthermore, the resources specified in the decision include the number of CPU cores, memory capacity, disk space, and network bandwidth.

[0093] Furthermore, the specified amount of resources includes CPU and memory.

[0094] The resource pool status update unit is used to trigger the update of the available resource pool in the current domain or related cross-domain resource availability information data after the operation is performed.

[0095] It should be noted that the closed-loop mechanism of decision reception, classified execution, and status update enables precise execution and real-time feedback of resource reclamation and expansion decisions. The resource reclamation decision execution unit selects the execution strategy of instance scaling down or quota reduction based on the instruction type, ensuring the accuracy of the reclamation operation. The resource expansion decision execution unit supports both local expansion and cross-domain allocation modes, and the atomicity and consistency of cross-domain allocation are guaranteed by the confirmation of source domain resource release. The resource pool status update unit synchronizes resource change information in a timely manner, providing accurate resource base data for the next round of decisions, ensuring the reliability and timeliness of resource allocation, and avoiding resource leakage or service interruption caused by execution failure.

[0096] In summary, this invention utilizes a load status assessment module to categorize microservice load status into three types: overload, normal, and light load, providing accurate judgment criteria for differentiated resource optimization strategies under different load conditions. The light load resource optimization module combines business priorities with available resources in the current domain to generate resource reclamation decisions, enabling proactive reclamation and reallocation of light load resources, improving overall resource utilization efficiency while ensuring business continuity. The overload resource optimization module, based on preset expansion strategies, business priorities, and resource demand assessment results, prioritizes resource expansion within the current domain, and then allocates resources from other domains with high resource sufficiency when resources are insufficient, and can also perform resource allocation based on business priorities. Arbitration prioritizes the resource expansion needs of high-priority business operations, effectively resolving issues of unreasonable resource allocation and inability to guarantee critical business operations, improving resource utilization efficiency, and enhancing system reliability and service quality. The execution module ensures the atomicity and consistency of resource reclamation and expansion operations, translating abstract resource optimization decisions into concrete resource operation instructions. It achieves end-to-end management from microservice registration, load status assessment, resource optimization decisions to execution feedback. Through cross-domain resource coordination and allocation mechanisms combined with business priority arbitration, it enhances system reliability and service quality while improving resource utilization efficiency, providing comprehensive elastic scaling capabilities for cross-domain PaaS applications, and resolving the problem of unreasonable resource allocation.

[0097] Example 3 is an embodiment of the present invention, which provides a method for elastic scaling of cross-domain PaaS applications based on a microservice architecture, including: receiving microservice information automatically registered by microservices in each domain at startup, and summarizing and storing the microservice information to form a microservice list; monitoring the indicator data of microservices in the microservice list, evaluating the load status of microservices based on the indicator data, and obtaining the load status evaluation result of microservices, including overload status, normal status, and light load status; when the load status evaluation result is light load status, generating a resource reclamation decision by combining business priority and available resources in the current domain; when the load status evaluation result is overload status, generating a resource expansion decision by combining business priority, available resources in the current domain, and available resources in other domains; and executing the resource reclamation decision and the resource expansion decision.

[0098] This embodiment also provides an electronic device suitable for a cross-domain PaaS application elastic scaling system based on a microservice architecture, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the cross-domain PaaS application elastic scaling system based on a microservice architecture as proposed in the above embodiment.

[0099] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, it implements the cross-domain PaaS application elastic scaling system based on microservice architecture as proposed in the above embodiments.

[0100] The storage medium proposed in this embodiment and the cross-domain PaaS application elastic scaling system based on microservice architecture proposed in the above embodiments belong to the same inventive concept. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0101] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0102] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A cross-domain PaaS application elastic scaling system based on microservice architecture, characterized in that, include: The system includes a service registry, a load status assessment module, a light-load resource optimization module, an overload resource optimization module, and an execution module. The service registry is used to receive microservice information that is automatically registered by microservices in each domain when they start up, and to summarize and store the microservice information to form a microservice list; The load status assessment module is used to monitor the indicator data of the microservices in the microservice list, assess the load status of the microservices based on the indicator data, and obtain the load status assessment result of the microservices. The load status assessment result includes overload status, normal status and light load status. The light-load state resource optimization module is used to generate a resource reclamation decision when the load state assessment result is light-load state, by combining business priority and the available resources in the current domain. The overload state resource optimization module is used to generate a resource expansion decision when the load state assessment result is an overload state, by combining business priority, available resources in the current domain, and available resources in other domains. The execution module is used to execute the resource reclamation decision and the resource expansion decision.

2. The cross-domain PaaS application elastic scaling system based on microservice architecture as described in claim 1, characterized in that... The service registration center includes: The cross-domain service registration interface is used to receive registration requests carrying microservice information that are automatically sent by microservices within each domain at startup. A cross-domain service information repository is used to aggregate and store microservice information from different domains, and to classify and index the microservice information according to domain and microservice type to form a microservice list; The cross-domain service discovery interface is used to provide instance information of a specified microservice in different domains based on the query requests from the light-load state resource optimization module, the overload state resource optimization module, and the execution module. The cross-domain heartbeat management unit is used to receive instance health status and active time from registered microservices in each domain, and update the registered microservices based on the instance health status and active time.

3. The cross-domain PaaS application elastic scaling system based on microservice architecture as described in claim 2, characterized in that... The load status assessment module includes: The data acquisition unit is used to locate and collect the indicator data of microservices in each domain based on the microservice list provided by the service registry. The load assessment unit is used to analyze preprocessed indicator data using machine learning algorithms to obtain the load status assessment results of microservices. The data transmission unit is used to send the indicator data and the load status evaluation results to the light load state resource optimization module and the overload state resource optimization module according to a preset data format and transmission protocol.

4. The cross-domain PaaS application elastic scaling system based on microservice architecture as described in claim 3, characterized in that... The light-load state resource optimization module includes: The data input unit is used to receive the indicator data and the load status assessment result sent by the load status assessment module; The business priority division unit is used to divide business priorities into high, medium, and low according to the service level agreement; The resource availability calculation unit is used to obtain the capacity data of the total resource pool in the current domain, and calculate the capacity data of the available resource pool in the current domain based on the CPU utilization, memory usage and the capacity data of the total resource pool in the current domain. The resource reclamation unit is used to calculate the number of instances that can be released and generate instance shrinkage instructions when the business priority is low; when the business priority is medium or high, it generates resource quota reduction instructions to proportionally reduce the upper limit of CPU and memory allocation.

5. The cross-domain PaaS application elastic scaling system based on microservice architecture as described in claim 4, characterized in that... The overload state resource optimization module includes: The business priority division unit is used to divide business priorities into high, medium, and low according to the service level agreement; The Current Domain Available Resources Detection Unit is used to detect the available computing, storage, and network resources within the current domain. The cross-domain resource detection unit is used to interact with the service registry to obtain information on the availability of resources in domains that can provide additional resource support. The resource demand assessment unit is used to calculate the total amount of new resources needed to alleviate the current overload state based on the number of requests and response time sent by the load status assessment module, and obtain the resource demand assessment result. The resource expansion decision-making unit is used to make resource expansion decisions based on preset expansion strategies, combined with business priorities, resource demand assessment results, current domain available resource assessment results, and other domain available resource assessment results. The decision output unit is used to encapsulate resource expansion decisions according to preset data formats and transmission protocols, and send them to the execution module.

6. The cross-domain PaaS application elastic scaling system based on microservice architecture as described in claim 5, characterized in that... The execution module includes: A decision receiving unit is used to receive the resource recovery decision and the resource expansion decision; The resource recycling decision execution unit is used to perform an instance scaling down operation or a resource quota reduction operation according to the type of the resource recycling decision. The resource expansion decision execution unit is used to determine the source domain for resource allocation and the target domain for resource expansion based on the assessment results of available resources in other domains, and to execute non-cross-domain resource allocation operations and cross-domain resource allocation operations according to the type of the resource expansion decision. The resource pool status update unit is used to trigger the update of the available resource pool in the current domain or related cross-domain resource availability information data after the operation is performed.

7. The cross-domain PaaS application elastic scaling system based on microservice architecture as described in claim 6, characterized in that... The resource expansion decision execution unit includes: Based on the assessment results of available resources in other domains, one or more domains with the highest resource sufficiency are selected as the source domains for resource allocation, and the domain where the microservice currently experiencing overload is located is determined as the target domain for resource expansion. When the resource expansion decision is to expand within the current domain and no cross-domain resource allocation is required, the resource management interface of the target domain is called to request and allocate the amount of resources specified in the decision. According to the microservice deployment template, a specified number of new microservice instances are created, and the new instances are automatically registered with the service registry. When the resource expansion decision is cross-domain resource allocation and a source domain and a target domain are specified, a resource release instruction and an allocation instruction are sent to the source domain to coordinate the release of the specified number of resources in the source domain. After confirming the successful release of resources in the source domain, the resource management interface is called in the target domain to apply for and allocate the allocated resources and the local resources specified in the decision. According to the deployment template of microservices in the target domain, a specified number of new microservice instances are created, and the new instances are automatically registered with the service registry.

8. A method for elastic scaling of cross-domain PaaS applications based on microservice architecture, using the cross-domain PaaS application elastic scaling system based on microservice architecture as described in any one of claims 1 to 7, characterized in that, include: Receive microservice information that is automatically registered by microservices in each domain at startup, and summarize and store the microservice information to form a microservice list; Monitor the indicator data of the microservices in the microservice list, evaluate the load status of the microservices based on the indicator data, and obtain the load status evaluation results of the microservices. The load status evaluation results include overload status, normal status and light load status. When the load status assessment result is a light load state, a resource reclamation decision is generated by combining the service priority and the available resources in the current domain. When the load status assessment result is an overload state, a resource expansion decision is generated by combining the service priority, the current available resources in the domain, and the available resources in other domains. Execute the resource recovery decision and the resource expansion decision.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the cross-domain PaaS application elastic scaling system based on microservice architecture as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the cross-domain PaaS application elastic scaling system based on microservice architecture as described in any one of claims 1 to 7.

Images